Control of Air Pollution from Motor
Vehicles: Tier 3 Motor Vehicle Emission
and Fuel Standards Final Rule
Regulatory Impact Analysis
&EPA
United States
Environmental Protection
Agency
-------
Control of Air Pollution from Motor
Vehicles: Tier 3 Motor Vehicle Emission
and Fuel Standards Final Rule
Regulatory Impact Analysis
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
&EPA
United States
Environmental Protection
Agency
EPA-420-R-14-005
March 2014
-------
Table of Contents
EXECUTIVE SUMMARY
Executive Summary ES-1
CHAPTER 1 VEHICLE PROGRAM TECHNOLOGICAL FEASIBILITY
1.1 Introduction 1-1
1.2 FTP NMOG+NOx Feasibility 1-1
1.3 SFTP NMOG+NOx Feasibility 1-28
1.4 Technology Description for NMOG+NOX Control 1-30
1.5 PM Feasibility 1-42
1.6 Evaporative Emissions Feasibility 1-52
1.7 ORVR for Complete HDGVs over 10,000 Ibs GVWR 1-63
1.8 Onboard Diagnostics for Vehicles less than 14,000 Ibs GVWR 1-64
CHAPTER 2 VEHICLE PROGRAM COSTS
2.1 Changes to Vehicle Costs between Proposed and Final Rules 2-1
2.2 General Methodology 2-2
2.3 Individual Technology Costs 2-11
2.4 Evaporative Emission, Canister Bleed, and Leak Controls 2-20
2.5 Vehicle Package Costs 2-27
2.6 Operating Costs 2-37
2.7 Vehicle Program Costs 2-41
CHAPTER 3 ESTABLISHING NEW EMISSION TEST FUEL PARAMETERS
3.1 Assessment of Current Gasoline Properties 3-1
3.2 Gasoline Emission Test Fuel Specifications 3-18
3.3 Changes to ASTM Test Methods 3-20
CHAPTER 4 FUEL PROGRAM FEASIBILITY
4.1 Overview of Refining Operations 4-1
4.2 Feasibility of Removing Sulfur from Gasoline 4-4
-------
4.3 Sulfur Credits 4-26
4.4 Sulfur Level Analysis 4-28
4.5 Lead Time Assessment 4-33
CHAPTER 5 FUEL PROGRAM COSTS
5.1 Methodology 5-1
5.2 Estimated Tier 3 Sulfur Control Costs 5-60
5.3 Other Cost Studies 5-73
5.4 Projected Energy Impacts and Impacts on Permitting 5-79
5.5 Fuel Quality Requirements for Denatured Fuel Ethanol 5-90
5.6 Gasoline Additives 5-90
5.7 Downstream Pentane Blending 5-91
CHAPTER 6 HEALTH AND ENVIRONMENTAL EFFECTS ASSOCIATED WITH
EXPOSURE TO CRITERIA AND TOXIC POLLUTANTS
6.1 Health Effects of Criteria and Toxic Pollutants 6-1
6.2 Environmental Effects of Criteria and Toxic Pollutants 6-19
CHAPTER 7 IMPACTS OF THE RULE ON EMISSIONS AND AIR QUALITY
7.1 Criteria and Toxic Pollutant Emission Impacts 7-1
7.2 Criteria and Toxic Pollutant Air Quality Impacts 7-52
7.3 Greenhouse Gas Emission Impacts 7-128
CHAPTER 7 APPENDIX: ADDITIONAL AIR TOXICS EMISSIONS AND AIR
QUALITY MODELING RESULTS
7A.1 Air Toxics Emissions 7A-1
7A.2 Seasonal Air Toxics Air Quality Modeling Results for 2018 7A-3
7A.3 Seasonal Air Toxics Air Quality Modeling Results for 2030 7A-10
CHAPTER 8 COMPARISON OF PROGRAM COSTS TO PROGRAM EMISSION
REDUCTIONS AND AIR QUALITY BENEFITS
8.1 Cost-Benefit Analysis 8-1
11
-------
CHAPTER 9 ECONOMIC IMPACT ANALYSIS
9.1 Introduction 9-1
9.2 Impacts on Vehicle Manufacturing Sector 9-2
9.3 Impacts on Petroleum Refinery Sector 9-8
CHAPTER 10 FINAL REGULATORY FLEXIBILITY ANALYSIS
10.1 Introduction 10-1
10.2 Overview of the Regulatory Flexibility Act 10-1
10.3 Need for the Rulemaking and Rulemaking Objectives 10-2
10.4 Definition and Description of Small Entities 10-3
10.5 Summary of Small Entities to Which the Rulemaking Will Apply 10-4
10.6 Related Federal Rules 10-5
10.7 Reporting, Recordkeeping, and Other Compliance Requirements 10-5
10.8 Regulatory Alternatives 10-6
10.9 Economic Effects 10-20
in
-------
List of Acronyms
A/F
AAM
ABT
ACS
AGO
AHS
AIRS
AML
ANPRM
API
ASTM
bbl
BCA
BenMAP
BTU
C-R
CAA
CAP
CARB
CASAC
CBI
DF
CG
CMAQ
CML
CO
C02
COI
COPD
cpsi
CR
CRC
CRDM
DMC
DOE
EO
E10
E15
ECA
EGR
EHC
EIA
EISA
EPA or Agency
air/fuel ratio
Alliance of Automobile Manufacturers
averaging, banking, and trading
American Cancer Society
atmospheric gasoil
U.S. Census Bureau's American Housing Survey
Aerometric Information Retrieval System
acute myeloid leukemia
Advanced Notice of Proposed Rulemaking
American Petroleum Institute
American Society for Testing and Materials
barrel
benefit-cost analysis
Environmental Benefits Mapping and Analysis Program
British Thermal Unit
concentration response
Clean Air Act
Compliance Assurance Program (2000)
California Air Resources Board
Clean Air Science Advisory Committee
confidential business information
Deterioration Factor
conventional gasoline
Community Multiscale Air Quality model
chronic myeloid leukemia
carbon monoxide
carbon dioxide
cost of illness
chronic obstructive pulmonary disease
cells per square inch
concentration-response
Coordinating Research Council
Climatological Regional Dispersion Model
direct manufacturing costs
U.S. Department of Energy
ethanol-free gasoline
gasoline containing 10 percent ethanol by volume
gasoline containing 15 percent ethanol by volume
Emission Control Area
exhaust gas recirculation
electrically heated catalyst
Energy Information Administration
Energy Independence and Security Act of 2007
U.S. Environmental Protection Agency
IV
-------
EPAct
ERIC
ESPN
EvOH
FBP
FCC
FTP
GC/MS
GDI
GDP
GPA
GVWR
HAP
HAPEM
HC
HCUP
HDGV
HDV
HEGO
HEI
I/M
IBP
ICI
ICM
IRFA
IRIS
LCO
LOT
LDV
LEV
LM
LML
LPG
MDPV
MECA
MLE
MRAD
MSAT
MSAT2
MSCF
MTBE
MY
NAAQS
NAC
NAICS
NAPAP
Energy Policy Act of 2005
Emissions Reduction and Intercept Control (system)
EPA speciation network
ethyl vinyl alcohol
final boiling point
fluidized catalytic cracker
Federal Test Procedure
gas chromatography/mass spectrometry
gasoline direct injection
gross domestic product
Geographic Phase-in Area
gross vehicle weight rating
Hazardous Air Pollutant
Hazardous Air Pollutant Exposure Model
hydrocarbon
Healthcare Cost and Utilization Program
heavy-duty gasoline vehicle
heavy-duty vehicle
heated exhaust gas oxygen (sensor)
Health Effects Institute
inspection/maintenance
initial boiling point
independent commercial importer
indirect cost multiplier
Initial Regulatory Flexibility Analysis
Integrated Risk Information System
light cycle oil
light-duty truck
light-duty vehicle
low emission vehicle
locomotive and marine diesel fuel
lowest measured level
liquid petroleum gas
medium-duty passenger vehicle
Manufacturers of Emission Controls Association
maximum likelihood estimate
minor restricted activity days
mobile source air toxic
Regulations for Control of Hazardous Air Pollutants from Mobile Sources,
72 FR 8428, 2/26/07
thousand standard cubic feet
methyl tertiary-butyl ether
model year
National Ambient Air Quality Standards
NOx adsorption catalyst
North American Industrial Classification System
National Acid Precipitation Assessment Program
-------
NATA
NEMA
NESHAP
NFRAQS
NGL
NLEV
NMHC
NMMAPS
NMOG
NO2
NOX
NPC
NPRA
NPRM
NRC
NSR
OAQPS
OAR
OBD
OC/EC
OMB
QMS
ORNL
OSC
OSTP
OTAG
PADD
PAN
PCM
Pd
PFI
PGM
PM
PNGV
POM
ppm
PSD
Pt
R+M/2
R&D
REL
RFA
RfC
RfD
RFG
RFS2
National-Scale Air Toxics Assessment
Northeast Mid-Atlantic
National Emissions Standards for Hazardous Air Pollutants
Northern Front Range Air Quality Study
natural gas liquids
national low emission vehicle
non-methane hydrocarbons
National Morbidity, Mortality, and Air Pollution Study
non-methane organic gases
nitrogen dioxide
oxides of nitrogen
National Petroleum Council
National Petrochemical & Refiners Association
Notice of Proposed Rulemaking
National Research Council
New Source Review
Office of Air Quality Planning and Standards
EPA's Office of Air and Radiation
on-board diagnostics
organic carbon/elemental carbon
Office of Management and Budget
Office of Mobile Sources
Oak Ridge National Laboratory
oxygen storage components
(White House) Office of Science and Technology Policy
Ozone Transport Assessment Group
Petroleum Administrative Districts for Defense
peroxy acetyl nitrate
powertrain control module
palladium
port fuel injection
platinum group metals
particulate matter
Partnership for a New Generation of Vehicles
polycyclic organic matter
part per million
Prevention of Significant Deterioration
platinum
average octane, or antiknock index
research and development
reference exposure level
Regulatory Flexibility Act
reference concentration
reference dose
reformulated gasoline
Renewable Fuel Standard Program, 75 FR 14670, 3/26/2010
VI
-------
Rh
ROI
ROTR
RPE
RRF
RVP
S-R
S&PDRI
SAB
SBA
SBAR or the Panel
SBREFA
SCR
SER
SFTP
SIC
SIGMA
SIP
SMAT
S02
SOX
SRU
SULEV
SVM
SVOC
SwRI
T10
T50
T90
TC
TGDI
THC
TOG
TW
UAM
UCL
UEGO
ULEV
UMRA
UV
VGO
VTB
VMT
VNA
voc
VSL
rhodium
return on investment
Regional Ozone Transport Rule
retail price equivalent
relative reduction factor
Reid vapor pressure
Source-Receptor Matrix
Standard & Poor's Data Research International
Science Advisory Board
U.S. Small Business Administration
Small Business Advocacy Review Panel
Small Business Regulatory Enforcement Fairness Act
selective catalytic reduction
Small Entity Representative
Supplemental Federal Test Procedure
Standard Industrial Classification
Society of Independent Gasoline Marketers of America
State Implementation Plan
Speciated Modeled Attainment Test
sulfur dioxide
oxides of sulfur
sulfur recovery unit
super ultra low emission vehicle
small volume manufacturer (of vehicles)
semivolatile organic compound
Southwest Research Institute
average temperature at which 10 percent of gasoline is distilled
average temperature at which 50 percent of gasoline is distilled
average temperature at which 90 percent of gasoline is distilled
total technology costs
turbocharged gasoline direct injection
total hydrocarbons
total organic gases
test weight
Urban Airshed Model
upper confidence limit
universal exhaust gas oxygen (sensor)
ultra low emission vehicle
Unfunded Mandates Reform Act
ultra violet
vacuum gasoil
vacuum tower bottoms
vehicle miles traveled
Voronoi Neighbor Averaging
volatile organic compound
value of a statistical life
vn
-------
WLD work loss days
WTP willingness to pay
Vlll
-------
This Page Intentionally Left Blank
IX
-------
Executive Summary
EPA is adopting a comprehensive program to address air pollution from passenger cars
and trucks. The final program, known as "Tier 3," will establish more stringent vehicle emissions
standards and reduce the sulfur content of gasoline, considering the vehicle and its fuel as an
integrated system The final Tier 3 standards will reduce concentrations of multiple air pollutants
(ambient concentrations of ozone, particulate matter (PM), nitrogen dioxide (NO2), and mobile
source air toxics (MSATs)) across the country and help state and local agencies in their efforts to
attain and maintain health-based National Ambient Air Quality Standards (NAAQS).
This Regulatory Impact Analysis provides technical, economic, and environmental
analyses of the new standards. Chapter 1 contains our technical feasibility justification for the
final vehicle emission standards, and Chapter 2 contains the estimated costs of the final vehicle
standards. In addition to the vehicle emission and gasoline standards, we are adopting an update
to the specifications of the emission test fuel with which vehicles demonstrate compliance with
emissions standards; our analysis of the emission test fuel parameter changes is found in Chapter
3. Chapters 4 and 5 contain our technical feasibility and cost analyses for the final gasoline
sulfur standards, respectively. Chapter 6 describes the health and welfare effects associated with
the air pollutants that will be impacted by the rule. Chapter 7 describes our analysis of the
emission and air quality impacts of the Tier 3 rule. Our estimates of the program-wide costs, the
societal benefits, and the cost per ton of emissions reduced due to the final Tier 3 program are
presented in Chapter 8. Chapter 9 contains our analysis of the final rule's economic impacts, and
Chapter 10 provides the results of our small business final regulatory flexibility analysis.
Tier 3 Standards
Vehicle Emission Standards
The Tier 3 standards include light- and heavy-duty vehicle tailpipe emission standards
and evaporative emission standards.
Light-Duty Vehicle, Light-Duty Truck, and Medium-Duty Passenger Vehicle Tailpipe
Emission Standards
The standards in this category apply to all light-duty vehicles (LDVs, or passenger cars),
light-duty trucks (LDTls, LDT2s, LDT3s, and LDT4s) and Medium-Duty Passenger Vehicles,
or MDPVs. The new standards are for the sum of NMOG and NOx emissions, presented as
NMOG+NOx, and for PM. For these pollutants, the standards are measured on test procedures
that represent a range of vehicle operation, including the Federal Test Procedure (or FTP,
simulating typical driving) and the Supplemental Federal Test Procedure (or SFTP, a composite
test simulating higher temperatures, higher speeds, and quicker accelerations).
The FTP and SFTP NMOG+NOx standards are fleet-average standards, meaning that a
manufacturer will calculate the weighted average emissions of the vehicles it sells in each model
year and compare that average to the applicable standard for that model year. The fleet average
ES-1
-------
standards for NMOG+NOx evaluated over the FTP will begin in MY 2017 and then decline
through MY 2025, as summarized in Table ES-1. Similarly, the NMOG+NOx standards
measured over the SFTP will also be fleet-average standards, declining from MY 2017 until MY
2025, as shown in Table ES-2.
Table ES-1 LDV, LDT, and MDPV Fleet Average NMOG+NOX FTP Standards (mg/mi)
LDV/LDTlb
LDT2,3,4 and
MDPV
Model Year
2017a
86
101
2018
79
92
2019
72
83
2020
65
74
2021
58
65
2022
51
56
2023
44
47
2024
37
38
2025
and later
30
30
" For vehicles above 6000 Ibs GVWR, the fleet average standards will apply beginning in MY 2018.
6 These standards will apply for a 150,000 mile useful life. Manufacturers can choose to certify their LDVs and
LDVls to a useful life of 120,000 miles. If any of these families are certified to the shorter useful life, a
proportionally lower numerical fleet-average standard will apply, calculated by multiplying the respective
150,000 mile standard by 0.85 and rounding to the nearest mg.
Table ES-2 LDV, LDT, and MDPV Fleet-Average NMOG+NOx SFTP Fleet Average
Standards (mg/mi)
NMOG + NOX
Model Year
2017a
103
2018
97
2019
90
2020
83
2021
77
2022
70
2023
63
2024
57
2025
and later
50
a For vehicles above 6000 Ibs GVWR, the fleet average standards will apply beginning in MY 2018.
The PM standard on the FTP for certification testing is 3 mg/mi for all vehicles and for
all model years. Manufacturers can phase in their vehicle models as a percent of sales through
MY 2022. The FTP PM standards will apply to each vehicle separately (i.e., not as a fleet
average). The program also includes a separate FTP PM requirement of 6 mg/mi for the testing
of in-use vehicles that will apply during the percent phase-in period only. Table ES-3 presents
the FTP certification and in-use PM standards and the phase-in percentages.
ES-2
-------
Table ES-3 Phase-In for FTP PM Standards
Phase-In
(percent of U.S. sales)
Certification Standard
(mg/mi)
In-Use Standard
(mg/mi)
2017a
20
3
6
2018
20
3
6
2019
40
3
6
2020
70
3
6
2021
100
3
6
2022 and
later
100
3
3
a Forvehicles above 6000 Ibs GVWR, the FTP PM standards will apply beginning in MY 2018.
The Tier 3 program also includes certification PM standards evaluated over the SFTP
(specifically the US06 component of the SFTP procedure) of 10 mg/mi for MYs 2017 and MY
2018, and a single final standard of 6 mg/mi for MY 2019 and later. For MYs 2019 through
2023, an in-use standard of 10 mg/mi will also apply.
Heavy-Duty Tailpipe Emission Standards
There are new Tier 3 exhaust emissions standards for complete heavy-duty vehicles
(HDVs) between 8,501 and 14,000 Ib GVWR. Vehicles in this GVWR range are often referred
to as Class 2b (8,501-10,000 Ib) and Class 3 (10,001-14,000 Ib) vehicles, and are typically full-
size pickup trucks and work vans. The key elements of these standards include a combined
NMOG+NOx declining fleet average standard, new stringent PM standards phasing in on a
separate schedule, extension of the regulatory useful life to 150,000 miles, and a new
requirement to meet standards over the SFTP that will address real-world driving modes not
well-represented by the FTP cycle alone. Table ES-4 presents the HDV fleet average
NMOG+NOx standard, which becomes more stringent in successive model years from 2018 to
2022, with voluntary standards available in 2016 and 2017.
The PM standards are 8 mg/mi for Class 2b vehicles and 10 mg/mi for Class 3 vehicles,
to be phased in on a percent-of-sales basis at 20-40-70-100 percent in 2018-2019-2020-2021,
respectively.
Table ES-4 HDV Fleet Average NMOG+NOx Standards (mg/mi)
Model Year
Class 2b
Class 3
Voluntary
2016
333
548
2017
310
508
Required Program
2018
278
451
2019
253
400
2020
228
349
2021
203
298
2022 and later
178
247
The new SFTP requirements for HDVs include NMOG+NOx, carbon monoxide (CO)
and PM standards. Compliance will be evaluated from a weighted composite of measured
emissions from testing over the FTP cycle, the SC03 cycle, and an aggressive driving cycle, with
the latter tailored to various HDV sub-categories: the US06 cycle for most HDVs, the highway
ES-3
-------
portion of the US06 cycle for low power-to-weight Class 2b HDVs, and the LA-92 cycle for
Class 3 HDVs.
Evaporative Emission Standards
To control evaporative emissions, more stringent standards will require covered vehicles
to have essentially zero fuel vapor emissions in use, including more stringent evaporative
emissions standards, new test procedures, and a new fuel/evaporative system leak standard. Tier
3 also includes refueling emission standards for complete heavy-duty gasoline vehicles
(HDGVs) over 10,000 Ibs GVWR. There are phase-in flexibilities as well as credit and
allowance programs. The standards, harmonized with California's zero evaporative emissions
standards, are designed to essentially eliminate fuel vapor-related evaporative emissions. The
Tier 3 evaporative emission standards will be phased in over a period of six MYs 2017-2022 as
shown in Table ES-5.
Table ES-5 Default Phase-in Schedule for Tier 3
Evaporative Emission Standards
Model year
2017
2018
2019
2020
2021
2022
Minimum
percentage of
vehicles subject to
the Tier 3 standards
40%M'J
60%
60%
80%
80%
100%
The phase-in percentage for model year 2017 applies
only for vehicles at or below 6,000 pounds GVWR.
2 The leak standard does not apply for model year 2017.
3 There are three options for the 2017 MY, only one is shown here.
Table ES-6 presents the evaporative hot soak plus diurnal emission standards by vehicle
class. Manufacturers may comply on average within each of the four vehicle categories but not
across these categories. Tier 3 also includes separate high altitude emission standards for these
vehicle categories.
Table ES-6 Evaporative Emission Standards (g/test)
Vehicle Category
LDV, LDT1
LDT2
LDT3, LDT4, MDPV
HDGVs
Highest Diurnal + Hot Soak Level
(over both 2-day and 3 -day diurnal tests)
0.300
0.400
0.500
0.600
ES-4
-------
There is a new testing requirement referred to as the bleed emission test procedure.
Manufacturers will be required to measure diurnal emissions over the 2-day diurnal test
procedure from just the fuel tank and the evaporative emission canister and comply with a 0.020
g/test standard for all LDVs, LDTs, and MDPVs without averaging. The canister bleed emission
standard test will apply only for low altitude testing conditions, but there is proportional control
at higher altitudes.
EPA is including these Tier 3 evaporative emission controls for HDGVs as part of the
overall scheme for LDVs and LDTs. The individual vehicle emission standard will be 0.600
g/test for both the 2-day and 3-day evaporative emission tests, the high altitude standard will be
1.75 g/test and the canister bleed test standard will be 0.030 g/test.
We are adding a new standard and test procedure related to controlling vapor leaks from
vehicle fuel and vapor control systems. The standard, which will apply to all LDVs, LDTs,
MDPVs, and Class 2b/3 HDGVs, will prohibit leaks larger than 0.02 inches of cumulative
equivalent diameter in the fuel/evaporative system. The Tier 3 evaporative emission standards
program requirements will be phased in over a period of six model years between MYs 2017 and
2022, with the leak test phasing in beginning in 2018 MY as a vehicle is certified to meet Tier 3
evaporative emission requirements.
There are new refueling emission control requirements for complete HDGVs equal to or
less than 14,000 Ibs GVWR (i.e., Class 2b/3 HDGVs), that start in the 2018 model year. For
complete HDGVs > 14,000 Ibs GVWR the refueling emission control requirement start in the
2022 model year.
We are adopting and incorporating by reference the current CARB onboard diagnostic
system (OBD) regulations effective for the 2017 MY plus two minor provisions to enable OBD-
based leak detection to be used in IUVP testing. EPA will retain the provision that certifying
with CARB's program will permit manufacturers to seek a separate EPA certificate on that basis.
Emissions Test Fuel Requirements
There are several changes to our federal gasoline emissions test fuel. Key changes
include:
• Moving away from "indolene" (EO) to a test fuel containing 10 percent ethanol by
volume (E10);
• Lowering octane to match regular-grade gasoline (except for premium-required
vehicles);
• Adjusting distillation temperatures, aromatics and olefms to better match today's in-
use fuel and to be consistent with anticipated E10 composition; and
• Lowering the existing sulfur specification and setting a benzene specification to be
consistent with proposed Tier 3 gasoline sulfur requirements and recent MSAT2
gasoline benzene requirements.
ES-5
-------
• E85 and premium test fuel specifications.
Gasoline Sulfur Standards
Under the Tier 3 fuel program, federal gasoline will contain no more than 10 parts per
million (ppm) sulfur on an annual average basis beginning January 1, 2017. There will be an
averaging, banking, and trading (ABT) program that would allow refiners and importers to
spread out their investments through an early credit program and rely on ongoing nationwide
averaging to meet the 10 ppm sulfur standard. There will be a three-year delay for small refiners
and "small volume refineries" (refiners processing less than or equal to 75,000 barrels per
calendar day). In addition, we are maintaining the current refinery gate and downstream sulfur
caps of 80 ppm and 95 ppm, respectively.
Projected Impacts
Changes to Analyses Since Proposal
Since the proposal, we have made several updates to the analyses that estimate the
projected impacts of the Tier 3 standards. We made several changes to our baseline (also
referred to as the "reference case"), which is our projection of future conditions if the Tier 3
standards were not finalized. Specifically, our baseline now accounts for the fact that California
and twelve additional states have adopted California's Low Emission Vehicle III (LEV III)
program. This change reduces the emissions and air quality impacts of the Tier 3 standards (and
thus the monetized benefits), and it also reduces the cost of the Tier 3 vehicle standards. In
addition, the baseline now accounts for the light-duty greenhouse gas emissions standards for
2017 and later model years, and the greenhouse gas emissions standards for medium- and heavy-
duty engines and vehicles. This update affects the per-vehicle technology costs but has little
impact on the emissions and air quality benefits of the Tier 3 program, because it is included in
both the baseline and control cases. Finally, the baseline now uses the U.S. Energy Information
Administration's Annual Energy Outlook 2013 (AEO2013) as the source for future renewable
fuel volumes and blends and future gasoline consumption. AEO2013 projects significantly lower
gasoline consumption than AEO2011 (which was used in the proposal's analysis), and this
reduces the total cost of the Tier 3 fuel program. There are a number of other updates to our
cost, emissions, air quality, and benefits analyses, as detailed in the RIA. Among the most
significant are the changes to the vehicle and fuel cost estimates, which have resulted in costs
that are lower than projected in the proposal. The updates with the most significant impacts on
the per-vehicle costs include a more robust estimate of catalyst loading costs and the new
baseline fleet that reflects implementation of the most recent greenhouse gas emissions
standards. Both of these updates reduced per-vehicle costs. Total vehicle program costs were
also significantly reduced because costs are no longer incurred for vehicles sold in states that
have adopted the California LEV III program. With respect to fuel costs, the change with the
most significant impact on per-gallon costs is the inclusion of nationwide credit trading (i.e.,
between companies). The proposal's primary cost analysis was based only on trading within
companies (although we also presented in the proposal the cost if trading between firms
occurred). The reduction in per-gallon costs, when combined with significantly lower
ES-6
-------
projections of gasoline consumption from AEO2013, resulted in lower fuel program costs than
the proposal had estimated.
Emission and Air Quality Impacts
The Tier 3 vehicle and fuel-related standards together will reduce emissions of NOx,
VOC, PM2.s, and air toxics. The gasoline sulfur standards, which will take effect in 2017, will
provide large immediate reductions in emissions from existing gasoline vehicles and engines.
The emission reductions will increase over time as newer vehicles become a larger percentage of
the fleet (e.g., in 2030, 70 percent of the miles travelled are from vehicles that meet the fully
phased-in Tier 3 standards). Projected emission reductions from the Tier 3 standards for 2018
and 2030 are shown in Table ES-7. We expect these reductions to continue beyond 2030 as more
of the fleet continues to turn over to Tier 3 vehicles.
Table ES-7 Estimated Emission Reductions from the Final Tier 3 Standards
(Annual U.S. short tons)
NOX
VOC
CO
Direct PM2.5
Benzene
S02
1,3-Butadiene
Formaldehyde
Acetaldehyde
Acrolein
Ethanol
2018
Tons
264,369
47,504
278,879
130
1,916
14,813
257
513
600
40
2,704
Percent of Onroad
Inventory
10%
3%
2%
0.1%
6%
56%
5%
2%
3%
3%
2%
2030
Tons
328,509
167,591
3,458,041
7,892
4,762
12,399
677
1,277
2,067
127
19,950
Percent of Onroad
Inventory
25%
16%
24%
10%
26%
56%
29%
10%
21%
15%
16%
We project that the Tier 3 vehicle and fuel standards will reduce nitrous oxide (N2O) and
methane (CH/i) emissions from vehicles. The reductions in these potent greenhouse gases will be
partially offset by the increase in CO2 emissions from refineries. The combined impact is a net
decrease on a CO2-equivalent basis (2.5 to 2.7 million metric tons of CO2-equivalent reduced in
2030).
Reductions in emissions of NOx, VOC, PM2.5 and air toxics are projected to lead to
nationwide decreases in ambient concentrations of ozone, PM2.s, NO2, CO, and air toxics.
Specifically, the Tier 3 standards will significantly decrease ozone concentrations across the
country, with an estimated population-weighted average decrease of 0.49 ppb in 2018 and 0.98
ppb in 2030. Few other strategies exist that would deliver the reductions needed for states to
meet the current ozone standards. The Tier 3 standards will decrease ambient annual PM2.s
concentrations across the county as well, with an estimated population-weighted average
decrease of 0.04 |ig/m3 by 2030. Decreases in ambient concentrations of air toxics are also
ES-7
-------
projected with the Tier 3 standards, including notable nationwide reductions in benzene
concentrations.
Costs and Benefits
The costs that will be incurred from our final program fall into two categories - costs
from the Tier 3 vehicle exhaust and evaporative standards and from reductions in sulfur content
of gasoline. All costs represent the fleet-weighted average of light-duty vehicles and trucks. All
costs are represented in 2011 dollars.
Vehicle Costs
The vehicle costs include the technology costs projected to meet the exhaust and
evaporative standards, as show in Table ES-8. The fleet mix of light-duty vehicles, light duty
trucks, and medium-duty trucks reflects the MY 2017-2025 light-duty and MY2014-2018 heavy-
duty GHG final rulemakings.
Table ES-8 Annual Vehicle Technology Costs, 2011$
Year
2017
2030
Vehicle Exhaust
Emission Control
Costs
(SMillion)
$268
$664
Vehicle Evaporative
Emission Control
Costs
($Million)
$26
$113
Operating Costs
($Million)
$0
-$19
Facility Costs
($Million)
$4
$4
Total Vehicle
Costs
($Million)a
$297
$761
These estimates include costs associated with the Tier 3 vehicle standards in all states except California and
states that have adopted the LEV III program.
Fuel Costs
The fuel costs consist of the additional operating costs and capital costs to the refiners to
meet the sulfur average of 10 ppm. The sulfur control costs assume a cost of 0.65 cents per
gallon which includes the refinery operating and capital costs. The annual fuel costs of the
program are listed in Table ES-9.
Table ES-9 Annual Fuel Costs, 2011$
Year
2017
2030
Fuel Sulfur Control Costs ($Million)a
$804
$696
These estimates include costs associated with the Tier 3 fuel
standards in all states except California.
Total Costs
The sum of the vehicle technology costs to control exhaust and evaporative emissions, in
addition to the costs to control the sulfur level in the fuel, represent the total costs of the
ES-8
-------
program, as shown in Table ES-10. The final fuel standards are projected to lead to an average
cost of 0.65 cents per gallon of gasoline, and the vehicle standards would have an average
technology cost of $72 per vehicle
Table ES-10: Total Annual Vehicle and Fuel Control Costs, 2011$
Year
2017
2030
Total Vehicle and Fuel Control Costs
($Million)a
$1,101
$1,457
These estimates include costs associated with both the Tier 3 vehicle
standards in all states except California and states that have adopted the
LEV III program, and the Tier 3 fuel standards in all states except
California.
Benefits
Exposure to ambient concentrations of ozone, PM2.5, and air toxics is linked to adverse
human health impacts such as premature deaths as well as other important public health and
environmental effects. The final Tier 3 standards are expected to reduce these adverse impacts
and yield significant benefits, including those we can monetize and those we are unable to
quantify.
The range of quantified and monetized benefits associated with this program are
estimated based on the risk of several sources of PM- and ozone-related mortality effect
estimates, along with other PM and ozone non-mortality related benefits information. Overall,
we estimate that the final rule will lead to a net decrease in PM2.5- and ozone-related health and
environmental impacts. The estimated range of total monetized ozone- and PM-related health
impacts is presented in Table ES-11.
Table ES-11: Estimated 2030 Monetized PM-and Ozone-Related Health Benefits
(Billions, 201 l$)a
Description
Total Estimated Health Benefits '°' 'e
3 percent discount rate
7 percent discount rate
2030
$7.4 -$19
$6.7 -$18
Notes:
a Totals are rounded to two significant digits and may not sum due to rounding.
b Total includes ozone and PM2 5 estimated benefits. Range was developed by adding the estimate from the
Bell et al., 2004 ozone premature mortality function to PM2 5-related premature mortality derived from the
American Cancer Society cohort study (Krewski et al., 2009) for the low estimate and ozone premature
mortality derived from the Levy et al., 2005 study to PM2 5-related premature mortality derived from the
Six-Cities (Lepeule et al., 2012) study for the high estimate.
0 Annual benefits analysis results reflect the use of a 3 percent and 7 percent discount rate in the valuation
of premature mortality and nonfatal myocardial infarctions, consistent with EPA and OMB guidelines for
preparing economic analyses.
ES-9
-------
d Valuation of premature mortality based on long-term PM exposure assumes discounting over the SAB
recommended 20-year segmented lag structure described in the Regulatory Impact Analysis for the 2006
PM National Ambient Air Quality Standards (September, 2006).
e Not all possible benefits are quantified and monetized in this analysis; the total monetized benefits
presented here may therefore be underestimated.
We estimate that by 2030, the annual emission reductions of the Tier 3 standards will
annually prevent between 660 and 1,500 PM-related premature deaths, between 110 and 500
ozone-related premature deaths, 81,000 work days lost, and approximately 1.1 million minor
restricted-activity days. The estimated annual monetized health benefits of the proposed Tier 3
standards in 2030 (2011$) will be between $7.4 and $19 billion, assuming a 3-percent discount
rate (or between $6.7 billion and $18 billion assuming a 7-percent discount rate).
Comparison of Costs and Benefits
Using a conservative benefits estimate, the 2030 benefits outweigh the costs by a factor
of 4.5. Using the upper end of the benefits range, the benefits could outweigh the costs by a
factor of 13. Thus, even taking the most conservative benefits assumptions, benefits of the final
standards are projected to outweigh the costs. The results are shown in Table ES-12.
Table ES-12 Summary of Annual Benefits and Cost Associated with the Final Tier 3
Program (Billions, 2011$)a
Description
Vehicle Program Costs
Fuels Program Costs
Total Estimated Costs
Total Estimated Health Benefitsc'd'e'f
3 percent discount rate
7 percent discount rate
Annual Net Benefits (Total Benefits - Total Costs)
3 percent discount rate
7 percent discount rate
2030
$0.76
$0.70
$1.5
$7.4 -$19
$6.7 -$18
$5.9 -$18
$5.2 -$17
Notes:
a All estimates represent annual benefits and costs anticipated for the year 2030. Totals are rounded to two
significant digits and may not sum due to rounding.
b The calculation of annual costs does not require amortization of costs over time. Therefore, the estimates of annual
cost do not include a discount rate or rate of return assumption (see Chapter 2 of the RIA for more information on
vehicle costs, Chapter 5 for fuel costs, and Section 8.1.1 for a summary of total program costs).
0 Total includes ozone and PM2 5 estimated benefits. Range was developed by adding the estimate from the Bell et
al., 2004 ozone premature mortality function to PM2 5-related premature mortality derived from the American
Cancer Society cohort study (Krewski et al., 2009) for the low estimate and ozone premature mortality derived from
the Levy et al., 2005 study to PM25-related premature mortality derived from the Six-Cities (Lepeule et al., 2012)
study for the high estimate.
d Annual benefits analysis results reflect the use of a 3 percent and 7 percent discount rate in the valuation of
premature mortality and nonfatal myocardial infarctions, consistent with EPA and OMB guidelines for preparing
economic analyses.
ES-10
-------
e Valuation of premature mortality based on long-term PM exposure assumes discounting over the SAB
recommended 20-year segmented lag structure described in the Regulatory Impact Analysis for the 2012 PM
National Ambient Air Quality Standards (December, 2012).
^Not all possible benefits or disbenefits are quantified and monetized in this analysis. Potential benefit categories
that have not been quantified and monetized are listed in Table 8-5.
Economic Impact Analysis
The rule will affect two sectors directly: vehicle manufacturing and petroleum refining.
The estimated increase in vehicle production cost because of the rule is expected to be small
relative to the costs of the vehicle. Some or all of this production cost increase will be expected
to be passed through to consumers. This increase in price is expected to lower the quantity of
vehicles sold, though because the expected cost increase is small, we expect the decrease in sales
to be negligible. This decrease in vehicle sales is expected to decrease employment in the
vehicle manufacturing sector. However, costs related to compliance with the rule should also
increase employment in this sector. While it is unclear which of these effects will be larger,
because the increase in vehicle production costs and the decrease in vehicle sales are minor, the
impact of the rule on employment in the vehicle manufacturing sector is expected to be small as
well. The key change for refiners from the proposed standards will be more stringent sulfur
requirements. Analogous to vehicle sales, this change to fuels is expected to increase
manufacturers' costs of fuel production. Some or all of this increase in production costs is
expected to be passed through to consumers which should lead to a decrease in fuel sales. As
with the vehicle manufacturing sector, we expect the decrease in fuel sales to negatively affect
employment in this sector, while the costs of compliance with the rule will be expected to
increase employment. It is not evident whether the rule will increase or decrease employment in
the refining sector as a whole. However, given the small anticipated increase in production costs
of less than one cent per gallon and the small likely decrease in fuel sales, we expect that the rule
will not have major employment consequences for this sector.
ES-11
-------
Chapter 1 Vehicle Program Technological Feasibility
1.1 Introduction
For the vehicles and emissions addressed in this final rule, EPA has comprehensively
assessed the technological phenomena related to the generation of emissions of interest, the
nature of the technological challenges facing manufacturers to produce emission reductions of
the scale described in the Preamble, and the technologies that we expect to be available to
manufacturers to meet those challenges during the rule implementation time frame. Our
feasibility assessment recognizes that the Tier 3 program is composed of several new
requirements for all types of new vehicles, including a range of vehicle classes from small cars to
large pick-up trucks and MDPVs, and even heavy-duty vehicles with diverse applications and
specific engine designs matched to the needs of the users. This assessment also recognizes the
critical role of gasoline sulfur content in making it possible for us to adopt emission standards at
these very stringent levels, particularly for certain vehicle types. We provide below a full
assessment of our current knowledge of the effects of gasoline sulfur on current vehicle
emissions as well as our projections of how sulfur can be expected to affect compliance with the
Tier 3 standards.
Since there are multiple aspects to the Tier 3 program, it is necessary to consider
technical feasibility in light of the different program requirements and their interactions with
each other. For example, the technical feasibility of the finalized Tier 3 FTP NMOG+NOx and
the PM standards is directly related to the specifications of the fuel, including fuel sulfur, RVP
and ethanol content. Additionally, as mentioned above, the feasibility assessment must consider
that different technologies may be needed on different types of vehicle applications (i.e., cars
versus trucks) and must consider the effectiveness of these technologies to reduce emissions for
the full useful life of the vehicle while operating on in-use fuels. Certain smaller vehicles with
correspondingly small engines may be less challenged to meet FTP standards than larger
vehicles with larger engines. Conversely, these smaller vehicles may have more difficulty
meeting the SFTP requirements than the larger and more powerful vehicles. Additionally, the
ability to meet the SFTP emission requirements can also be impacted by the path taken to meet
the FTP requirements (i.e., larger volume catalysts for US06 emissions control vs. smaller
catalysts for improved FTP cold-start emissions control).
The rule also contains revised evaporative emission standards to be met on 9RVPE10 test
fuel for LDVs, LDTs, MDPVs, and HDGVs, as well as leak standard for all gasoline-powered
LDVs, LDTs, MDPVs, and HDGVs rated at or below 14,000 Ibs GVWR and new OBD
requirements for LDVs, LDTs, MDPVs, and HDVs rated at or below 14,000 Ibs GVWR. The
feasibility of these standards is discussed below as well.
1.2 FTP NMOG+NOx Feasibility
The new emission requirements include stringent NMOG+ NOx standards over the FTP
that would require new vehicle hardware and additional control of gasoline sulfur levels in order
to achieve the 30 mg/mi fleet average level in 2025. The type of new hardware that would be
required would vary depending on the specific application and emission challenges. Smaller
1-1
-------
vehicles with corresponding smaller engines would generally need less new hardware while
larger vehicles and other vehicles with larger engines may need additional hardware and
improvements beyond what would be needed for the smaller vehicles with smaller engines.
Additionally, the fleet-average nature of the standards allows more challenged vehicles to be
offset by vehicles that could outperform the required fleet averages.
In order to assess the technical feasibility of a 30 mg/mi NMOG+NOx national fleet
average FTP standard and a 3 mg/mi PM vehicle standard, EPA conducted two primary analyses.
The initial analyses performed were of the current Tier 2 and LEV II fleets. This provided a
baseline for the current federal fleet emissions performance, as well as the emissions
performance of the California LEV II fleet. The second consideration was a modal analysis of
typical vehicle emissions under certain operating conditions. In this way EPA determined the
specific emissions performance challenges that vehicle manufacturers would face in meeting the
lower fleet average emission standards. Each of these considerations is described in greater
detail below.
1.2.1 Assessment of the Current Federal Fleet Emissions
The current federal fleet is only required to be certified to an average of Tier 2 Bin 5,
equivalent to 160 mg/mi NMOG+NOx A For example, in MY 2009 92 percent of passenger cars
and LDTls were certified to Tier 2 Bin 5 and 91 percent of LDT2s through LDT4s were certified
to Tier 2 Bin 5. This was not an unexpected result as there was no motivation for vehicle
manufacturers to produce a federal fleet that over-complied with respect to the Tier 2 standards.
By comparison, in the MY 2009 California fleet, where compliance with the LEVII declining
NMOG requirement and the "PZEV" program encouraged manufacturers to certify to cleaner
levels, only 30 percent of the passenger cars and LDTls were certified to Tier 2 Bin 5 and 60
percent were certified to Tier 2 Bin 3. The situation regarding the truck fleet in California was
similarly stratified, with 37 percent of the LDT2s through LDT4s being certified to Tier 2 Bin 5
and 55 percent being certified to Tier 2 Bin 3. In many cases, vehicles were being certified to a
lower standard in California and a higher standard federally. In the proposal, EPA stated a belief
that the patterns described above indicated that much of the Tier 2 fleet could be certified to a
lower federal fleet average immediately, with no major feasibility concerns.6
For the final rule, we have looked at MY2013 certification data. The MY 2013 data
indicate that more engine families are being certified to cleaner Tier 2 bins than what was
observed in previous MYs. In fact, in MY2013, while only 68% of passenger car and LDT1
families and 65% of LDT2s through LDT4s are certified to Tier 2 Bin 5, 31% of passenger car
and LDT1 and 29% of LDT2 through LDT4 are now certified to Tier 2 Bins 2 through 4. This
supports our stated belief that the Tier 2 fleet could be certified to a lower federal fleet average
without feasibility concerns. Table 1-1 shows that 58 MY2013 engine families have certified
A The current Tier 2 program does not combine NMOG and NOX emissions into one fleet-average standard. The
fleet-average standard in that program is for NOX emissions alone.
B Compliance with full useful life standards in California occurs at much lower in-use gasoline sulfur levels than is
the case with federally certified vehicles. For further discussion of the impact of gasoline fuel sulfur on light-vehicle
emissions feasibility and in-use compliance, please refer to Section 1.2.4.
1-2
-------
emission levels well below the Tier 2 bin to which they have been certified. The table also
shows that these engine families have emissions at or below the Tier 3 30 mg/mi NMOG+NOx
level.
Table 1-1 MY2013 Certified Engine Families with Certified Emission Levels at or Below
the Tier 3 NMOG+NOx 30 mg/mi Standard
Mfr
Audi
BMW
BMW
BMW
BMW
BMW
Chrysler
Chrysler
Chrysler
Ford
Ford
Ford
Ford
Ford
GM
GM
GM
GM
Honda
Honda
Honda
Jaguar Cars
Mazda
Mercedes-Benz
Mercedes-Benz
Mercedes-Benz
Nissan
Nissan
Nissan
Subaru
Subaru
Suzuki
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Model
A3
John Cooper Works A114
Countryman
328i xDrive
328 Ci Convertible
ActiveHybrid 7
ActiveHybrid 7
Patriot 4wd
Patriot 2wd
Dart
Fusion FWD
FOCUS
MKZ (HEV)
MKZ (HEV)
C-Max (PHEV)
REGAL
MALffiU
MALffiU
XTS
INSIGHT EX
ILX HYBRID
CIVIC HYBRID
XJ3.0
Mazda 3 DI 5 -Door
GLK 350 4MATIC
S 400 HYBRID
E 400 HYBRID
NISSAN TITAN 4WD
Crew Cab XE LWB
NISSAN VERSA S
NISSAN ALTIMA2.5SL
OUTBACK WAGON
AWD
OUTBACK WAGON
AWD
GRAND VITARA 4WD
TACOMA 2WD
RX450hAWD
RX450hAWD
FJ CRUISER 4WD
LX570
iQ
NMOG
(g/mi)
0.005
0.017
0.010
0.010
0.012
0.012
0.003
0.003
0.003
0.006
0.004
0.011
0.011
0.008
0.020
0.005
0.005
0.008
0.009
0.011
0.012
0.008
0.008
0.008
0.005
0.007
0.024
0.018
0.010
0.005
0.005
0.024
0.012
0.008
0.007
0.019
0.019
0.017
NOx
(g/mi)
0.011
0.010
0.010
0.010
0.010
0.010
0.020
0.010
0.010
0.010
0.010
0.000
0.000
0.010
0.000
0.010
0.010
0.020
0.010
0.010
0.010
0.020
0.010
0.010
0.010
0.010
0.000
0.000
0.000
0.020
0.020
0.005
0.010
0.000
0.000
0.010
0.010
0.010
Certified
NMOG
+NOx
(g/mi)
0.016
0.027
0.020
0.020
0.022
0.022
0.023
0.013
0.013
0.016
0.014
0.011
0.011
0.018
0.020
0.015
0.015
0.028
0.019
0.021
0.022
0.028
0.018
0.018
0.015
0.017
0.024
0.018
0.010
0.025
0.025
0.029
0.022
0.008
0.007
0.029
0.029
0.027
NMOG
+NOx of
Tier 2 bin
(g/mi)
0.085
0.160
0.160
0.160
0.160
0.160
0.160
0.160
0.110
0.085
0.085
0.085
0.085
0.085
0.110
0.110
0.110
0.110
0.085
0.085
0.085
0.160
0.160
0.110
0.110
0.110
0.160
0.160
0.160
0.110
0.110
0.160
0.160
0.085
0.085
0.160
0.160
0.160
Certified
emissions
at or below
Tier 3 bin
20
30
30
30
30
30
30
20
20
20
20
20
20
20
30
20
20
30
20
30
30
30
20
20
20
20
30
20
20
30
30
30
30
20
20
30
30
30
1-3
-------
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Volkswagen
Volkswagen
Volkswagen
Volkswagen
Volkswagen
Volkswagen
Volkswagen
Volkswagen
YARIS
PRIUSc
PRIUSv
PRIUS
PRIUS Plug-in Hybrid
ES 300h
ES 300h
Camry Hybrid XLE
CAMRY
GS 450h
LS 460 L AWD
LS600hL
Jetta Hybrid
Eos
Eos
Eos
Jetta
Jetta
Jetta
Jetta
0.018
0.006
0.004
0.005
0.005
0.008
0.007
0.006
0.006
0.006
0.018
0.006
0.008
0.004
0.004
0.004
0.010
0.010
0.010
0.010
0.010
0.000
0.000
0.000
0.000
0.000
0.000
0.010
0.020
0.010
0.010
0.000
0.007
0.008
0.008
0.008
0.002
0.002
0.002
0.002
0.028
0.006
0.004
0.005
0.005
0.008
0.007
0.016
0.026
0.016
0.028
0.006
0.015
0.012
0.012
0.012
0.012
0.012
0.012
0.012
0.160
0.085
0.085
0.085
0.085
0.085
0.085
0.085
0.160
0.085
0.160
0.085
0.085
0.085
0.085
0.085
0.085
0.085
0.085
0.085
30
20
20
20
20
20
20
20
30
20
30
20
20
20
20
20
20
20
20
20
Looking at the MY2013 certification data more closely and in the context of Tier 3
compliance, we find that 40 engine families are certified to emission levels low enough to
provide sufficient compliance margin (30% compliance margin, or 70% of the standard) to
enable Tier 3 compliance today.0 Table 1-2 shows these 40 engine families, their certified
emission levels, and the Tier 3 bin into which they could be certified while providing 30%
compliance margin.
Table 1-2 MY2013 Certified Engine Families with Certified Emission Levels that Could be
Certified to Tier 3 Today*
Mfr
Audi
BMW
BMW
Chrysler
Chrysler
Ford
Ford
Ford
Ford
Model
A3
328i xDrive
328 Ci Convertible
Patriot 2wd
Dart
Fusion FWD
FOCUS
MKZ (HEV)
MKZ (HEV)
NMOG
(g/mi)
0.005
0.010
0.010
0.003
0.003
0.006
0.004
0.011
0.011
NOx
(g/mi)
0.011
0.010
0.010
0.010
0.010
0.010
0.010
0.000
0.000
Certified
NMOG +NOx
(g/mi)
0.016
0.020
0.020
0.013
0.013
0.016
0.014
0.011
0.011
Possible Tier
3 bin*
30
30
30
20
20
30
20
20
20
c We believe that manufacturers will target compliance margins of 20-40% under Tier 3, as discussed in section
1.2.3 of this chapter. Here we have used 30% as it represents the midpoint of that expected range.
1-4
-------
Ford
GM
GM
GM
Honda
Honda
Mazda
Mercedes-Benz
Mercedes-Benz
Mercedes-Benz
Nissan
Nissan
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Toyota
Volkswagen
Volkswagen
Volkswagen
Volkswagen
Volkswagen
Volkswagen
Volkswagen
Volkswagen
C-Max (PHEV)
REGAL
MALffiU
MALffiU
INSIGHT EX
ILX HYBRID
Mazda 3 DI 5 -Door
GLK 350 4MATIC
S 400 HYBRID
E 400 HYBRID
NISSAN VERSA S
NISSAN ALTIMA2.5SL
RX450hAWD
RX450hAWD
PRIUSc
PRIUSv
PRIUS
PRIUS Plug-in Hybrid
ES 300h
ES 300h
Camry Hybrid XLE
GS 450h
LS600hL
Jetta Hybrid
Eos
Eos
Eos
Jetta
Jetta
Jetta
Jetta
0.008
0.020
0.005
0.005
0.009
0.011
0.008
0.008
0.005
0.007
0.018
0.010
0.008
0.007
0.006
0.004
0.005
0.005
0.008
0.007
0.006
0.006
0.006
0.008
0.004
0.004
0.004
0.010
0.010
0.010
0.010
0.010
0.000
0.010
0.010
0.010
0.010
0.010
0.010
0.010
0.010
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.010
0.010
0.000
0.007
0.008
0.008
0.008
0.002
0.002
0.002
0.002
0.018
0.020
0.015
0.015
0.019
0.021
0.018
0.018
0.015
0.017
0.018
0.010
0.008
0.007
0.006
0.004
0.005
0.005
0.008
0.007
0.016
0.016
0.006
0.015
0.012
0.012
0.012
0.012
0.012
0.012
0.012
30
30
30
30
30
30
30
30
30
30
30
20
20
20
20
20
20
20
20
20
30
30
20
30
20
20
20
20
20
20
20
including at least a 20% compliance margin (i.e., emissions 70% of the standard).
To support the FTP emission levels finalized for heavy duty vehicles, we analyzed the
certification emission results from the 2010 through 2013 MY vehicles0. The new Tier 3 fleet
average NMOG+NOx standard in 2022 for Class 2b vehicles is 178 mg/mi while the level for
Class 3 vehicles is 247 mg/mi. Shown in Table 1-3 below are the emission levels of 2010 and
2011 MY heavy-duty vehicle models operating on various fuels. It is important to note that while
we are finalizing a useful life of 150,000 miles, the current heavy duty vehicle requirements and
therefore the reported emission results represent the 120,000 miles deteriorated results either
calculated using deterioration factors applied to the 4,000 mile test or actual aged vehicles and
components. It will be important for manufacturers to carefully manage emissions deterioration
throughout the useful life of the vehicle to meet useful life emission requirements, consistent
with the challenge for light-duty applications.
D Manufacturers will regularly carry-over data for several model years. Where available, the latest reported
certification data was used for this analysis.
1-5
-------
Table 1-3: 2010 thru 2013 MY Heavy Duty Vehicle FTP 120k Certification Results
Class 2b
b
Class 3
b
Manufacturer
Altech
Chrysler
Chrysler
Chrysler
Ford
General
Motors
Mercedes-
Benz
Nissan
Nissan
Baytech
Chrysler
Chrysler
Ford
Ford
Ford
Ford
Ford
General
Motors
General
Motors
Mercedes-
Benz
Models
F250
Ram 2500
Ram 2500
Ram 2500
F250
Silverado
2500
Sprinter
NV3500
4.0L
NV3500
5.6L
Silverado
3500
Ram 3500
Ram 3500
E350C
E350C
F350
F350
F350
Silverado
3500
Silverado
3500
Sprinter
Fuel
Type
CNG
Gasoline
Diesel
CNG
Diesel
Diesel
Diesel
Gasoline
Gasoline
CNG
Gasoline
Diesel
Gasoline
E85
Diesel
Gasoline
E85
Gasoline
Diesel
Diesel
NMOG
Level a
(mg/mi)
10
118
63
24
104
79
4
57
66
11
133
52
51
70
89
79
76
131
54
11
NOX
Level
(mg/mi)
100
100
200
100
200
200
100
0
100
100
200
400
82
65
300
130
83
150
200
100
NMOG
+NOX
(mg/mi)
110
218
263
124
304
279
104
57
166
111
333
452
133
135
389
209
159
281
254
111
CO
(g/mi)
5.9
1.6
.2
.8
.9
.7
.1
1.4
1.3
1.3
2.6
.2
2
1.1
.9
3.2
1.8
3.4
.5
.2
PM
(mg/
mi)
-
-
0
-
10
1
10
-
-
-
3
-
-
20
-
-
-
0
0
Notes:
" Diesel reported as NMHC
* Gasoline Class 2b models from General Motors and Ford certified using worst case Class 3 data
c Tested at LVW with MDPVs
1.2.2 NMOG and NOX Emissions on the FTP
To understand how several currently-used technologies described below could be used by
manufacturers to reach the stringent Tier 3 NMOG+NOx standards, it is helpful to consider
1-6
-------
emissions formation in common modes of operation for gasoline engines, or modal analysis.E
As previously stated during the discussion of the NMOG+NOx standard, many gasoline engines
produce very little NOx over the FTP. Thus, the challenge faced by manufacturers for producing
Tier 3 compliant gasoline powertrains is to reduce the NMOG portion of the emissions. Based
on modal analysis of a gasoline powered vehicle being operated on the FTP cycle, approximately
90 percent of the NMOG emissions occur during the first 50 seconds after a cold start. In
addition, about 60 percent of the NOx emissions occur in this same period. The remainder of the
emissions, particularly NOx emissions, are made during warmed up operation when the emission
controls rely primarily on very high conversion efficiency of the catalyst. This is possible when
catalyst performance, both on warm-up and during hot operation is not impeded by sulfur in the
fuel. Figure 1-1 below shows the second-by-second emissions for NMOG and NOx following a
cold start.
Thus, effective control of these cold-start emissions, especially NMOG emissions, would
be the primary technological goal of manufacturers complying with the Tier 3 FTP standards.
As discussed below, manufacturers are already applying several technologies capable of
significant reductions in these cold start emission to vehicles currently on the road.
Cumulative NOx + THC (Normalized)
Time (sec)
Figure 1-1 Modal Analysis of NMOG and NOx Emissions (LA92)
1.2.3 Compliance Margin
Vehicle manufacturers have historically designed vehicles to meet emissions targets
which are 50-70 percent of the emission standards after the catalytic converters have been
thermally aged and exposed to expected normal levels observed in-use of catalyst poisons (e.g.,
E A modal analysis provides a second-by-second view of the total amount of emissions over the entire cycle being
considered.
1-7
-------
sulfur from fuel, phosphorous from lubricating oil, etc.) out to the vehicle's full useful life. This
difference is referred to as "compliance margin" and is a result of manufacturers' efforts to
address all the sources of variability and emissions control system degradation that could occur
during the certification or in-use testing processes and during in-use operation. We believe that
manufacturers will continue require a compliance margin however the combined MOG+NOx
approach for Tier 3 will allow for some flexibility in the compliance margin targets. We expect
that compliance margins for the Tier 3 combined standard will range from 20% to 40%, because
the combined standard generally allows an increase in one emission constituent to be offset by a
decrease in the other. Thus, the emission design targets for Tier 3 standards at full useful life
would be approximately 18 to 24 mg/mi MOG+NOx for a Bin 30 certified vehicle at full useful
life. These sources of variability include:
• Test-to-test variability (within one test site and lab-to-lab)
• Build variation and manufacturing tolerances
• Vehicle operation (for example: driving habits, ambient temperature, etc.)
• Fuel composition
o The deleterious effects of fuel sulfur on exhaust catalysts and oxygen
sensors
o Other fuel composition impacts
• Oil consumption
o The impact of direct emission of lubricating oil on semi-volatile organic
PM emissions
o The impact of oil additives and other components (e.g., phosphorous) and
oil ash on exhaust catalysts and oxygen sensors
For MY 2013, the compliance margin for a Tier 2 Bin 5 vehicle averaged approximately
65 percent. In other words, actual vehicle emissions performance was on average about 35
percent of a 160 mg/mi NMOG+NOx standard, or about 56 mg/mi. By comparison, for MY
2013 California-certified vehicles, the average SULEV compliance margin was somewhat less
for the more stringent standards, approximately 50 percent. We believe that the recent California
experience is a likely indicator of the smaller compliance margins that manufacturers would
design for in order to comply with the Tier 3 FTP standards. Thus, a typical Tier 2 Bin 5
vehicle, performing at 35 percent of the current standard (i.e., at about 56 mg/mi) would need
improvements sufficient to achieve the Tier 3 targets for the 30 mg/mi combined NMOG+NOx
standard.
1.2.4 Impact of Gasoline Sulfur Control on the Feasibility of the Vehicle Emission Standards
In this section, we discuss the impact of gasoline sulfur control on the feasibility of the
Tier 3 vehicle emissions standards and on the exhaust emissions of the existing in-use vehicle
fleet. Section 1.2.4.1 describes the chemistry and physics of the impacts of gasoline sulfur
compounds on exhaust catalysts. Sections 1.2.4.2, 1.2.43, and 1.2.4.4 summarize research on the
impacts of gasoline sulfur on vehicles utilizing various degrees of emission control technology,
with Section 1.2.4.2 summarizing historical studies on the impact of gasoline sulfur on vehicle
emissions, Section 1.2.4.3 describing impacts on Tier 2 vehicles and the existing light-duty
vehicle fleet, and Section 1.2.4.4 describing impacts on vehicles using technology consistent
1-8
-------
with what we expect to see in the future Tier 3 vehicle fleet. Section 1.2.4.5 provides EPA's
assessment of the level of gasoline sulfur control necessary for light-duty vehicles to comply
with Tier 3 exhaust emission standards.
EPA's primary findings are:
• Reducing gasoline sulfur content to a 10 ppm average will provide immediate and
significant exhaust emissions reductions to the current, in-use fleet of light-duty
vehicles.
• Reducing gasoline sulfur content to an average of 10 ppm will enable vehicle
manufacturers to certify their entire product lines of new light-duty vehicles to the
final Tier 3 Bin 30 fleet average standards. Without such sulfur control it would not
be possible for vehicle manufacturers to reduce emissions sufficiently below Tier 2
levels to meet the new Tier 3 standards because it would require offsetting
significantly higher exhaust emissions resulting from the higher sulfur levels. EPA
has not identified any existing or developing technologies that would compensate for
or offset the higher exhaust emissions resulting from higher fuel sulfur levels.
1.2.4.1 Gasoline Sulfur Impacts on Exhaust Catalysts
Modern three-way catalytic exhaust systems utilize platinum group metals (PGM), metal
oxides and other active materials to selectively oxidize organic compounds and carbon monoxide
in the exhaust gases. These systems simultaneously reduce NOx when air-to-fuel ratio control
operates in a condition of relatively low amplitude/high frequency oscillation about the
stoichiometric point. Sulfur is a well-known catalyst poison. There is a large body of work
demonstrating sulfur inhibition of the emissions control performance of PGM three-way exhaust
catalyst systems. 1'2!3'4'5'6'7'8'9'10 The nature of sulfur interactions with washcoat materials, active
catalytic materials and catalyst substrates is complex and varies with catalyst composition,
exhaust gas composition and exhaust temperature. The variation of these interactions with
exhaust gas composition and temperature means that the operational history of a vehicle is an
important factor; continuous light-load operation, throttle tip-in events and enrichment under
high-load conditions can all impact sulfur interactions with the catalyst.
Sulfur from gasoline is oxidized during spark-ignition engine combustion primarily to
862 and, to a much lesser extent, SCV2. Sulfur oxides selectively chemically bind (chemisorb)
with, and in some cases react with, active sites and coating materials within the catalyst, thus
inhibiting the intended catalytic reactions. Sulfur oxides inhibit pollutant catalysis chiefly by
selective poisoning of active PGM, ceria sites, and the alumina washcoating material (see Figure
1-2).11 The amount of sulfur retained by an exhaust catalyst system is primarily a function of the
concentration of sulfur oxides in the incoming exhaust gases, air-to-fuel ratio feedback and
control by the engine management system, the operating temperature of the catalyst and the
active materials and coatings used within the catalyst.
In their supplemental comments on the proposed Tier 3 rule, API criticized the use of
emissions data generated using gasoline with sulfur content outside of the range of 10 ppm to 30
1-9
-------
ppm within EPA and other analyses of the impacts of gasoline sulfur on exhaust emissions from
current in-use (Tier 2) and future (Tier 3) light-duty vehicles. Specific examples include:
• Comparisons of exhaust emissions at 5 ppm and 28 ppm gasoline sulfur levels
within the recent EPA study of emissions from Tier 2 vehicles12
• Comparison of exhaust emissions of a SULEV vehicle at 8 ppm and 33 ppm
gasoline sulfur levels within the Takei et al. study13
• Comparison of exhaust emissions of a PZEV vehicle at 3 ppm and 33 ppm
gasoline sulfur levels within the Ball et al. study.14
The relationship between changes in gasoline sulfur content and NOx, HC, NMHC and
NMOG emissions is typically linear. The linearity of sulfur impacts on NOx, NMHC and
NMOG emissions is supported by past studies with multiple fuel sulfur levels all of which
compare gasoline with differing sulfur levels that are below approximately 100 ppm (e.g. CRC
E-60 and 2001 AAM/AIAM programs as well as by comments on this rulemaking submitted by
MECA).15'16'17 An assumption of linearity of the effect of gasoline sulfur level on catalyst
efficiency between any two test fuels with differing sulfur levels is reasonable given that the
mass flow rate of sulfur in exhaust gas changes in proportion to its concentration in the fuel, and
that the chemistry of adsorption of sulfur on the active catalyst sites is an approximately-first-
order chemisorption until all active sites within a catalyst reach an equilibrium state relative to
further input of sulfur compounds. The relative linearity of the effect of gasoline sulfur level on
NMOG and NOx emissions allows exhaust emissions results generated within EPA and other
studies of gasoline sulfur at levels immediately above or below either 10 ppm or 30 ppm to be
normalized to either 10 ppm sulfur (Tier 3 gasoline) or to 30 ppm sulfur (Tier 2 gasoline, which
used in the analysis of the impacts of the Tier 3 gasoline standards on existing in-use vehicles
and future Tier 3 vehicles.
In their supplemental comments to the Tier 3 proposal, API also commented that EPA
did not show the sulfur impact on exhaust emissions at intermediate sulfur levels between 10
•I Q
ppm and 30 ppm. In response, based on the relative linearity of the effect of gasoline sulfur
level on NMOG and NOx emissions allowing exhaust emissions to be estimated for gasoline
sulfur levels between 10 and 30 ppm, data in EPA's analysis increased NMOG+NOx emissions
(as fuel sulfur increases) that becomes more severe (i.e., higher percentage increase in
NMOG+NOx emissions) for vehicles with extremely low19 exhaust emission (SULEV, PZEV,
LEVIII, Tier 3) as described in further detail in Sections 1.2.4.4 and 1.2.4.5.
1-10
-------
Active Catalytic Site
Catalytic Site Deactivated by Sulfur Poisoning
Alumina Washcoat
Monolith Substrate
\f- 100 A
$"::•:;::::':;::::::::::Ł—*
Figure 1-2 Functional schematic showing selective poisoning of active catalytic sites by
sulfur compounds. Adapted from Heck and Farrauto 2002.
20
Selective sulfur poisoning of platinum (Pt) and rhodium (Rh) is primarily from surface-
layer chemisorption. Sulfur poisoning of palladium (Pd) and ceria appears to be via
chemisorption combined with formation of more stable metallic sulfur compounds, e.g. PdS and
€62628, present in both surface and bulk form (i.e., below the surface layer).21'22'23'24 Ceria,
zirconia and other oxygen storage components (OSC) play an important role that is crucial to
NOx reduction over Rh as the engine air-to-fuel ratio oscillates about the stoichiometric closed-
9S
loop control point. Ceria sulfation interferes with OSC functionality within the catalyst and
thus can have a detrimental impact on the catalyst's ability to effectively reduce NOx emissions.
Water-gas-shift reactions are important for NOx reduction over catalysts combining Pd and ceria.
This reaction can be blocked by sulfur poisoning and may be responsible for observations of
reduced NOx activity over Pd/ceria catalysts even with exposure to fairly low levels of sulfur
(equivalent to 15 ppm in gasoline).26'2? Pd is also of increased importance for meeting Tier 3
standards due to its unique application in the close-coupled-catalyst location required for
vehicles certifying to very stringent emission standards. Close-coupling means that the exhaust
catalyst is moved as close as possible to the engine's exhaust ports within the packaging
constraints of an engine compartment. This ensures that the catalyst reaches its minimal
operational, or "light-off," temperature as quickly as possible after the vehicle is started. It also
means, however, that the exhaust catalyst(s) in the close-coupled location(s) are subject to higher
exhaust temperatures during fully-warmed up operation. Pd is required in closed-coupled
catalysts due to its resistance to high-temperature thermal sintering thereby maintaining
sufficient durability of the emissions control system over the useful life of a vehicle. Sulfur
1-11
-------
removal from Pd requires rich operation at higher temperatures than required for sulfur removal
from other PGM catalysts.27
In addition to its interaction with catalyst materials, sulfur can also react with the wash-
coating itself to form alumina sulfate, which in turn can block coating pores and reduce gaseous
diffusion to active materials below the coating surface (see Figure 1-2).28 This may be a
significant mechanism for the observed storage of sulfur compounds at light and moderate load
operation with subsequent, rapid release as sulfate particulate matter emissions when high-load,
9Q
high-temperature conditions are encountered.
Operating the catalyst at a sufficiently high temperature under net reducing conditions
(e.g., air-to-fuel equivalence that is net fuel-rich of stoichiometry) can effectively release the
sulfur oxides from catalyst components. Thus, regular operation at sufficiently high
temperatures at net fuel-rich air-to-fuel ratios can minimize the effects of fuel sulfur levels on
catalyst active materials and catalyst efficiency; however, it cannot completely eliminate the
effects of sulfur poisoning. In current vehicles, desulfurization conditions occur typically at high
loads when there is a degree of commanded enrichment (i.e., fuel enrichment commanded by the
engine management system primarily for protection of engine and/or exhaust system
components). A study of Tier 2 vehicles in the in-use fleet recently completed by EPA30 shows
that emission levels immediately following high speed/load operation is still a function of fuel
sulfur level for the gasoline used following desulfurization. If a vehicle operates on gasoline
with less than 10 ppm sulfur, exhaust emissions stabilize over repeat FTP tests at emissions near
those of the first FTP that follows the high speed/load operation and catalyst desulfurization. If
the vehicle continues to operate on higher sulfur gasoline following desulfurization, exhaust
emissions creep upward until a new equilibrium exhaust emissions level is established. This
suggests that lower fuel sulfur levels achieve emission benefits unachievable by catalyst
desulfurization procedures alone. Continued operation on gasoline with a 10 ppm average
sulfur content or lower is necessary after catalyst desulfurization in order to achieve emissions
reductions with the current in-use fleet.31 Furthermore, regular operation at the high exhaust
temperatures and rich air-to-fuel ratios necessary for catalyst desulfurization is not desirable and
may not be possible for future Tier 3 vehicles for several reasons:
• Thermal sintering and resultant catalyst degradation: The temperatures necessary
to release sulfur oxides are high enough to lead to thermal degradation of the
catalyst over time via thermal sintering of active materials. Sintering reduces the
surface area available to participate in reactions and thus reduces the overall
effectiveness of the catalyst.
• Operational conditions: It is not always possible to maintain fuel-rich operational
conditions and exhaust catalyst temperatures that are high enough for sulfur
removal because of cold weather, idle conditions and light-load operation.
• Increased emissions: In order to achieve greater emission reductions across a
fuller range of in-use driving conditions, vehicle manufacturers' use of
commanded enrichment, which has been beneficial for sulfur removal, will be
greatly reduced or eliminated under Tier 3. Additionally, the fuel-rich air-to-fuel
ratios necessary for sulfur removal from active catalytic surfaces would result in
1-12
-------
increased PM, NMOG, CO and air toxic emissions, particularly at the high-
temperature, high load conditions (e.g., US06 or comparable) necessary for sulfur
removal. Previously used levels of commanded enrichment (e.g., under Tier 2)
would interfere with the strategies necessary to comply with more stringent Tier 3
SFTP exhaust emissions standards. There are also additional provisions within the
Tier 3 standards that further restrict the use of US06 and off-cycle commanded
enrichment in an effort to reduce high-load and off-cycle PM, NMOG, CO and air
toxic emissions.32
• Expected changes to engine performance necessary to reduce fuel consumption
and greenhouse gas emissions will improve the thermal efficiency of engines and
may result in reduced exhaust temperatures.
1.2.4.2 Previous Studies of Gasoline Sulfur Impacts
This section summarizes studies to provide historical context regarding what is known
about the direct impacts of gasoline sulfur on vehicle exhaust emissions. Reducing fuel sulfur
levels has been the primary regulatory mechanism EPA has used to minimize sulfur
contamination of exhaust catalysts and to ensure optimum emissions performance over the useful
life of a vehicle. The impact of gasoline sulfur on exhaust catalyst systems has become even
more important as vehicle emission standards have become more stringent. Studies have
suggested a progressive increase in catalyst sensitivity to sulfur (in terms of percent conversion
efficiency) when standards increase in stringency and emissions levels decrease. Emission
standards under the programs that preceded the Tier 2 program (Tier 0, Tier 1 and National LEV,
or NLEV) were high enough that the impact of sulfur was considered of little importance. The
Tier 2 program recognized the importance of sulfur and reduced the sulfur levels in the fuel from
around 300 ppm to 30 ppm in conjunction with the new emission standards.33 At that time, very
little work had been done to evaluate the effect of further reductions in fuel sulfur, especially on
in-use vehicles that may have some degree of catalyst deterioration due to real-world operation
or on vehicles with extremely low tailpipe emissions, as described earlier.
In 2005, EPA and several automakers jointly conducted a research program, the Mobile
Source Air Toxics (MSAT) Study that examined the effects of sulfur and other gasoline
properties such as benzene and volatility on emissions from a fleet of nine Tier 2 compliant
vehicles.34 The study found significant reductions in NOx, CO and total hydrocarbons (HC)
when the vehicles were tested on low sulfur fuel, relative to 32 ppm fuel. In particular, the study
found a 48 percent increase in NOx over the FTP when gasoline sulfur was increased from 6
ppm to 32 ppm. Given the preparatory procedures related to catalyst clean-out and loading used
by these studies, these results may represent a "best case" scenario relative to what would be
expected under more typical driving conditions. Nonetheless, these data suggested the effect of
in-use sulfur loading was largely reversible for Tier 2 vehicles, and that there were likely to be
significant emission reductions possible with further reductions in gasoline sulfur level. More
recently, EPA completed a comprehensive study on the effects of gasoline sulfur on the exhaust
emissions of Tier 2 vehicles at low to moderate mileage levels.35 Further details of this study are
summarized in Section 1.2.4.3of this preamble.
1-13
-------
In the NPRM, we summarized the limited data available regarding the impact of gasoline
sulfur on the near-zero exhaust emission vehicle technologies that will be necessary for Tier 3
compliance. Vehicles certified to California LEV II SULEV and PZEV standards and federal
Tier 2 Bin 2 standards achieve levels of exhaust emissions control consistent with the levels of
control that will be necessary for Tier 3 compliance. While these vehicles represent only a
relatively small subset (e.g., typically small light-duty vehicles and light-duty trucks with limited
GVWR or towing utility) of the broad range of vehicles that will need to comply with Tier 3
standards as part of a fleet-wide average, data on these vehicles provides an opportunity to study
the impact of gasoline sulfur on near-zero exhaust emission technologies and is generally
representative of technology that are expected to be used with mid-size and smaller light-duty
vehicles for Tier 3 compliance. Vehicle testing by Toyota (Takei et al.) of LEV I, LEV IIULEV
and prototype SULEV vehicles showed larger percentage increases in NOx and HC emissions
for SULEV vehicles as gasoline sulfur increased from 8 ppm to 30 ppm, as compared to other
LEV vehicles they tested36. Ball et al. of Umicore Autocat USA, Inc. studied the impact of
gasoline fuel sulfur levels of 3 ppm and 33 ppm on the emissions of a 2009 Chevrolet Malibu
PZEV37 Umicore's testing of the Malibu PZEV vehicle showed a pronounced and progressive
trend of increasing NOx emissions (referred to as "NOx creep") when switching from a 3 ppm
sulfur gasoline to repeated, back-to-back FTP tests using 33 ppm sulfur gasoline. The PZEV
Chevrolet Malibu, after being aged to an equivalent of 150,000 miles, demonstrated emissions at
a level consistent with the Tier 3 Bin 30 NMOG+NOx standards when operated on 3 ppm sulfur
fuel and for at least one FTP test after switching to 33 ppm certification fuel. Following
operation over 2 FTP cycles on 33 ppm sulfur fuel, NOx emissions alone were more than double
the Tier 3 30 mg/mi NMOG+NOx standard.10 This represents [a 70% NOx increase between 3
ppm sulfur and 33 ppm sulfur gasolines, approximately 2-3 times of what has been previously
reported for similar changes in fuel sulfur level for Tier 2 and older vehicles.38'39
Both the Umicore and Toyota studies suggest that the emissions from vehicles using
near-zero exhaust emissions control technology similar to what is expected for compliance with
the Tier 3 standards are more sensitive to changes in gasoline sulfur content at low (sub-30 ppm)
sulfur concentrations than technology used to meet the higher Federal Tier 2 and California LEV
II standards. The Umicore and Toyota studies clearly indicate that a progressive increase in
catalyst sensitivity to sulfur continues as exhaust emissions decrease from levels required by
federal Tier 2 and California LEV II emissions standards to the lower levels required by Tier 3
emissions standards. In addition, although vehicles with Tier 2 technology have somewhat less
sulfur sensitivity compared to future Tier 3 vehicles, there is still significant opportunity for
further emissions reductions from the existing in-use fleet by reducing gasoline sulfur content
from 30 ppm to 10 ppm. The results of recent testing demonstrating the potential for in-use
emissions reductions from further gasoline sulfur control are summarized in Section 1.2.43).
Recent data on the impact of gasoline sulfur on vehicles with exhaust emission control
technologies that we expect to be used with Tier 3 vehicles is summarized in Sections 1.2.4.4
and 1.2.4.5.
1.2.4.3 EPA Testing of Gasoline Sulfur Effects on Tier 2 Vehicles and the In-Use Fleet
Both the MSAT40 and Umicore41 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
1-14
-------
time impacts emissions, and what the state of catalyst sulfur loading is for the typical vehicle in
the field. In response to these data needs, 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
49
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 93 cars and light trucks recruited
from owners in southeast Michigan, covering model years 2007-9 with approximately 20,000-
40,000 odometer miles.43 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 test fuel, one at a sulfur level of 5
ppm and the other at 28 ppm. 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 as well as 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, demonstrating that sulfur loadings have a large effect on exhaust catalyst
performance, and that Tier 2 vehicles can achieve significant reductions based on removing, at
least in part, the negative impact of the sulfur loading on catalyst efficiency (Table 1-4). For
example, Bag 2 NOx emissions dropped 31 percent between the pre- and post-cleanout tests on
28 ppm fuel.
1-15
-------
Table 1-4 Percent Reduction in In-Use Emissions After the Clean-out
Using 28 ppm Test Fuel"
Bagl
Bag 2
Bag3
FTP
Composite
Bag 1 - Bag 3
NOX
(p-value)
31.4%
(0.0003)
35.4%
(O.OOOl)
11.4%
(0.0002)
THC
(p-value)
14.9%
(0.0118)
20.4%
(O.OOOl)
3.8%
(0.0249)
CO
(p-value)
6.0%
(0.0151)
—
21.5%
(0.0001)
6.8%
(0.0107)
7.2%
(0.0656)
NMHC
(p-value)
18.7%
(0.0131)
27.7%
(O.OOOl)
3.5%
(0.0498)
CH4
(p-value)
14.4%
(0.0019)
10.3%
(O.OOOl)
6.0%
(0.0011)
PM
(p-value)
15.4%
(<
0.0001)
—
24.5%
(O.OOOl)
13.7%
(O.OOOl)
a The clean-out effect is not significant at a = 0.10 when no reduction estimate is provided.
To assess the impact of lower sulfur fuel on in-use emissions, further testing was
conducted on a representative subset of vehicles on 28 ppm and 5 ppm fuel with accumulated
mileage. A first step in this portion of the study was to assess the differences in the effectiveness
of the clean-out procedure under different fuel sulfur levels. Table 1-5 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 34 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. This further
indicates that current sulfur levels in gasoline continue to have a long-term, adverse effect on
exhaust emissions control that is not fully removed by intermittent clean-out procedures that can
occur in day-to-day operation of a vehicle and demonstrates that lowering sulfur levels to 10
ppm on average will significantly reduce the effects of sulfur impairment on emissions control
technology.
Table 1-5 Percent Reduction in Exhaust Emissions When Going from 28 ppm to 5 ppm
Sulfur Gasoline for the First Three Repeat FTP Tests Immediately Following Clean-out
Bagl
Bag 2
Bag 3
FTP Composite
NOX
(p-value)
5.3%
(0.0513)
34.4%
(0.0036)
42.5%
(O.OOOl)
15.0%
THC
(p-value)
6.8%
(0.0053)
33.9%
(O.OOOl)
36.9%
(O.OOOl)
13.3%
CO
(p-value)
6.2%
(0.0083)
a
14.7%
(0.0041)
8.5%
NMHC
(p-value)
5.7%
(0.0276)
26.4%
(0.0420)
51.7%
(O.OOOl)
10.9%
CH4
(p-value)
14.0%
(O.OOOl)
49.4%
(O.OOOl)
28.5%
(O.OOOl)
23.6%
PMa
—
—
—
—
1-16
-------
Bag 1 - Bag 3
(0.0002)
a
(O.OOOl)
a
(0.0050)
a
(0.0012)
a
(O.OOOl)
a
-
"The effectiveness of clean-out cycle is not significant at a = 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
Table 1-6. Reductions for Bag 2 NOx were particularly high, estimated at 52 percent between 28
ppm and 5 ppm overall. For all pollutants, 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.
Table 1-6 Percent Reduction in Emissions from 28 ppm to 5 ppm Fuel Sulfur
on In-Use Tier 2 Vehicles
Bagl
Bag 2
Bag3
FTP Composite
Bag 1 - Bag 3
NOX
(p-value)
7.1%
(0.0216)
51.9%
(< 0.0001)
47.8%
(< 0.0001)
14.1%
(0.0008)
a
THC
(p-value)
9.2%
(0.0002)
43.3%
(< 0.0001)
40.2%
(< 0.0001)
15.3%
(< 0.0001)
5.9%
(0.0074)
CO
(p-value)
6.7%
(0.0131)
a
15.9%
(0.0003)
9.5%
(< 0.0001)
a
NMHC
(p-value)
8.1%
(0.0017)
42.7%
(0.0003)
54.7%
(< 0.0001)
12.4%
(< 0.0001)
b
CH4
(p-value)
16.6%
(< 0.0001)
51.8%
(< 0.0001)
29.2%
(< 0.0001)
29.3%
(< 0.0001)
b
NOx+NMOG
(p-value)
N/A
N/A
N/A
14.4%
(< 0.0001)
N/A
PMa
-
-
-
-
-
a Sulfur level not significant at a =
b Inconclusive because the mixed
= 0.10.
model did not converge.
Major findings from this study include:
• Largely 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.
1-17
-------
• Reducing fuel sulfur levels from current levels to levels in the range of the Tier 3
gasoline sulfur standards is expected to achieve significant reductions in emissions of
NOx, hydrocarbons, and other pollutants of interest in the current in-use fleet.
• Assuming that the emissions impacts vs. gasoline sulfur content are approximately
linear, changing gasoline sulfur content from 30 ppm to 10 ppm would result in
NMOG+NOx emissions decreasing from 52 mg/mi to 45 mg/mi, respectively (a 13%
decrease), and NOX emissions decreasing from 19 mg/mi to 16 mg/mi, respectively (a
16% decrease), for the vehicles in the study.
To evaluate the robustness of the statistical analyses assessing the overall in-use
emissions reduction between operation on high and low sulfur fuel (Table 1-6), a series of
sensitivity analyses were performed to assess the impacts on study results of measurements from
low-emitting vehicles and influential vehicles, as documented in detail in the report.44 The
sensitivity analyses showed that the magnitude and the statistical significance of the results were
not impacted and thus demonstrated that the results are statistically robust. We also subjected the
design of the experiment and data analysis to a contractor-led independent peer-review process
in accordance with EPA's peer review guidance. The results of the peer review45'46 largely
supported the study design, statistical analyses, and the conclusions from the program and raised
only minor concerns that have not changed the overall conclusions and have subsequently been
addressed in the final version of the report.47
Overall, the reductions found in this study are in agreement with other low sulfur studies
conducted on Tier 2 vehicles, namely MSAT and Umicore studies mentioned above, in terms of
4R
the magnitude of NOx and HC reductions when switching from 28 ppm to 5 ppm fuel. We
have reviewed the results of the emission effects study performed by SGS, which was included
with API's comments on the Tier 3 proposal, and have concluded that these results are also
consistent with the findings of EPA's Tier 2 in-use study, specifically that exhaust emissions
performance is sensitive to fuel sulfur level.49 The SGS study also suggests that negative effects
of exposure to a somewhat higher sulfur level (80 ppm in this case) are largely reversible for Tier
2 vehicles, meaning that reducing fuel sulfur levels nationwide will bring significant immediate
benefits by reducing emissions of the existing fleet. For further details regarding the Tier 2 In-
Use Gasoline Sulfur Effects Study, see the final report.50
As a follow-on phase to the Tier 2 in-use study, EPA analyzed five vehicles51 certified to
Tier 2 Bin 4, LEV IIULEV and LEV II SULEV exhaust emissions standards to assess the
gasoline sulfur sensitivity of Tier 2 and California LEV II vehicles with emission levels
approaching or comparable to the Tier 3 standards. The analysis found that these low-emitting
Tier 2 vehicles showed similar or greater sensitivity to fuel sulfur levels compared to the original
Tier 2 test fleet - for example, a 24 percent reduction in FTP composite NOx emissions when
sulfur is reduced from 28 ppm to 5 ppm.52 Test results discussed below in Section 1.2.4.4 also
confirm that there is significantly increased sensitivity of exhaust emissions to gasoline sulfur as
vehicle technologies advance towards exhaust emissions approaching near-zero emissions (e.g.,
Tier 3 Bin 50 and lower). The impact of fuel sulfur on vehicles with exhaust emission control
technologies that we expect to be used with Tier 3 vehicles is summarized in the next two
Sections (1.2.4.4 and 1.2.4.5).
1-18
-------
EPA believes that the studies by EPA and others described in this Section strongly
support our conclusion that reducing gasoline sulfur content to a 10 ppm average will result in
significant exhaust emissions reductions from the current in-use fleet. However, some
commenters have expressed concerns about the relevance and appropriateness of the data, as
well as the conclusions drawn from them. The Summary and Analysis of Comments document,
available in the docket for this rulemaking, provides our responses to those comments.
1.2.4.4 Testing of Gasoline Sulfur Effects on Vehicles with Tier 3/LEV III Technology
The Tier 3 fleet average exhaust emissions standards of 30 mg/mi NMOG+NOx will
require large reductions of emissions across a broad range of light-duty vehicles and trucks with
differing degrees of utility. Previous studies of sulfur impacts on extremely low exhaust
emission vehicles (e.g., Toyota, Umicore) were limited to mid-size or smaller light-duty
vehicles. There are currently nonhybrid LDT3 or any LDT4 vehicles certified at or below
Federal Tier 2 Bin 3 or to the California LEV II SULEV exhaust emission standards. At the time
of the Tier 3 NPRM, EPA was not aware of any existing data demonstrating the impact of
changes in gasoline sulfur content on larger vehicles with technology comparable to what would
be expected for compliance with Tier 3 exhaust emission standards. In their supplemental
comments to the Tier 3 proposal, API criticized EPA's reliance on emissions data from older
vehicles that were not considered to be examples of future Tier-3-like vehicles. In order to
further evaluate this issue, the Agency initiated a test program at EPA's National Vehicle and
Fuel Emissions Laboratory (NVFEL) in Ann Arbor, Michigan. The Agency obtained a heavy-
light-duty truck and applied changes to the design and layout of the exhaust catalyst system and
to the calibration of the engine management system consistent with our engineering analyses of
technology necessary to meet Tier 3 Bin 30 emissions with a 20 to 40% compliance margin at
150,000 miles. EPA also requested that Umicore loan the Agency the vehicle tested in their
study to undergo further evaluation of gasoline sulfur impacts on exhaust emissions. In addition,
Ford Motor Company completed testing of fuel sulfur effects on a Tier 3/LEV III developmental
heavy-light-duty truck and submitted a summary report of their findings as part of their
supplemental comments to the Tier 3 NPRM. The results of these three test programs are
summarized below.
1.2.4.4.1 Ford Motor Company Tier 3 Sulfur Test Program
Ford Motor Company recently completed testing of a heavy-light-duty truck (i.e.,
between 6000 and 8500 pounds GVWR) under development to meet the Tier 3 Bin 50 standards
on two different fuel sulfur levels and submitted the resulting data to EPA as part of its
supplemental comments.53'54 The test results from this vehicle are particularly important when
considering the following factors:
• These are the first detailed emissions data submitted by a vehicle manufacturer to
the Agency demonstrating emissions of a heavy-light-duty-truck consistent with
Tier 3 Bin 50 or lower emissions levels.
• The truck tested uses a version of Ford's 2.0 L GTDI engine, an engine with high
BMEP (approximately 23-bar) that can allow significant engine displacement
1-19
-------
downsizing while maintaining the truck's utility. This is a key enabling GHG
reduction strategy analyzed by EPA in the 2017-2025 GHG Final Rule.55
• The vehicle was specifically under development by a vehicle manufacturer with
an engineering target of meeting Tier 3 Bin 50 and LEV III ULEV50 exhaust
emissions standards.
Turbocharged, downsized engines are key technologies within Ford's strategy to reduce GHG
emissions.56 EPA expects that trucks with configurations similar to this developmental Ford
Explorer (downsized engines with reduced GHG emissions and very low emissions of
NMOG+NOx) will become increasingly prevalent within the timeframe of the implementation of
the Tier 3 regulations.
The developmental truck used close-coupling of both catalyst substrates and relatively
high PGM loading (150 g/ft3). Ford used accelerated aging of the catalysts and 62 sensors to an
equivalent of 150,000 miles (the Tier 3 full useful life). The developmental hardware and engine
management calibration configuration of this truck was designed to meet federalTier 3 Bin 50
and California LEV III ULEV50 standards of 50 mg/mi NMOG+NOx at 150,000 miles. The
emissions data submitted by Ford included NOx and NMHC emissions during operation on E10
California LEV III certification fuel at two different sulfur levels, 10 ppm and 26.5 ppm. Ford
did not provide NMOG emissions data but there was sufficient information for EPA to calculate
NMOG emissions from the provided NMHC data using calculations from Title 40 CFR §
1066.665.
The truck demonstrated average FTP NMOG+NOX emissions of 37 mg/mi on the 10 ppm
E10 California LEV III fuel, emissions that are consistent with compliance with Bin 50 and
ULEV50 standards with a reasonable margin of compliance (emissions at approximately 70% of
the standard). Retesting of the same vehicle on LEV3 E10 blended57 to 26.5 ppm S resulted in
average NMOG+NOx emissions of 53 mg/mi, 6% above the Tier 3 Bin 50 standard. Ford found
a high level of statistical significance with respect to the increase of emissions with increasing
fuel sulfur. Assuming a linear effect of sulfur on emissions performance, NMOG+NOx
emissions would be approximately 56 mg/mi at 30 ppm sulfur, which is approximately 12%
above the Bin 50 exhaust emissions standard. This also represents an increase in NMOG+NOx
emissions of 53% with an approximate doubling of NOx emissions and a 13 % increase in
NMOG for 30 ppm sulfur gasoline vs. 10 ppm sulfur gasoline.
The advanced technology Ford truck, which was shown to be capable of complying with
the Tier 3 Bin 50 standard with a reasonable margin of compliance on 10 ppm sulfur gasoline, in
effect reverted to approximately LEV IIULEV exhaust emissions levels when tested on higher
sulfur gasoline, equivalent to the previous level of emissions control to which earlier models of
this vehicle were certified for MY 2013. The effect of increasing gasoline sulfur levels from 10
ppm to 30 ppm58 on this vehicle essentially negated the entire benefit of the advances in
emissions control technology that were applied by the vehicle manufacturer to meet
developmental goals for compliance with Tier 3 standards. This clearly indicates, for this
vehicle model using technology representative of what would be expected for compliance with
Tier 3 Bin 50 and post 2017 GHG standards, reducing gasoline sulfur to 10 ppm is needed for
the advances in technology to achieve their intended effectiveness in reducing NMOG+NOx
1-20
-------
emissions. The advances in vehicle technology and the reduction in gasoline sulfur clearly are
both needed to achieve the emissions reductions called for by Tier 3.
1.2.4.4.2 EPA Re-test of Umicore 2009 ChevroletMalibu PZEV
Ball et al. of Umicore Autocat USA, Inc. previously studied the impact of gasoline fuel
sulfur levels of 3 ppm and 33 ppm on the emissions of a 2009 Chevrolet Malibu PZEV.59 In
their supplemental comments to the Tier 3 proposal, API commented hat the composition of the
two test fuels outside of sulfur content was not held constant and thus the exhaust emissions
differences attributed to the difference in gasoline sulfur levels may have been due to other fuel
property differences.60 For example, the 3 ppm fuel used by Ball et al. was nonoxygenated EEE
Clear test fuel (essentially, Tier 2 Federal certification gasoline except with near-zero sulfur)
while the 33 ppm fuel was an oxygenated California Phase 2 LEV II certification fuel. Thus it
was not entirely clear if the changes in NOX emissions observed between tests with the two fuels
were significantly impacted by fuel composition variables other than gasoline sulfur content.
EPA obtained the same test vehicle from Umicore for retesting at the EPA NVFEL facility using
the 5 ppm and 28 ppm sulfur EO test fuels and vehicle test procedures used in EPA gasoline
sulfur effects testing on Tier 2 vehicles (see Section 1.2.4.2).
In EPA's retest of the 2009 Chevrolet Malibu PZEV, when sulfur was the only difference
between the test fuels, the gasoline with higher sulfur resulted in significantly higher increases in
NOx emissions with increasing fuel sulfur content than was observed in the previous testing by
Ball et al. at Umicore. Assuming emissions impacts vs. gasoline sulfur content are
approximately linear, the original data from Ball et al. result in a predicted increase in NOx
emissions of approximately 40% when increasing gasoline sulfur from 10 ppm to 30 ppm. The
EPA re-testing of the same vehicle that controlled for other fuel composition differences result in
a predicted increase in NOx emissions of 93% when increasing gasoline sulfur from 10 ppm to
30 ppm, with NOx emissions approximately doubling from 22 g/mi to 43 g/mi, with no
statistically significant difference in NMOG emissions and with an increase in NMOG+NOx
emissions of 56%. The approximate doubling in NOx emissions with the Malibu PZEV between
10 ppm and 30 ppm sulfur was nearly identical to the results found during testing of the Tier 3
Bin 50 developmental Ford Explorer discussed above. The results confirm that fuel
compositional differences other than sulfur may have impacted exhaust emissions results in the
Ball et al. study by masking a substantial portion of the effect of increased fuel sulfur on NOx
emissions. When controlling for other fuel composition differences, the resultant increase in
NOx exhaust emissions due to increasing gasoline sulfur was more than double that of the
original Ball et al. study. The observed increase in NMOG+NOx emissions during EPA testing
of the Malibu PZEV was also comparable to results found with the developmental Tier 3 Bin 50
Ford Explorer. There was also a much higher increase in NOx and NMOG+NOx emissions for
both the Malibu PZEV and the Tier 3 Bin 50 Explorer with increased gasoline sulfur than was
observed with Tier 2 vehicles in the EPA Tier 2 in-use study.61
1.2.4.4.3 EPA Prototype Tier 3 Heavy-light-duty Truck Test Program
EPA purchased a 2011 Chevrolet Silverado heavy-light-duty (LDT4) pickup truck with a
developmental goal of modifying the truck to achieve exhaust emissions consistent with
compliance with the Tier 3 Bin 30 emissions standards. The truck was equipped with a 5.3L V8
1-21
-------
with General Motors' "Active Fuel Management" cylinder deactivation system. This particular
truck was chosen in part because cylinder deactivation is a key technology for light-truck
compliance with future GHG standards and in part because it achieved very low emissions in its
OEM, Tier 2-compliant configuration (certified to Tier 2 Bin 4). A prototype exhaust system
was obtained from MECA consisting of high-cell-density (900 cpsi) thin-wall (2.5 mil), high-
PGM, close-coupled Pd-Rh catalysts with an additional under-body Pd-Rh catalyst. The total
catalyst volume was approximately 116 in3 with a specific PGM loading of 125 g/ft3 and
approximate loading ratio of 0:80:5 (Pt:Pd:Rh). Third-party (non-OEM) EMS calibration tools
were used to modify the powertrain calibration in an effort to improve catalyst light-off
performance. The final test configuration used approximately 4 degrees of timing retard and
approximately 200 rpm higher idle speed relative to the OEM configuration during and
immediately following cold-start. The exhaust catalyst system and HEGO sensors were bench
aged to an equivalent 150,000 miles using standard EPA accelerated catalyst bench-aging
procedures.62 The truck was tested on California LEV III E10 certification fuel at 9 and 29 ppm
gasoline sulfur levels.
The EPA Tier 3 prototype Silverado achieved NMOG+NOx emissions of 18 mg/mi on
the 9 ppm S fuel. The NMOG+NOx emissions were approximately 60% of the Bin 30 standard
and thus are consistent with meeting the Tier 3 Bin 30 exhaust emissions standard with a
moderate compliance margin. NMOG+NOx emissions increased to 29 mg/mi on the 29 ppm S
fuel and one out of four tests exceeded the Bin 30 exhaust emissions standards. NMOG+NOx
emissions would be at 19 mg/mi and 30 mg/mi with 10 ppm and 30 ppm gasoline sulfur,
respectively, assuming a linear effect of sulfur on emissions performance. This represents an
increase in NMOG+NOx emissions of approximately 55%, comparable to increases observed
with both the EPA-tested Chevrolet Malibu PZEV and the developmental Tier 3 Bin 50 Ford
Explorer. The impact of increased gasoline sulfur on NMOG+NOx emissions was due to
comparable increases (on a percentage basis) in both NMOG and NOx emissions. This effect of
gasoline sulfur on the Prototype Silverado truck's emissions differed from the sulfur impacts
observed on the developmental Ford Explorer, which primarily affected NOx emissions, and the
Malibu PZEV, where the impact was entirely on NOx emissions.
1.2.4.5 Gasoline Sulfur Level Necessary for New Light-duty Vehicles to Achieve Tier 3
Exhaust Emissions Standards
Meeting Tier 3 NMOG+NOx standards will require major reductions in exhaust
emissions across the entire fleet of new light-duty vehicles. As discussed in previous sections,
the Tier 3 program will require reductions in fleet average NMOG+NOX emissions of over 80
percent for the entire fleet of light-duty vehicles and light-duty trucks. This significant level of
fleet average emission reduction will require reductions from all parts of the fleet, including
vehicles models with exhaust emissions currently at or near the level of the fully phased-in Tier 3
FTP NMOG+NOx fleet average standard of 30 mg/mi.
Compliance with the more stringent Tier 3 fleet average standards will require vehicle
manufacturers to certify a significant amount of vehicles to bin standards that are below the Bin
30 fleet average standard to offset other vehicles that are certified to bin standards that remain
somewhat above the Bin 30 fleet average even after significantly reducing their emissions. At
the same time, the stringency of the Tier 3 standards will push almost all vehicle models to be
1-22
-------
close to or below the Bin 30 fleet average standard. There are only 2 compliance bins below Bin
30, i.e., Bin 20 and Bin 0, available to offset emissions of vehicles certifying above Bin 30.
There is also very limited ability for vehicle manufacturers to certify vehicles below the stringent
Tier 3 fleet average exhaust emissions standard since Bin 20 and Bin 30 standards for individual
vehicle certification families are approaching the engineering limits of what can be achieved for
vehicles using an internal combustion engine and Bin 0 can only be achieved by electric-only
vehicle operation. The result is that there is a very limited ability to offset sales of vehicles
certified above the 30 mg/mi fleet average emission standard. This means in general that vehicle
models currently with higher emissions will have to achieve significant emissions reductions to
minimize the gap, if any, between their certified bin levels under Tier 3 and the Tier 3 Bin 30
fleet average standard, and vehicle models currently at or below Bin 30 will also have to achieve
further emissions reductions under Tier 3 to offset the vehicles that remain certified to bin
standards somewhat above Bin 30 . The end result is a need for major reductions from all types
of vehicles in the light-duty fleet, including those above as well as most vehicles that are already
near, at, or below the Tier 3 Bin 30 fleet average standard.
Achieving exhaust emissions reductions of over 80% for the fleet, with major reductions
across all types of light-duty vehicles and light-duty trucks, will be a major technological
challenge. Vehicles already have made significant advances in controlling cold start emissions
and maximizing exhaust catalyst efficiency (e.g., improving warm-up and catalyst light-off after
cold starts and maintaining very high catalyst efficiency once warmed up) in order to meet Tier 2
and LEV II emissions standards. There are no "low-hanging fruit" remaining for additional
NMOG+NOx reductions from light-duty vehicles from a technology perspective, meaning that
vehicle manufacturers cannot merely change one aspect of emissions control and thereby achieve
all of the required reductions. Instead, compliance with light-duty Tier 3 exhaust emissions
standards will require significant improvements in all areas of emissions control - with further
improvements in fuel-system management and mixture preparation during cold start,
improvements in achieving catalyst light-off immediately after cold start, and improved catalyst
efficiency during stabilized, fully-warmed-up conditions. Manufacturers will need further
improvements in each of these areas with nearly every vehicle in order to comply with the fleet-
average Tier 3 standards.
From a technology perspective, the most likely control strategies will involve using
exhaust catalyst technologies and powertrain calibration primarily focused on reducing cold-start
emissions of NMOG, and on reducing both cold-start and warmed-up (running) emissions of
NOx. An important part of this strategy, particularly for larger vehicles having greater difficulty
achieving cold-start NMOG emissions control, will be to reduce NOx emissions to near-zero
levels. This will involve controlling engine-out NOx emissions during cold start, shortening the
cold start period prior to catalyst light-off of NOx reduction reactions, and better controlling
NOx emissions once the catalyst is fully warmed up. This is needed to allow a sufficient NMOG
compliance margin so that vehicles can meet the combined NMOG+NOx emissions standards
for their full useful life.
While significant NMOG+NOx emissions reductions can be achieved from better control
of cold start NMOG emissions, there are practical engineering limits to NMOG control for larger
displacement vehicles (e.g., large light-duty trucks with significant payload and etrailer towing
capabilities). This is based in part on the impact on NMOG emissions of the larger engine
1-23
-------
surface-to-volume ratio and resultant heat conduction from the combustion chamber during
warm-up. There are also tradeoffs between some cold-start NMOG controls and cold-start NOx
control. For example, secondary air injection and/or leaner fueling strategies improve catalyst
light-off for NMOG after a cold-start but also place OSC components in an oxidation state that
limits potential for NOx reduction and thus often result in higher cold-start NOx emissions.
Some applications achieve lower NMOG+NOx emissions without the use of secondary air
injection by careful calibration, changes to the catalyst formulation and balancing of catalyst HC
and NOx activity. The EPA Prototype Silverado and the developmental Ford Explorer are
specific examples of this approach.
Because of engineering limitations with large vehicles, heavy-light-trucks and other
vehicles with significant utility, we expect many applications will need close to 100% efficiency
in NOx control under fully warmed-up conditions and very fast light-off of NOx reduction
reactions over the exhaust catalyst almost immediately after cold-start for those applications.
This will require significant improvements in catalytic and engine-out NOx reduction compared
with Tier 2 vehicles and will be especially important for heavier vehicles due to the challenges of
achieving low NMOG.
These technology improvements - improving warm-up and catalyst light-off after cold
starts and maintaining very high catalyst efficiency - - once warmed up - all rely on 10 ppm
average sulfur fuel to achieve the very significant emissions reductions required for the fleet to
achieve the Tier 3 Bin 30 fleet average emissions standard. The evidence from the test results
and specific vehicle examples discussed above clearly indicate that leaving the gasoline sulfur
level at 30 ppm would largely negate the benefits of key technology improvements expected to
be used for compliance with Tier 3 exhaust emissions standards. Without the lower 10 ppm
gasoline sulfur content, the Tier 3 exhaust fleet average emissions standards would not be
achievable across the broad range of vehicles that must achieve significant exhaust emissions
reductions.
One aspect of the need for sulfur levels of 10 ppm average stems from the fact that
achieving the Tier 3 emission standards will require very careful control of the exhaust chemistry
and exhaust temperatures to ensure high catalyst efficiency. The impact of sulfur on OSC
components in the catalyst makes this a challenge even at relatively low (10 ppm) gasoline sulfur
levels. NOx conversion by exhaust catalysts is strongly influenced by the OSC components like
ceria. Ceria sulfation may play an important role in the large degradation of NOx emission
control with increased fuel sulfur levels observed in the MSAT, Umicore and EPA Tier 2 In-Use
Gasoline Sulfur Effects studies and the much more severe NOx emissions degradation observed
in recent test data from PZEV and prototype/developmental Tier 3/LEV III vehicles.63
The importance of lower sulfur gasoline is also demonstrated by the fact that vehicles
certified to California SULEV are typically certified to higher bins for the federal Tier 2
program. Light-duty vehicles certified to CARB SULEV and federal Tier 2 Bin 2 exhaust
emission standards accounted for approximately 3.1 percent and 0.4 percent, respectively, of
vehicle sales for MY2009. Light-duty vehicles certified to SULEV under LEV II are more
typically certified federally to Tier 2 Bin 3, Bin 4 or Bin 5, and vehicles certified to SULEV and
Tier 2 Bins 3-5 comprised approximately 2.5 percent of sales for MY2009. In particular,
nonhybrid vehicles certified in California as SULEV are not certified to federal Tier 2 Bin 2
1-24
-------
emissions standards even though the numeric limits for NOx and NMOG are shared between the
California LEV II and federal Tier 2 programs for SULEV and Bin 2. Confidential business
information shared by the auto companies indicate that the primary reason is an inability to
demonstrate compliance with SULEV/Bin 2 emission standards after vehicles have operated in-
use on gasoline with greater than 10 ppm sulfur and with exposure to the higher sulfur gasoline
sold nationwide. While vehicles certified to the LEV II SULEV and Tier 2 Bin 2 standards both
demonstrate compliance using certification gasoline with 15-40 ppm sulfur content, in-use
compliance of SULEV vehicles in California occurs after significant, sustained operation on
gasoline with an average of 10 ppm sulfur and a maximum cap of 30 ppm sulfur while federally
certified vehicles under the Tier 2 program operate on gasoline with an average of 30 ppm sulfur
and a maximum cap of 80 ppm sulfur. Although the SULEV and Tier 2 Bin 2 standards are
numerically equivalent, the increased sulfur exposure of in-use vehicles certified under the
federal Tier 2 program results in a need for a higher emissions compliance margin to take into
account the impact of in-use gasoline sulfur on full useful life vehicle emissions. As a result,
vehicles certified to California SULEV typically certify to emissions standards under the federal
Tier 2 program that are 1-2 certification bins higher (e.g., SULEV certified federally as Tier 2
Bin 3 or Bin 4) in order to ensure in-use compliance with emissions standards out to the full
useful life of the vehicle when operating on higher-sulfur gasoline.
There are currently no LDTs larger than LDT2 with the exception of a single hybrid
electric SUV certified to Tier 2 Bin 2 or SULEV emissions standards. We expect that additional
catalyst technologies, for example increasing catalyst surface area (volume or substrate cell
density) and/or increased PGM loading, will need to be applied to larger vehicles in order to
achieve the catalyst efficiencies necessary to comply with the Tier 3 standards, and any sulfur
impact on catalyst efficiency will have a larger impact on vehicles and trucks that rely more on
very high catalyst efficiencies in order to achieve very low emissions. The vehicle emissions data
referenced in Section 1.2.4.4 represents the only known data on non-hybrid vehicles spanning a
range from mid-size LDVs to heavy-light-trucks at the very low criteria pollutant emissions
levels that will be needed to comply with the Tier 3 exhaust emissions standards. The
developmental Ford Explorer, Chevrolet Malibu PZEV and EPA prototype Chevrolet Silverado
vehicles described in Section 1.2.4.3 also represent a range of different technology approaches to
both criteria pollution control and GHG reduction (e.g., use of secondary air vs. emphasizing
cold-start NOx control, use of engine downsizing via turbocharging vs. cylinder deactivation for
GHG control, etc.) and represent a broad range of vehicle applications and utility (mid-size
LDV, LDT3, LDT4). All of the vehicles with Tier 3/LEV III technology demonstrated greater
than 50% increases in NMOG+NOx emissions when increasing gasoline sulfur from 10 ppm to
30 ppm. Two of the vehicles showed a doubling of NOx emissions when increasing gasoline
sulfur from 10 ppm to 30 ppm. Both of the heavy-light-duty trucks with specific engineering
targets of meeting Tier 3 emissions were capable of meeting their targeted emission standards
with a sufficient compliance margin on 10 ppm sulfur gasoline and could not meet their targeted
emissions standards or could not achieve a reasonable compliance margin when tested with 30
ppm sulfur gasoline.
The negative impact of gasoline sulfur on catalytic activity and the resultant loss of
exhaust catalyst effectiveness to chemically reduce NOx and oxidize NMOG occur across all
vehicle categories. However, the impact of gasoline sulfur on the effectiveness of exhaust
1-25
-------
catalysts to control NOx emissions in the fully-warmed-up condition is particularly of concern
for larger vehicles (the largest LDVs and LDT3s, LDT4s, and MDPVs). Manufacturers face the
most significant challenges in reducing cold-start NMOG emissions for these vehicles. Because
of the need to reach near-zero NOx emissions levels in order to offset engineering limitations on
further NMOG exhaust emissions control with these vehicles, any significant degradation in
NOx emissions control over the useful life of the vehicle would likely prevent some if not most
larger vehicles from reaching a combined NMOG+NOx low enough to comply with the 30
mg/mi fleet-average standard. Any degradation in catalyst performance due to gasoline sulfur
would reduce or eliminate the margin necessary to ensure in-use compliance with the Tier 3
emissions standards. Certifying to a useful life of 150,000 miles versus the current 120,000
miles will further add to manufacturers' compliance challenge for Tier 3 large light trucks (See
Section IV.7.b of the preamble for more on the useful life requirements.) These vehicles
represent a sufficiently large segment of light-duty vehicle sales now and for the foreseeable
future such that their emissions could not be sufficiently offset (and thus the fleet-average
standard could not be achieved) by certifying other vehicles to bins below the fleet average
standard.
As discussed above, achieving Tier 3 levels as an average across the light-duty fleet will
require fleet wide reductions of approximately 80%. This will require significant reductions
from all light duty vehicles, with the result that some models and types of vehicles will be at
most somewhat above the Tier 3 level, and all other models will be at or somewhat below Tier 3
levels. Achieving these reductions presents a major technology challenge. The required
reductions are of a magnitude that EPA expects manufacturers to employ advances in technology
in all of the relevant areas of emissions control - reducing engine-out emissions, reducing the
time to catalyst lightoff, improving exhaust catalyst durability at 120,000 or 150,000 miles and
improving efficiency of fully warmed up exhaust catalysts. All of these areas of emissions
control need to be improved, and gasoline sulfur reduction to a 10 ppm average is a critical part
of achieving Tier 3 levels through these emissions control technology improvements.
The use of 10 ppm average sulfur fuel is an essential part of achieving Tier 3 levels while
applying an array of advancements in emissions control technology to the light-duty fleet. The
testing of Tier 2 and Tier 3 type technology vehicles, as well as other information, shows that
sulfur has a very large impact on the effectiveness of the control technologies expected to be
used in Tier 3 vehicles. Without the reduction in sulfur to a 10 ppm average, the major
technology improvements projected under Tier 3 would only result in a limited portion of the
emissions reductions needed to achieve Tier 3 levels. For example, without the reduction in
sulfur from a 30 ppm to 10 ppm average, the technology improvements would not come close to
achieving Tier 3 levels, and in some cases might have no more effectiveness than current Tier 2
technology and achieve only approximately Tier 2 levels of exhaust emissions control.
Achieving Tier 3 levels without a reduction in sulfur to 10 ppm levels would only be
possible if there were technology improvements significantly above and beyond those discussed
above. Theoretically, without reducing sulfur levels to 10 ppm average, emissions control
technology improvements would need to provide upwards of twice as much, and in some cases
significantly more than twice as much, emissions control effectiveness as the Tier 3 technology
improvements discussed above in Section 1.2.4.4. EPA has not identified technology
improvements that could provide such a large additional increase in emissions control
1-26
-------
effectiveness, across the light-duty fleet, above and beyond that provided by the major
improvements in technology discussed above, without any additional gasoline reductions in
gasoline sulfur content. The impact of sulfur reduction on the effectiveness of the available
technology improvements plays such a large role in achieving the Tier 3 levels that there would
be no reasonable basis to expect that technology would be available, at the 30 ppm sulfur level,
to fill the emission control gap left from no sulfur reduction, and achieve the very significant
fleetwide reductions needed to meet the Tier 3 fleet average standards. In effect reducing sulfur
from 30 ppm to 10 ppm has such a large impact on the ability of the technology improvements to
achieve Tier 3 emissions levels that absent these sulfur reductions there is not a suite of
technology advancements available to fill the resulting gap in emissions reductions. Technology
would not be available that would achieve the Tier 3 Bin 30 average standard, across the fleet,
with sulfur at 30 ppm levels, and as a result Tier 3 levels would not be technically feasible and
achievable.
This analysis also applies to gasoline sulfur levels between 10 and 30 ppm, e.g. 20 ppm.
The Tier 3 required emissions reductions are so large and widespread across the fleet, and the
technology challenges are sufficiently high, especially for heavier vehicles, that the large
increase in emissions that would occur from a higher average sulfur level compared to a 10 ppm
average would lead to an inability for vehicle technologies to widely achieve Tier 3 levels as a
fleet wide average in order to meet the Tier 3 Bin 30 fleet average standard.
EPA acknowledges that some models in the light-duty fleet, when viewed in isolation,
may be able to achieve Tier 3 levels at current sulfur levels of 30 ppm average. Under the Tier 3
fleet average standards, it is not sufficient for one or a few of a manufacturer's vehicle models to
meet Tier 3 levels because the manufacturer's light-duty vehicle fleet as a whole must achieve
the Tier 3 30 mg/mi exhaust emissions standard as a fleet-wide average. As discussed above, all
vehicle models will need to achieve further reductions and be either below or no more than
somewhat above Tier 3 levels to achieve the Tier 3 standard as a fleet wide average. Absent the
reductions in sulfur levels to 10 ppm average, this is not achievable from a technology
perspective.
As discussed in Section V.A.2 of the preamble, the 10 ppm standard for sulfur in
gasoline represents the lowest practical limit from a standpoint of fuel production, handling and
transport. While lowering gasoline sulfur to average levels below 10 ppm would further help
ensure in-use vehicle compliance with the Tier 3 standards, the Agency believes that a gasoline
sulfur standard of 10 ppm, combined with the advances in emissions control technology
discussed above, will enable vehicle manufacturers to achieve compliance with a national fleet
average standard of 30 mg/mi NMOG+NOx. Not only will a 10 ppm sulfur standard enable
vehicle manufacturers to certify their entire product line of vehicles to the Tier 3 fleet average
standards, but based on the results of testing both Tier 2 vehicles and SULEV vehicles as
discussed above, reducing gasoline sulfur to a 10 ppm average should enable these vehicles to
maintain their emission performance in-use over their full useful life.
1-27
-------
1.3 SFTP NMOG+NOx Feasibility
The new Tier 3 emission requirements include stringent NMOG+ NOx composite
standards over the SFTP that will generally only require additional focus on fuel control of the
engines and diligent implementation of new technologies like gasoline direct injection (GDI) and
turbocharged engines. Additionally, the fleet-average nature of the standards will allow more
challenged vehicles to be offset by vehicles that could outperform the required fleet averages.
In order to assess the technical feasibility of a 50 mg/mi NMOG+NOx national fleet
average SFTP composite standard EPA conducted an analysis of data from the in-use
verification program (IUVP). The IUVP vehicles are tested by manufacturers at various ages and
mileages and the results are reported to EPA. The analysis was performed on Tier 2 and LEV II
vehicles. This provided a baseline for the current federal fleet emissions performance, as well as
the emissions performance of the California LEV II fleet.
1.3.1 Assessment of the Current Federal Fleet Emissions
To investigate feasibility, we acquired and analyzed IUVP certification results for model
years 2010 and 201 1 which represent the most recent model years of which complete IUVP data
sets are currently available. These data included FTP composites, as well as results for the US06,
and SC03 cycles. We focused on results for hydrocarbons (HC) and NOx. For the FTP results
HC represents non-methane organic gases (NMOG). The US06 and SC03 results represent
NMHC+NOx.
As a first step, we averaged the results by model year and test group (engine family).
After compiling results on all three cycles for each test group, we calculated SFTP composite
estimates for each engine family as
SFTP = 0.35 - (ĄTPN
NOx
) + 0.28 • US06 + 0.37 • SC03
As a second step, we then averaged the SFTP composite results by standard level and
vehicle class, focusing on results in Bins 2, 3 and 5, as well as vehicles certified to LEV-II LEV
and SULEV standards. In averaging, we treated Bin 2 and LEV-II/SULEV standards as
equivalent, and accordingly, pooled their results. Table 1-7 shows the numbers of test groups in
each combination of standard level and vehicle class.
Table 1-7 Numbers of Test Groups Certified to Selected Tier-2 and LEV-II Standards in
Model Years 2010 and 2011
Standard Level
Bin 2 + LEV-II/SULEV
Bin3
Bin 5
LEV-II/LEV
Vehicle Class
LDV-LDTl
88
26
331
124
LDT2
3
1
37
17
LDT3
13
4
LDT4
1
14
4
1-28
-------
Figure 1-3 shows results for Bin-5 and LEV-II/LEV vehicles. It is clear that vehicles in
all four vehicle classes, from LDV to LDT4, are certified to these standards. The means show a
modest, but not striking increase with vehicle class, from approximately 30 mg/mi for LDV to
approximately 50 mg/mi for LDT4. However, an equivalent trend among the maxima is not
evident. The results also show that assuming equivalence between these two standards is
reasonable. On average it is clear that test groups certified at the Bin-5 level are capable of
meeting the target level of 50 mg/mi, although with small compliance margins. However,
relatively small numbers of families exceed this level, ranging to over 100 mg/mi.
Additionally, Figure 1-4 shows results for test groups certified to Bin-2 and Bin-3
standards. For these test groups, a trend with vehicle class is not evident, although very small
numbers of test groups are certified as trucks. In contrast to the Bin-5 vehicles, most families
certified at the Bin-2 and Bin-3 levels are well below the 50-mg/mi level, and maxima are no
higher than 7 percent below this level.
LDV-T1
LDT2
LDT3
LDT4
Figure 1-3: Mean and Maximum Composite SFTP Results for HC+NOx for Test Groups
certified to Bin-5 and LEV-II/LEV Standards (bars and error-bars represent means and
maxima for sets of test groups, respectively)
1-29
-------
Bin 2/LEVII-SULEV
Bin3
LDV-T1
LDT2
LDT3
LDT4
Figure 1-4: Mean and Maximum Composite SFTP Results for HC+NOX for Test Groups
certified to Bin-2 and Bin-3 Standards (bars and error-bars represent means and maxima
for sets of test groups, respectively)
1.4 Technology Description for NMOG+NOX Control
A range of technology options exist to reduce NMOG and NOx emissions from both
gasoline fueled spark ignition and diesel engines below the current Tier 2 standards. Available
options include modifications to the engine calibration, engine design, exhaust system, and after
treatment systems. The different available options have specific benefits and limitations. This
section describes the technical challenges to reducing emissions from current levels, describes
available technologies for reducing emissions, estimates the potential emissions reduction of the
different technologies, describes if there are other ancillary benefits to engine and vehicle
performance with the technology, and reviews the limits of each technology. Except where
noted, these technologies are applicable to all gasoline vehicles covered by this rule. Unique
diesel technologies are addressed in Section 1.4.2.
1.4.1 Summary of the Technical Challenge for NMOG+NOx control
The Tier 3 emission standards will require vehicle manufacturers to reduce the level of
both NMOG and NOx emissions from the existing Tier 2 fleet by approximately 80 percent over
the FTP by 2025. The FTP measures emissions during cold start, hot start, and warmed-up
vehicle city driving. The majority of NMOG and NOx emissions from gasoline fueled vehicles
measured during the FTP test historically occur during the cold start phase however emissions
during warmed-up and hot operation cannot be ignored and must be limited in order to meet Tier
3 standards. Figure 1-1, above, graphically demonstrates when NMOG and NOx emissions are
produced during a cold start. As shown in the figure, approximately 90 percent of the NMOG
emissions occur during the first 50 seconds after the cold start. In addition, about 60 percent of
the NOx emissions occur during this same 50 second period. Unlike NMOG which is mostly
controlled after the first 50 seconds, NOx emissions tend to be released throughout the remainder
of the FTP test and are particularly sensitive to fuel sulfur content. Achieving the Tier 3
1-30
-------
NMOG+NOx FTP emissions standard may require manufacturers to reduce both cold start
NMOG and NOx emissions and further reduce NOx emissions when the vehicle is warmed up.
The Tier 3 emission standards will also require manufacturers to maintain their current
vehicle high load NMOG+NOx emission performance as measured during the US06 operation of
the Supplemental Federal Test Procedure (SFTP). The US06 component of the SFTP is
designed to simulate higher speeds and acceleration rates during warmed up vehicle operation.
Significant quantities of NMOG and NOx emissions are produced during the US06 portion of the
SFTP if enrichment events occur to reduce exhaust temperatures during high-load operation.
Most vehicles are now avoiding these enrichment events during the US06 and achieve relatively
low NMOG+NOx emissions.
It is anticipated that manufacturers will change the design of their exhaust and catalyst
systems to reduce catalyst light-off times and reduce warmed-up and hot running emissions,
particularly NOx to almost zero, in order to achieve the Tier 330 mg/mi FTP NMOG+NOx
standard. Design changes to reduce catalyst light-off time can also result in higher catalyst
temperatures during high-load operation as seen during the US06 test. To achieve the
NMOG+NOx Tier 3 SFTP standard manufacturers will need to develop and implement
technologies to manage catalyst temperatures during high-load operation without using fuel
enrichment.
In addition, it is anticipated that the technologies manufacturers will use for reducing
warmed up NOx emissions during the FTP will also reduce NOx emissions during warmed up
operation on the US06.
For the catalyst to effectively reduce NMOG+NOx emissions it must reach the light-off
temperature of approximately 250 °C. Emissions during the catalyst warm up period can be
reduced by reducing the emissions produced by the engine during the catalyst warm up phase.
Emissions can also be reduced by shortening the time period required for the catalyst to reach the
light-off temperature. Reducing warmed-up NOx emissions requires improving the efficiency of
the catalyst system which will generally require little to no presence of sulfur contaminants in the
fuel.
To achieve the Tier 3 NMOG+NOx FTP emissions standards it is anticipated that vehicle
manufacturers will focus on three areas to reduce emissions:
• minimizing the emissions produced by the engine before the catalyst reaches the
light-off temperature;
reducing the time required for the catalyst to reach the light-off temperature; and,
improving the NOx efficiency of the catalyst during warmed-up operation.
It is anticipated that improvements in all three areas will be required particularly for
heavier passenger cars, light-duty trucks in classes LDT3 and LDT4, and MDPVs. The NOx
efficiency during warmed-up operation of vehicles certified to the Tier 2 Bin 4 emission level
and operating on low sulfur fuel (i.e. 10 ppm or lower) are such that it is anticipated that
1-31
-------
reductions in cold start emissions are primarily what will be required to achieve the Tier 3
NMOG+NOx standard.
Heavy-duty trucks (8,501 up to 14,000 Ibs) will have a similar challenge to meet their
Tier 3 standards along with the new SFTP requirements for this vehicle class. In addition to the
new test requirements and tighter standards, these vehicles useful life is being extended from
120,000 to 150,000 miles. Unlike lighter weight vehicles, heavy-duty trucks tend to operate at
higher loads for greater periods of time and therefore have different constraints to meet the new
requirements and more stringent standards.
For spark-ignition engines, the higher operating load of these vehicles limits the ability to
move the catalyst close to the cylinder head due to durability concerns from higher thermal
loading. This limit will constrain the ability of these trucks to quickly light-off the catalyst, it
will, however, allow them to stay out of fuel-enriched operation to maintain catalyst
temperatures when the vehicle is being operated under high load. The emissions produced
during fuel-enrichment events, which occur at high loads can be significantly greater than the
reductions achievable during the cold start and idle phase. Fuel enrichment events cause criteria
pollutant and CO2 emission rates to increase and also reduces the vehicle's fuel economy. To
achieve the NMOG+NOx FTP emissions standards while also meeting the new SFTP
requirements it is anticipated that heavy-duty vehicle manufacturers will focus on four areas for
spark ignition engines:
• reducing the emissions produced by the engine before the catalyst reaches light-
off temperature;
• reducing the time required for the catalyst to reach the light-off temperature;
• improving the NOx efficiency of the catalyst during warmed up operation; and,
• minimizing the time spent in fuel enrichment to reduce the operating temperature
of the catalyst.
Compression ignition or diesel engines also have limitations with thermal goals and
location of the emission control system on the vehicle. With the similar goal of providing engine
exhaust heat to the catalysts, SCR and DPF, these emission control systems may compete with
each other for thermal energy. Additionally, the SCR system and the DPF generally require
sufficient capacity or size to handle the emissions from the engine which may limit the ability to
locate them in the optimal location.
To meet Tier 3 NMOG+NOx FTP emissions standards while also meeting the new SFTP
requirements it is anticipated that heavy-duty vehicle manufacturers will focus on three areas for
compression ignition:
• reducing the emissions produced by the engine while the catalysts and SCR
system are being brought to proper operating temperature;
1-32
-------
• reducing the time required for the catalysts and SCR system to reach the proper
operating temperature;
• improving the NOx efficiency of the SCR during warmed up operation through
refinement in engine out emission controls and SCR strategies.
1.4.1.1 Reducing Engine Emissions Produced Before Catalyst Light-Off
During the first 50 seconds of the cold start phase of the FTP the engine is operating
either at idle or low speed and load in non-hybrid vehicles. The engine temperature is between
20 and 30 °C (68 and 86 °F). At these temperatures and under these low loads the cold engine
produces lower concentrations of NOx than NMOG. As the engine warms up and as the load
increases the concentration of NOx produced by the engine increases and the concentration of
NMOG decreases.
The design of the air induction system, combustion chamber, spark plug, and fuel
injection system determines the quantity of fuel required for stable combustion to occur in the
cold engine. Optimizing the performance of these components can provide reductions in the
amount of fuel required to produce stable combustion during these cold operating conditions.
Reductions in the amount of fuel required leads to reductions in cold start NMOG emissions.
The design considerations to minimize cold start emissions are also dependent on the fuel
injection method. Port fuel injected (PFI) engines have different design constraints than gasoline
direct injection (GDI) spark ignition engines. For both PFI and GDI engines, however, attention
to the details affecting the in cylinder air/fuel mixture can reduce cold start NMOG emissions.
It has been shown that cold start NMOG emissions in PFI engines can be reduced by
reducing the size of the fuel spray droplets and optimizing the spray targeting. Fuel impinging
on cold engine surfaces in the cylinder does not readily vaporize and does not combust.
Improving injector targeting to reduce the amount of fuel reaching the cylinder walls reduces the
amount of fuel needed to create a combustible air fuel mixture. Reducing the size of the spray
droplets improves the vaporization of the fuel and creation of a combustible mixture.64
Droplet size can be reduced by modifying the injector orifice plate and also by increasing
the fuel pressure. Reducing droplet size and improving fuel vaporization during cold start has
been shown to reduce cold transient emissions by up to 40 percent during the cold start phase of
the FTP emission test.65 This and other PFI injector technology improvements have been used to
optimize the cold start performance of today's vehicles certified to the CA LEV II SULEV
standards.
The mixture formation process in a DISI engine is different than a PFI engine. In a PFI
engine the fuel is injected during the intake stroke of the engine in the intake runner. The fuel
has time to evaporate during the intake stroke as the fuel and air are drawn into the cylinder. In
addition, as the engine warms up the fuel can be injected into the intake runner and engine heat
can assist in evaporating the fuel prior to the intake valve opening.
1-33
-------
The DISI engine injects fuel at higher fuel pressures than PFI engines directly into the
combustion chamber. In a DISI engine the fuel droplets need to evaporate and mix with the air
in the cylinder in order to form a flammable mixture.
Injecting directly into the cylinder reduces the time available for the fuel to evaporate and
mix with the intake air in a DISI engine compared to a PFI engine. An advantage of the DISI
design is that the fuel spray does not impinge on the walls of the intake manifold or other
surfaces in the cylinder.
DISI systems have the ability to split the injection timing event. At least one study has
indicated that significant reductions in hydrocarbon emissions can be achieved by splitting the
injections during the cold start of a DISI engine. An initial injection occurs during the intake
stroke and a second injection is timed to occur during the compression stroke. This injection
method reduced unburned hydrocarbon emissions 30 percent compared to a compression stroke
only injection method.66
These are two examples of specific engine design characteristics, fuel injector design and
fuel system pressure on PFI engines and injection timing on GDI engines which can be used to
reduce cold start NMOG emissions significantly during the engine warm up prior to the catalyst
reaching the light-off temperature.
Optimizing the fuel injection system design and calibration is anticipated to be used in all
vehicle classes, including heavy-duty vehicles. It is anticipated that these described
improvements, along with improvements to other engine design characteristics, will be used to
reduce cold start emissions for passenger cars, LDTs, MDPVs, and HDTs.
Because the engine is relatively cold and the operating loads are low during the first 50
seconds of the FTP the engines typically do not produce significant quantities of NOx emissions
during this phase. In addition manufacturers tend to retard the combustion timing during the
catalyst warm up phase. Retarding combustion timing has been shown also to reduce the
concentration of NMOG in the exhaust. This calibration method further reduces peak
combustion temperatures while increasing the exhaust gas temperature compared to optimized
combustion timing. The increased exhaust gas temperature leads to improved heating of the
catalyst and reduced catalyst light-off times. Retarding combustion and other technologies for
reducing catalyst light-off time are discussed in the following section.
1.4.1.2 Reducing Catalyst Light-Off Time
The effectiveness of current vehicle emissions control systems depends in large part on
the time it takes for the catalyst to light-off, which is typically defined as the catalyst reaching a
temperature of 250°C. In order to reduce catalyst light-off time, it is expected manufacturers will
use technologies that will improve heat transfer to the catalyst during the cold start phase and
improve catalyst efficiency at lower temperatures. Technologies to reduce catalyst light-off time
include calibration changes, thermal management, close-coupled catalysts, catalyst PGM
loading, and secondary air injection. It is anticipated that in some cases where the catalyst light-
off time and efficiency are not sufficient to reduce cold start NMOG emissions, hydrocarbon
adsorbers may be utilized. The adsorbers trap hydrocarbons until such time that the catalyst is
1-34
-------
fully warmed up and the emissions can be oxidized by the catalyst. Note that with the exception
of hydrocarbon adsorbers each of these technologies addresses NMOG and NOx performance.
The technologies are described in greater detail below.
1.4.1.2.1 Calibration Changes to Reduce Catalyst Light-Off Time
These include calibration changes to increase the temperature and mass flow of the
exhaust prior to the catalyst reaching the light-off temperature. By reducing the time required for
the catalyst to light-off engine calibration changes can affect NMOG and NOx emissions.
Retarding combustion in a cold engine by retarding the spark advance is a well known
method for reducing the concentration of NMOG emissions in the exhaust and increasing the
exhaust gas temperature.67'68 The reduction in NMOG concentrations is due to a large fraction of
the unburned fuel within the cylinder combusting before the flame is extinguished on the
cylinder wall. Reductions of total hydrocarbon mass of up to 40 percent have been reported
from studies evaluating the effect of spark retard on exhaust emissions.
In addition to reducing the NMOG exhaust concentrations retarding the spark advance
reduces the torque produced by the engine. In order to produce the same torque and maintain the
engine speed and load at the desired level when retarding the spark advance, the air flow into the
engine is increased causing the manifold pressure to increase which can also improve
combustion stability. Retarding the combustion process also results in an increase in the exhaust
gas temperature. The retarded ignition timing during the cold start phase in addition to reducing
the NMOG emissions increases the exhaust mass flow and exhaust temperature. These changes
lead to a reduction in the time required to heat the catalyst.
The torque produced by the engine will begin to vary as the spark retard amount reaches
engine combustion limits. As the torque variations increase, the combustion process is
deteriorating and the engine performance begins to degrade due to the partial burning. It is the
level of this variability which defines the absolute maximum reduction in spark advance that can
be utilized to reduce NMOG emissions and reduce the catalyst light-off time.
Retarding combustion during cold start can be applied to spark-ignition engines in all
vehicle classes. The exhaust temperatures and NMOG emission reductions will vary based on
engine design. This calibration methodology is anticipated to be used to improve catalyst warm-
up times and reduce cold start NMOG emissions for all vehicle classes, passenger cars, LDTs,
MDPVs, and HDTs.
With the penetration of variable valve timing technology increasing in gasoline-fueled
engines additional work is being performed to characterize the impact of valve timing on cold
start emissions. The potential exists that calibration changes to the valve timing during the cold
start phase will lead to additional reductions in cold start NMOG emissions.69
1.4.1.2.2 Exhaust System Thermal Management to Reduce Catalyst Light-Off Time
This category of technologies includes all design attributes meant to conduct combustion
heat into the catalyst with minimal cooling. This includes insulating the exhaust piping between
1-35
-------
the engine and the catalyst, reducing the wetted area of the exhaust path, reducing the thermal
mass of the exhaust system, and/or using close-coupled catalysts (i.e., the catalysts are packaged
as close to the engine cylinder head as possible to mitigate the cooling effects of longer exhaust
piping). By reducing the time required to light-off the catalyst, thermal management
technologies reduce NMOG and NOx emissions.
Moving the catalyst closer to the cylinder head is a means manufacturers have been using
to reduce both thermal losses and the catalyst light-off time. Many vehicles today use close-
coupled catalysts, a catalyst which is physically located as close as possible to the cylinder head.
Moving the catalyst from an underbody location to within inches of the cylinder head reduces the
light-off time significantly.
Another means for reducing heat losses are to replace cast exhaust manifolds with thin-
wall stamped manifolds. Reducing the mass of the exhaust system reduces the heat losses of the
system. In addition an insulating air gap can be added to the exhaust system which further
reduces the heat losses from the exhaust system, insulating air gap manifolds are also known as
dual-wall manifolds.
With thin- and dual-wall exhaust manifolds, close-coupled catalyst housings can be
welded to the manifold. This reduces the needed for manifold to catalyst flanges which further
reduces the thermal inertia of the exhaust system. Close coupling of the catalyst and reducing
the thermal mass of the exhaust system significantly reduces the light-off time of the catalyst
compared to an underbody catalyst with flanges and pipes connected to a cast exhaust manifold.
Using close-coupled catalysts reduces the heat losses between the cylinder head and
catalyst. While reducing the time required to light-off the catalyst the close-coupled catalyst can
be subject to higher temperatures than underbody catalysts during high load operating
conditions. To ensure the catalyst does not degrade manufacturers currently use fuel enrichment
to maintain the exhaust temperatures below the levels which would damage the catalyst. It is
anticipated that to meet the Tier 3 SFTP standards, manufacturers will need to ensure that fuel
enrichment is not required on the US06 portion of the FTP. Calibration measures, other than fuel
enrichment, may be required to ensure the catalyst temperature does not exceed the maximum
limits.
Another technology beginning to be used for both reducing heat loss in the exhaust and
limiting exhaust gas temperatures under high load conditions is integrating the exhaust manifold
into the cylinder head. Honda utilized this technology on the Insight's 1.0 L VTEC-E engine.
The advantage of this technology is that it minimizes exhaust system heat loss during warm-up.
In addition with the exhaust manifold integrated in the cylinder head the cooling system can be
used to reduce the exhaust temperatures during high load operation. It is anticipated that
manufacturers will further develop this technology as means to both quickly light-off the catalyst
and reduce high-load exhaust temperatures.
To achieve the Tier 3 NMOG and NOx emissions standards it is expected that
manufacturers will optimize the thermal inertia of the exhaust system to minimize the time
needed for the catalyst to achieve the light-off temperature. In addition, the manufacturers will
need to ensure the high load performance does not cause thermal degradation of the catalyst
1-36
-------
system. It is expected that methods and technologies will be developed to reduce the need to use
fuel enrichment to reduce high load exhaust temperatures.
Optimizing the catalyst location and reducing the thermal inertia of the exhaust system
are design options manufacturers can apply to all vehicle classes (PCs, LDTs, MDPVs, and
HDTs) for improving vehicle cold start emission performance.
It is not anticipated HDTs with spark-ignition engines will utilize catalysts close-coupled
to the exhaust manifold. The higher operating loads of these engines results in durability
concerns due to high thermal loading. It is expected that manufacturers will work to optimize
the thermal mass of the exhaust systems to reduce losses along with optimizing the underbody
location of the catalyst. These changes are expected to improve the light-off time while not
subjecting the catalysts to the higher thermal loadings from a close coupled location.
1.4.1.2.3 Catalyst Design Changes
A number of different catalyst design changes can be implemented to reduce the time for
the catalyst to light-off. Changes include modifying the substrate design, replacing a large
volume catalyst with a cascade of two or more catalysts, and optimizing the loading and
composition of the platinum group metals (PGM).
Progress continues to be made in the development of the catalyst substrates which
provide the physical support for the catalyst components which typically include a high surface
area alumina carrier, ceria used for storing oxygen, PGM catalysts, and other components. A
key design parameter for substrates is the cell density. Today catalyst substrates can be
fabricated with cell densities up to 900 and 1,200 cells per square inch (cpsi) with wall
thicknesses approaching 0.05 mm.
Increasing the surface area of the catalyst improves the performance of the catalyst.
Higher substrate cell densities increases the surface area for a given catalyst volume. Higher
surface areas improve the catalyst efficiency and durability reducing NMOG and NOx emissions.
The limitation of the higher cell density substrates include increased exhaust system
pressures at high load conditions. The cell density and substrate frontal area are significant
factors that need to be considered to optimize the catalyst performance while limiting flow loss
at high load operation.
During the cold start phase of the FTP the engine speeds and load are low during the first
50 seconds of the test. One method for reducing the catalyst light-off time is to replace a larger
volume catalyst with two catalysts which total the same volume as the single catalyst. The
reduced volume close-coupled catalyst reduces the heat needed for this front catalyst to reach the
light-off temperature. The front catalyst of the two catalyst system will reach operating
temperature before the larger volume single catalyst, reducing the light-off time of the system.
All other parameters held constant, increasing the PGM loading of the catalyst also
improves the efficiency of the catalyst. The ratio of PGM metals is important as platinum,
palladium, and rhodium have different levels of effectiveness promoting oxidation and reduction
1-37
-------
reactions. Therefore, as the loading levels and composition of the PGM changes the light-off
performance for both NMOG and NOx need to be evaluated. Based on confidential
conversations with manufacturers it appears that there is an upper limit to the PGM loading,
beyond which further increases do not improve light-off or catalyst efficiency.
To achieve the Tier 3 NMOG and NOx emission standards it is anticipated that
manufacturers will make changes to catalyst substrates and PGM loadings. 70 To achieve the
emission levels required to meet the Tier 3 NMOG+NOx standard of 30 mg/mi with a
compliance margin will require very low sulfur levels in the fuel. As described in Section
1.2.3.3 even low levels (greater than 10 ppm) of sulfur in gasoline inhibit the ability of PGM
catalysts to achieve the low levels NOx emission levels of the Tier 3 standard. For the Tier 3
FTP emission standards to be achieved and maintained, particularly in use, it is required that the
sulfur content of the fuel be reduced to 10 ppm or lower.
Manufacturers will optimize the design of their aftertreatment systems for their different
vehicles. Primary considerations include cost, light-off performance, warmed-up conversion
efficiency and the exhaust temperatures encountered by the vehicle during high load operation.
Vehicles having low power to weight ratios will tend to have higher exhaust gas temperatures
and exhaust gas flow which will result in a different design when compared to vehicles having
higher power to weight ratios.
Manufacturers and catalyst suppliers perform detailed studies evaluating the cost and
emission performance of aftertreatment systems. It is anticipated that manufacturers will
optimize their aftertreatment systems to achieve the Tier 3 standards and meet the durability
criteria for all vehicle classes (PCs, LDTs, MDPVs, and HDTs).
1.4.1.2.4 Secondary Air Injection
By injecting air directly into the exhaust stream, close to the exhaust valve, combustion
can be maintained within the exhaust, creating additional heat thereby further increasing the
catalyst temperature. The air/fuel mixture must be adjusted to provide a richer exhaust gas for
the secondary air to be effective.
Secondary air injection has been used by a variety of passenger vehicle manufacturers to
assist with achieving the emission levels required of the CA LEV II SULEV standard.
Secondary air injection systems are used after the engine has started and once exhaust port
temperatures are sufficiently high to sustain combustion in the exhaust port. When the
secondary air pump is turned on the engine control module increases the amount of fuel being
injected into the engine. Sufficient fuel is added so that the air/fuel ratio in the cylinder is rich of
stoichiometry. The exhaust contains significant quantities of CO and hydrocarbons. The rich
exhaust gas mixes with the secondary air in the exhaust port and the combustion process
continues increasing the temperature of the exhaust and rapidly heating the manifold and close-
coupled catalyst.71' 2
Engines which do not use secondary air injection cannot operate rich of stoichiometry as
the added enrichment would cause increased NMOG emissions. The richer cold start calibration
used with vehicles that have a secondary air injection system provides a benefit as combustion
1-38
-------
stability is improved. In addition, the richer calibration is not as sensitive to changes in fuel
volatility. Less volatile fuels found in the market may result in poor start and idle performance
on engines calibrated to run lean during the cold operation. Engines which use secondary air and
have a richer warm up calibration would have a greater combustion stability margin.
Manufacturers may perceive this to be a benefit for the operation of their vehicles during the cold
start and warm up phase.
Installing a secondary air injection system combined with calibration changes can be used
by manufacturers to reduce the cold start emissions and improve the catalyst light-off on existing
engine designs. It is anticipated that manufacturers will utilize this technology to improve the
cold start performance on heavier vehicles and vehicles with low power to weight ratios.
Secondary air injection has been used on vehicles to achieve the CA LEVII SULEV emission
standards. This technology can be used on engines in all vehicle classes.
It is anticipated that secondary air injection will be used primarily in combination with
close coupled catalysts. Therefore, it is not anticipated that this technology will be used with
HDTs as it is not expected that the catalyst in these vehicles will be moved to a location
sufficiently close to the exhaust manifold to provide any improvement in catalyst light-off.
HDTs tend to operate at higher loads and catalyst durability is a concern due to the
increased thermal loading as the catalyst is moved closer to the cylinder head. Moving the
catalyst closer to the exhaust manifold would result in increasing the time spent in fuel
enrichment modes to ensure the temperatures are maintained below the threshold which would
reduce the durability of the catalyst. Using fuel enrichment to control catalyst temperature
causes significant increases in criteria pollutant emissions, CC>2 emissions and reductions in fuel
economy.
1.4.1.2.5 Hydrocarbon Adsorbers
Hydrocarbon adsorbers trap hydrocarbons emitted by the engine when the adsorber is at
low temperatures. As the temperature of the hydrocarbon adsorber increases the trapped
hydrocarbons are released. Passive adsorbers use an additional washcoat on an existing three-
way catalyst. The adsorber is a zeolite-based material which absorbs hydrocarbons at low
exhaust temperatures and desorbs hydrocarbons as the temperature increases. A significant
technical challenge to using a passive adsorber is to design the system such that the three-way
catalyst has reached the light-off temperature prior to the adsorber coating releasing the adsorbed
hydrocarbons.
Active adsorbers use a substrate with an adsorber washcoat over which the exhaust is
directed when the exhaust temperature is below the desorption temperature of the material. Once
the exhaust temperature reaches the desorption temperature the exhaust is routed such that it no
longer passes over the adsorber. As the adsorber continues to heat in the exhaust the captured
hydrocarbons are released and oxidized by the warmed-up catalyst system.
Adsorbers have been used to reduce cold start NMOG emissions on CA LEV II SULEV
vehicles. Additional work is being performed to further improve the performance of
hydrocarbon adsorbers.
1-39
-------
It is anticipated that if manufacturers have difficulty achieving the Tier 3 NMOG+NOx
emission standards because of challenging NMOG emission levels on the cold start,, they may
evaluate hydrocarbon adsorbers as an option to further reduce the NMOG emissions during the
cold start. One manufacturer used this approach to achieve the CA LEV II SULEV standard on a
large displacement V-8 engine with the application of an active hydrocarbon adsorber.73
Hydrocarbon adsorbers can be used on all spark-ignition engines and all classes of
vehicles. It is anticipated that these technologies may be required for engines with larger
displacement and in some of the larger vehicles. It is anticipated that HDTs will be able to
achieve the emissions levels required without the use of hydrocarbon adsorbers to meet the
standard.
1.4.1.3 Improving Catalyst NOx Efficiency during Warmed-up Operation
Significant quantities of NOx emissions are produced by vehicles during warmed-up
vehicle operation on the FTP for Tier 2certified vehicles. The stabilized NOx emission levels
will need to be reduced to achieve the Tier 3 NMOG+NOx emission standard. Improving the
NOx performance of the vehicle can be achieved by improving the catalyst efficiency during
warmed-up operation. As previously described the performance of the catalyst can be improved
by modifications to the catalyst substrate, increasing cell density, increasing PGM loadings and
particularly important, reducing the sulfur level of gasoline. Three-way catalyst efficiency is
also affected by frequency and amplitude of the air/fuel ratio. For some vehicles warmed-up
catalyst NOx efficiency can be improved by optimizing the air/fuel ratio control and limiting the
amplitude of the air fuel ratio excursions. It is anticipated that a combination of changes will be
made by manufacturers including further improvements to air/fuel ratio calibration and catalyst
changes including cell density and PGM loadings.
A requirement to ensure that the NOx emission performance of the vehicles is maintained
at or below the 30 mg/mi NMOG+NOx emission standard is reduced fuel sulfur concentrations.
As described in detail in Section 1.2.3.3 further reductions in fuel sulfur concentration are
required to ensure the catalyst performance is not degraded which causes increases in NOx
emissions beyond the Tier 3 standard.
It is anticipated that manufacturers will use these catalyst and calibration technologies to
improve the warmed up NOx emissions performance of vehicles in all classes, passenger cars,
LDTs, MDPVs, and HDTs.
1.4.1.4 EPA Estimates of Technology Improvements Required for Large Light-
Duty Trucks
Discussions with and comments from vehicle manufacturers indicated that large light-
duty trucks (e.g., pickups and full-size SUVs in the LDT3 and LDT4 categories) will likely be
the most challenging light-duty vehicles to bring into compliance with the Tier 3 NMOG+NOx
standards at the 30 mg/mi corporate average emissions level. A similar challenge was addressed
when large light-duty trucks were brought into compliance with the Tier 2 standards in the
previous decade. Figure 1-5 provides a graphical representation of the effectiveness of Tier 3
technologies when combined with gasoline sulfur control for large light-duty truck applications.
1-40
-------
The Tier 3 technologies shown are those that can be utilized on existing vehicles and do not
require engine design changes. A compliance margin is shown in both cases. Note that the
graphical representation of the effectiveness of catalyst technologies on NOx and NMOG when
going from Tier 2 to Tier 3 also includes a reduction in gasoline sulfur levels from 30 ppm to 10
ppm.
180*
170'
160*
150*
MO-
ISC'
120*
110*
100*
90*
80
70*
60*
50*
40*
30
20
10'
_OJ
1
o
§
on
V)
E
'5
.2
LU
X
O
}"""] NMOG+NOx Emissions of Tier 2 Bin 5 LDT3/4
JH NMOG+NOx Emissons of Tier 3 LD3/T4
60%
Compliance
Margin
Catalyst
PGM Loading
Thermal Management
67% Emissions
Reduction
Secondary
Air Injection
I Passive HC Adsorber I
Tier 3 NMOG+NOx Standard
20-40% Compliance Margin
Tier 2 Bin
w/30 ppm
5 LDT3/4
S Gasoline
Tier 3 LDT3/4
w/10 ppm S Gasoline
Figure 1-5: Contribution of the expected Tier 3 technologies to large light-duty truck
compliance with the Tier 3 standards with a comparison to Tier 2 Bin 5. The technologies
and levels of control are based on a combination of confidential business information
submitted by auto manufacturers and suppliers, public data and EPA staff engineering
judgment.
1.4.2 Diesel Technologies for Achieving Tier 3 NMOG and NOx Emission Requirements
Compared to spark-ignition engines, diesel engines typically produce very low NMOG
emissions. However, diesel engines do not operate at stoichiometry preventing them from using
emission control approaches similar to spark-ignition engines to control NOx emissions. The
technical challenge for diesel engines to achieve the Tier 3 NMOG+NOx emission levels will be
to obtain significant NOx emission reductions. It is anticipated that improvements in NOx
emissions performance of diesel exhaust catalysts during the cold start phase will be a major
1-41
-------
technical challenge. Depending on the performance of the exhaust catalyst system, additional
reductions in warmed-up NOx emissions may also be required to achieve the Tier 3 emission
levels.
It is not anticipated that diesel engines will have difficulty achieving the Tier 3 SFTP
emission standards. The exhaust catalyst system is fully warmed up and operational on the high
load portion of the SFTP, the US06. It is anticipated that manufacturers may need to optimize
the calibration of the selective catalytic reduction (SCR) system or the NOx adsorption catalyst
(NAC) system to ensure the systems achieve the required performance.
The technical task for achieving the Tier3 standards on all diesel engines in all vehicle
classes will be to have the exhaust catalysts reach operating temperatures early in the cold-start
phase of the FTP. To achieve these improvements it is anticipated that diesel manufacturers will
focus on means to reduce NOx emissions during the engine warm-up phase after the cold start
and reducing the time required for the SCR or NAC system to begin reducing (SCR) or capturing
and reducing (NAC) the NOx.
By controlling the timing of the fuel injection event, the number of fuel injection events
and the timing of intake and exhaust valve events, the temperature of the exhaust can be
increased. Diesel engine manufacturers will optimize the injection and valvetrain calibration to
increase the exhaust temperature after the engine is started and before it has reached operating
temperature.
As with gasoline engines, locating the exhaust catalyst system closer to the cylinder head
and air-gap insulating the exhaust system or reducing the mass of the exhaust components
upstream of the catalysts can be used to build and maintain heat in the exhaust system. A variety
of technologies are available to conduct combustion heat into the exhaust catalyst system with
minimal cooling. This includes uses of dual-wall, air-gapped exhaust piping between the engine
and the catalyst or trap; reducing the wetted area of the exhaust path; and reducing the thermal
mass of the exhaust system through use of thinner wall materials. By reducing the time required
to light-off the catalysts, thermal management technologies can reduce NOx emissions from
diesel engines. Once light-off has been achieved, NOx emissions reduction for modern, base-
metal zeolite SCR systems approaches that of modern three-way catalyst systems used for
stoichiometric gasoline spark-ignition applications.74
1.5 PM Feasibility
Particulate matter emitted from internal combustion engines is a multi-component
mixture composed of elemental carbon (or soot), semi-volatile organic compounds, sulfate
compounds (primarily sulfuric acid) with associated water, nitrate compounds and trace
quantities of metallic ash. At temperatures above 1,300K, fuel hydrocarbons without access to
oxidants can pyrolize to form particles of elemental carbon. Fuel pyrolysis can occur as the result
of operation at richer than stoichiometric air-to-fuel ratio (primarily PFI gasoline GDI engines),
direct fuel impingement onto surfaces exposed to combustion (primarily GDI and diesel engines)
and non-homogeneity of the air-fuel mixture during combustion (primarily diesel engines).
Elemental carbon particles that are formed can be oxidized during later stages of combustion via
in-cylinder charge motion and reaction with oxidants. Semi-volatile organic compounds (SVOC)
1-42
-------
are composed primarily of organic compounds from lubricant and partial combustion products
from fuel. PM emissions from SVOC are typically gas phase when exhausted from the engine
and contribute to PM emissions via particle adsorption and nucleation after mixing with air and
cooling. Essentially, PM-associated SVOC represent the condensable fraction of NMOG
emissions. Sulfur and nitrogen compounds are emitted primarily as gaseous species (862, NO
and NO2). Sulfate compounds can be a significant contributor to PM emissions from stratified
lean-burn gasoline engines and diesel engines, particularly under conditions where PGM-
containing exhaust catalysts used for control of gaseous and PM emissions oxidize a large
fraction of the 862 emissions to sulfate (primarily sulfuric acid). Sulfate compounds do not
significantly contribute to PM emissions from spark-ignition engines operated at near
stoichiometric air-fuel ratios due to insufficient availability of oxygen in the exhaust for
oxidation of SC>2 over PGM catalysts.
Elemental carbon PM emissions can be controlled by:
• Reducing fuel impingement on piston and cylinder surfaces
• Inducing charge motion and air-fuel mixing via charge motion (e.g., tumble and
swirl) or via multiple injection (e.g., GDI and diesel/common rail applications)
• Reducing or eliminating operation at net-fuel-rich air-to-fuel ratios (PFI gasoline
and GDI applications)
• Use of wall-flow or partial-wall-flow exhaust filters (diesel applications)
SVOC PM emissions can be controlled by:
• Reducing lubricating oil consumption
• Improvements in exhaust catalyst systems used to control gaseous NMOG
emissions (e.g., increased PGM surface area in the catalyst, improvements in
achieving catalyst light-off following cold-starts, etc.)
Sulfate PM emission can be controlled by:
• Reducing or eliminating sulfur from fuels
1.5.1 PM Emissions from Light-duty Tier 2 Vehicles
In order to establish the feasibility of the Tier 3 PM emission standards, EPA conducted a
test program to measure PM emissions from Tier 2 light-duty vehicles. The test program was
designed to measure PM emissions from late model year vehicles that represented a significant
volume of annual light duty-sales and included vehicles that ranged from small cars through
trucks. In addition, GDI vehicles were included in the program as were vehicles with known
high oil consumption.
1-43
-------
The Agency investigated PM emissions from Tier 2 light-duty vehicles. Seventeen model
year 2005-2010 Tier 2 Bin 4, 5, and 8 vehicles were tested at the U.S. EPA National Vehicle and
Fuel Emissions Laboratory (NVFEL) facility. A summary of their characteristics are provided in
Table 1-8. They included ten cars and seven trucks. Fifteen of these vehicles had accumulated
102,000-124,000 miles prior to the launch of the test program. One vehicle had accumulated
75,000 miles and another accumulated 21,000. Three cars and one truck were equipped with
GDI engines. Twelve of the fifteen test vehicles had previously been used in the DOE V4
Program. The remaining five vehicles were recruited in southeastern Michigan. One vehicle
(vehicle K) was suspected of having atypically high oil consumption and had only 37,000 miles
of mileage accumulation. Vehicle K was a duplicate of Vehicle C and was determined to
consume two and one half times the average oil consumption of vehicle C and three vehicles of
the same make, model and model year when tested within the DOE V4 program.17
The twelve vehicles acquired from the DOE V4 Program were selected to represent a
broad cross section of some of the highest sales vehicles in the U.S. market for model years
2005-2009. These vehicles had originally been purchased by DOE with odometer readings
ranging from 10,000-60,000 miles, placed in a mileage accumulation program and operated over
the EPA Standard Road Cycle on a test track or on mileage accumulation dynamometers to
110,000-120,000 miles.75 Immediately prior to inclusion in the EPA PM Test Program, the test
vehicles were serviced per the manufacturer's published service schedule and maintenance
procedures and underwent engine oil aging over a distance of 1,000 miles accumulated over the
EPA Standard Road Cycle to stabilize engine oil contribution to PM emissions76.
Three recruited test vehicles were selected because they used GDI technology. An
additional GDI equipped vehicle was obtained from the DOE V4 Program An attempt was made
to only recruit vehicles approaching the 120,000 mile useful life level. Testing was completed
for two of the four vehicles prior to the proposal of this rule. All of the recruited test vehicles
were thoroughly inspected, but otherwise tested as received.
All vehicles were tested on an El5 fuel with RVP, aromatic content, sulfur content, T50
and T90 of 9.1 psi, 23.8 vol%, 7 ppm, 160F and 31 IF, respectively. The properties of this fuel
approximated those of a projected El 5 market fuel.
The test program included three cold start and three hot start UDDS tests and three US06
tests conducted on each vehicle. FTP results were calculated for gaseous and PM emissions by
applying the cold-start and hot-start weighting factors to the complete cold and hot UDDS
results, respectively. This eliminated separate analysis of the typically very low concentration
FTP phase-2 gaseous and PM emissions samples and represented one method proposed within
40 CFR 1066 for increasing sample integration of measured gaseous and PM mass. During these
tests, triplicate PM samples were collected in parallel on PTFE membranes and single
(composite) PM samples were collected on primary and secondary quartz filters for TOT/TOR
OC/EC PM speciation analysis. Additional quartz filters were collected to determine the
F Vehicle K consumed approximately 1 quart per 3,000 miles vs. an average of approximately 1 quart per 8,000
miles for the other four vehicles of this make, model and year tested within the DOE V4 program.
1-44
-------
contribution of gas-phase artifact to the OC collected on the quartz filter samples. The
compositing of quartz filters over three repeats of each test was done to enhance the precision of
subsequent OC/EC thermogravimetric measurements. Single background (dilution air) PM
samples were also taken during each emissions test. Weekly tunnel blank and field blank PM
samples were also collected.
The following parameters were measured: NOx, NMHC, NMOG, alcohols, carbonyls,
CO, CH4, CC>2 and fuel consumption and PM mass as per the 40 CFR 1065 and the proposed 40
CFR part 1066 test procedures. Limited exhaust HC speciation was also performed.
PM composition was determined from filter samples taken on both quartz filters and
PTFE membranes. PM compositional analyses include determination of the contribution of
elemental and organic carbon to PM mass, 7 elemental analysis via EDXRF, sulfate analysis via
ion chromatography and determination of the contribution of unresolved complex organic PM
compounds by GC/MS.
Note that during the compositional analysis of the PM, EPA discovered a significant
amount of silicon deposited on some of the filters. The source of the silicon was determined to
be a silicone elastomer transfer tube used to connect vehicles to the emissions measurement
equipment. The data below reflect test results that are not subject to silicone contamination. For
additional information, refer to our memo to the docket78 which describes the original analysis
and corrective actions in greater detail.
1-45
-------
Table 1-8: Vehicles Tested as part of EPA's Light-Duty Vehicle PM Test Program
Vehicle Make, Model
and Designation
Honda Civic
Toyota Corolla
Honda Accord
Dodge Caliber
Chevrolet Impala
Ford Taurus
Toyota Tundra
Chrysler Caravan
Jeep Liberty
Ford Explorer
Honda Accord
Ford F 150
Chevrolet
Silverado
VW Passat
Manufacturer' s
Development
Vehicle13
Saturn Outlook
Cadillac STS4
A
B
C
D
E
F
G
H
I
J
K
L
P
M
N
O
Q
Model
Year
2009
2009
2007
2007
2006
2008
2005
2007
2009
2009
2007
2005
2006
2006
PC
2009
2010
Certified to
Emissions
Standard
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 5
Tier 2/Bin 4
Tier 2/Bin 5
Tier 2/Bin 8
Tier 2/Bin 8
Tier 2/Bin 5
None (Tier
2/LEVII ^
Prototype)*
Tier 2/Bin 5
Tier 2/Bin 5
Odometer at
Start of
Program, miles
121,329
120,929
123,695
114,706
114,284
115,444
121,243
116,742
121,590
121,901
36,958
111,962
110,898
102,886
120,011
123,337
21,266
Fuel
Delivery51
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
PFI
TGDI
TGDI
GDI
GDI
Used in
DOEV4
Program?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
Yes
No
Notes:
a PFI is Naturally aspirated, port fuel injected; GDI is Naturally aspirated, gasoline direct
injection; TGDI is Turbocharged, gasoline direct injection
Manufacturer's developmental vehicle. Vehicle used a spray-guided GDI fuel system
with a centrally-mounted injector. Emissions were targeted at Tier 2 Bin 5 or better.
1.5.1.1 PM Emissions Test Results
The results of exhaust emissions tests conducted in this program are summarized in Table
1-9 and Table 1-10 for the FTP and US06 test cycles, respectively. FTP NMOG +NOX and PM
results are also plotted in Figure 1-7 and Figure 1-8, respectively. The US06 NMOG+NOx and
PM results are shown in Figure 1-9 and Figure 1-10, respectively. In all figures, the vehicles are
divided into two groups: PFI and GDI. Within each group they are listed in the sequence of
increasing CO2 emissions on the FTP test cycle. The bars shown in the figures represent the
means of triplicate measurements. The individual data points are indicated in all figures together
with the corresponding standard deviations. Vehicle Q only had one valid PM test on the FTP
test cycle and no error bars are plotted.
1-46
-------
Table 1-9: FTP Composite Emissions Results
Vehicle
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
C02
g/mile
284.6
286.3
324.4
364.8
410.8
419.2
447.2
462.9
495.7
554.8
332.5
568.8
365.2
411.0
505.2
547.0
465.0
CO
g/mile
0.358
0.434
0.382
6.740
0.571
0.271
0.626
1.617
0.719
1.072
0.202
2.264
0.346
0.735
0.599
0.649
0.475
NOX
g/mile
0.0239
0.0461
0.0231
0.1432
0.0600
0.0151
0.0424
0.0507
0.0317
0.0281
0.0165
0.1024
0.0342
0.0279
0.0173
0.3578
0.0279
NMOG
g/mile
0.0316
0.0408
0.0299
0.0663
0.0359
0.0206
0.0439
0.0493
0.0429
0.0525
0.0171
0.0822
0.0261
0.0258
0.0399
0.0429
0.0221
NOx+NMOG
g/mile
0.056
0.087
0.053
0.210
0.096
0.036
0.086
0.100
0.075
0.081
0.034
0.185
0.060
0.054
0.057
0.401
0.050
PM
mg/mile
0.27
0.22
0.18
0.45
0.14
0.11
0.36
0.40
1.36
0.10
0.93
0.39
_
2.55
4.72
0.18
7.15
1-47
-------
Table 1-10: US06 Emissions Results
Vehicle
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
CO2
g/mile
289.0
312.8
318.2
413.7
393.3
422.8
490.9
467.0
516.0
555.9
320.4
595.6
352.8
401.7
547.4
529.1
436.6
CO
g/mile
7.092
9.315
1.293
9.077
0.660
1.237
3.462
1.128
0.833
3.015
1.800
5.519
9.225
0.330
9.862
2.728
2.595
NOX
g/mile
0.0212
0.0530
0.0257
0.1324
0.1019
0.0274
0.0369
0.0910
0.1852
0.1121
0.0247
0.0036
0.0481
0.1614
0.0377
0.1427
0.0265
NMOG
g/mile
0.0162
0.0248
0.0105
0.0127
0.0163
0.0124
0.0172
0.0134
0.0037
0.0159
0.0079
0.0125
0.0297
0.0048
0.0282
0.0116
0.0204
NOx+NMOG
g/mile
0.0374
0.0779
0.0362
0.1451
0.1183
0.0398
0.0540
0.1044
0.1889
0.1280
0.0326
0.0160
0.0779
0.1662
0.0659
0.1543
0.0470
PM
mg/mile
0.76
2.05
1.05
_
0.46
1.61
_
2.04
3.31
0.27
2.84
2.13
_
2.37
_
1.83
_
As shown in Figure 1-6, with the exception of one PFI passenger car (vehicle D), the FTP
NMOG+NOx emissions of all tested vehicles remained below their respective fleet average 2017
standards, but none performed below the 2025 standard.
The FTP PM from PFI vehicles remained well below the Tier 3 PM standard of 3
mg/mile, confirming that most current light duty vehicles are already capable of meeting the Tier
3 PM standard (Figure 1-7). Two GDI vehicles demonstrated FTP PM emissions above final
Tier 3 standard, indicating that additional combustion system development would be necessary
in some vehicles to achieve compliance.
As shown in Figure 1-8, with the exception of two LDTs (vehicles I and J), all vehicles
met their respective fleet average 2017 (for vehicles below 6,000pounds GVWR) or 2018 (for
1-48
-------
vehicles above 6,000 pounds GVWR) US06 NMOG+NOX standards. Five vehicles, four
passenger cars (vehicles A, B, F and L) and one LDT (vehicle L), produced US06 NMOG+NOx
emissions lower than the 2025 standard.
As in the case of FTP results, all PFI passenger cars remained well below the proposed
US06 10 mg/mile standard (Figure 1-9). One GDI passenger car (vehicle N) performed well
below its respective US06 PM standard and achieved PM emissions over the US06 comparable
to its performance over the FTP. In summary, all of the vehicles tested met the Tier 3 US06 PM
standards.
The suspected high oil consumption vehicle (vehicle K) emitted 3 and 2.3 times more PM
in this program than a comparable vehicle with average oil consumption (vehicle C) in the FTP
and US06 tests, respectively.
PFI (PC)
PFI (LPT2)
PFI (LDT3/4)
GDI
0.5 -
CD
'Ł0.4
o
0.3 -
r-
-z.
Q.0.1 -
LJ_
o.o -
PC
LDT2
LDT3/4
PC PC PC LDT3
I I I I I I I III
ABCKDEF GHI
P J L
i i i i
M N Q O
Figure 1-6: Composite FTP NMOG+NOx Emissions Results
1-49
-------
PFI (PC)
PFI (LPT2)
PFI (LDT3/4)
GDI
PC
LDT2
LDT3/4
O)
•4 -
Q.
Q_
0 -
1
PC PC LDT3
i i i i i i i
A B C K D E F
GHI
PJL NQO
Figure 1-7: Composite FTP PM Emission Results
PFI (PC)
PFI (LPT2)
PFI (LDT3/4)
GDI
,
0.5 -
.BO. 4 -
O)
O0.3 -
X0.2 -\
O
§0.1 -\
o.o -
PC
LDT2
LDT3/4
PC PC PC LDT3
i i i i i i iii
BCKDEF GHI
PJL
i i i i
MNQO
Figure 1-8 US06 NMOG+NOX Emissions Results
1-50
-------
PFI (PC)
PFI (LPT2)
PFI (LDT3/4)
GDI
E
E
^4H
Q.
CD
O
C/)2 -
0 -
PC
LDT2
LDT3/4
PC
i i i
B C K
H I
Figure 1-9: US06 PM Emissions Results
1.5.2 FTP PM Feasibility
With regard to the feasibility of the light-duty fleet to meet the Tier 3 PM standards over
the FTP and US06, we based our conclusions on the PM performance of the existing fleet. Data
on both low and high mileage light-duty vehicles demonstrate that the majority of vehicles are
currently achieving levels in the range of the Tier 3 FTP standards. A small number of vehicles
are at or just over the finalized Tier 3 standard at low mileage and would require calibration
changes, catalyst changes and/or further combustion system improvements to meet the new
standards. It is our expectation that the same calibration and catalyst changes required to address
NMOG would also provide some additional PM control. Vehicles that are currently
demonstrating higher PM emissions over the FTP at higher mileages would likely be required to
control oil consumption and combustion chamber deposits.
1.5.3 SFTPPM
Also, US06 test data shows that many vehicles are already at or below the Tier 3
standards for US06. Vehicles that are demonstrating high PM on the US06 would need to
control enrichment and oil consumption. The oil consumption strategies are much like that
described above for controlling oil consumption on the FTP. However, given the higher engine
RPMs experienced on the US06 and the commensurate increase in oil consumption,
manufacturers will most likely focus on oil sources stemming from the piston to cylinder
interface and positive crankcase ventilation (PCV). With respect to enrichment, changing the
fuel/air mixture by increasing the fuel fraction is no longer the only tool that manufacturers have
available to them to protect engine and exhaust system components from over-temperature
conditions. With application of electronic throttle controls on nearly every light-duty vehicle,
1-51
-------
the manufacturer has the option to richen the air/fuel mixture by maintaining the amount of fuel
being delivered and closing the throttle plate. Previously, on manual throttle control vehicles,
the throttle plate position was established by the driver and the engine controls had no capability
to change the amount of air in the intake. While it is possible that this solution may result in a
small reduction in vehicle performance we believe that it is an effective way to reduce PM
emissions over the US06 cycle.
1.5.4 Full Useful Life: Durability and Oil Consumption
Manufacturers have informed us that they have already or are planning to reduce oil
consumption by improved sealing of the paths of oil into the combustion chamber, including
improved piston-to-cylinder interfaces. They are taking or considering these actions to address
issues of customer satisfaction, cost of ownership and improved emission control system
performance as vehicles age.
Over the past decade, many manufacturers have extended oil change intervals from the
historically required 3,000 miles interval to a now typical 10,000 mile interval or more in some
vehicle models. In order to allow for these longer intervals, improvements were made to limit
pathways for oil to enter the combustion chamber, resulting in significantly reduced oil
consumption. While customer satisfaction and longer oil change intervals, particularly for leased
vehicles where owners may be less inclined to perform frequent oil changes, were a motivation
for reducing oil consumption, improvements in the performance of the emission control system
are a secondary benefit of reduced oil consumption. Oil consumption can damage catalytic
converters by coating the areas of the catalyst that convert and oxidize the pollutants. Over time,
this can cause permanent inactivity of those areas, resulting in reduced catalytic conversion
efficiency. Reductions in oil consumption can extend the life of the catalytic converter and help
manufacturers meet longer useful life requirements. This is particularly important on vehicles
meeting the most stringent emission standards, because they will need to maintain high catalyst
efficiencies in order to meet the stringent emission standards at higher mileage.
1.6 Evaporative Emissions Feasibility
The basic technology for controlling evaporative emissions was first introduced in the
1970s. Manufacturers routed fuel tank and carburetor vapors to a canister filled with activated
carbon, where vapors were stored until engine operation allowed for purge air to be drawn
through the canister to extract the vapors for delivery to the engine intake. Over the past 30
years, evaporative emission standards and test procedures have changed several times, most
notably in the mid-1990s when enhanced evaporative controls were required to address 2- and 3-
day diurnal emissions and running losses. Refueling emission controls were added with phase-in
beginning in the 1998 MY. Almost universally manufacturers elected to integrate evaporative
and refueling emission control systems. In the mid-2000s more stringent evaporative emission
standards with E10 durability gasoline led to the development and adoption of technology to
identify and eliminate permeation of fuel through fuel tanks, fuel lines, and other fuel-system
components.
The Tier 3 evaporative emission requirements include more stringent hot-soak plus
diurnal standards that are expected to require new vehicle hardware and improved fuel system
1-52
-------
designs. The type of new hardware that will be required will vary depending on the specific
application and emission challenges and are described in the following section. Additionally, the
fleet-average nature of the standards would allow more challenged vehicles to be offset by
vehicles that could outperform the required fleet averages.
In order to assess the technical feasibility of the evaporative emission standards, EPA
conducted three analyses. The first analysis was a review of technology in the fleet as well as a
literature review of the technologies used to meet PZEV "zero evap" requirements in California.
The second analysis performed was based on the certification results for the current EPA-
certified evaporative families. This provided a baseline for the current fleet emissions
performance. The third analysis looked at the list of PZEV-certified vehicles in the California
LEV II/ZEV fleets. The Tier 3 evaporative emission standards are similar to the current
evaporative requirements for PZEVs in California. These analyses are described in greater detail
below.
1.6.1 Tier 3 Evaporative Emissions/Leak Control Technology Approaches
Vehicles designed to meet the Tier 3 evaporative emission standards for the full useful
life will incorporate a variety of technologies. The choice of technologies will be based on three
key elements. The first is related to hardware and designs focused on reducing emissions to
achieve "zero" fuel vapor emissions. While the emission standards are not numerically zero, the
2 and 3-day hot soak plus diurnal standards are intended primarily to allow for the non-fuel
hydrocarbons which arise from the vehicle and its interior components. The push for "zero" fuel
vapor emissions is emphasized by the inclusion of the canister bleed emission standard which is
less than 10 percent of the hot soak plus diurnal standard. Thus, we expect the technology to
focus on the largest remaining sources of emissions. The second element is related to full-life
durability. Maintaining "zero" evaporative emission levels over many years and many miles of
driving will require a focus on preventing the deterioration in fuel/evaporative control system
performance which arises from factors such as vibration, environmental conditions, and fuel
effects. The new leak standard and the related OBD evaporative control system leak monitoring
requirement are intended to get focus on elements of technology and design which will reduce
the impacts of these factors on in-use emissions by encouraging the use of more durable
integrated technologies and systems. The third element is related to fuel effects. While EPA has
kept the RVP of the fuel at 9 psi, the Tier 3 certification fuel includes 10 percent ethanol which
will have to be further considered in choices of fuel system materials and vapor lines. EPA does
not expect the change in certification fuel to affect refueling, spit back, or running loss
compliance technology or strategies.
While the three elements discussed above are important considerations in the evaporative
emission control system design, there are two other factors which come into play when
considering which technologies will come in to the fleet and on which vehicles. First, in many
cases a given technology will provide emission reduction benefits against more than one
emission standard. For example, improved activated carbon canister technology to meet the
canister bleed standard will help to meet the hot soak plus diurnal standard or reducing
fuel/evaporative system connection points to meet the leak standard will help to meet the hot
soak plus diurnal standard. Second, to varying degrees, the technologies discussed below are in
use in the fleet today, resulting in reduced emissions relative to the current requirements for
1-53
-------
evaporative emission standards and improved durability. Thus, not every vehicle model will use
every technology either because it is already being implemented or the required reductions to
meet the Tier 3 emission targets are not large enough to warrant its application.
In preparing this assessment EPA considered two key sources of information. The first
was the developmental studies in the literature to identify the technologies best capable of
reducing emissions.79'80 Second, we examined the technologies used on various PZEV zero evap
vehicles certified in the CARB ZEV program. The technologies identified as a result of our
review are summarized below first for technologies expected to see widespread use and then for
technologies with a more limited application because they are in common use today.
1.6.1.1 Technologies expected to see widespread use
Engine/fuel system conversion: As projected in our final rule RIA for the 2017-2025 GHG
emissions, EPA projects a significant movement from port fuel injection (PFI) engines to
gasoline direct injection (GDI) engines. This ranges from 60-100 percent of products for all
categories except gasoline-powered trucks over 14,000 Ibs GVWR. This reduces air induction
systems emissions by 90 percent since the GDI uses a different fuel injection timing strategy
than the PFI.81
Air Induction System (AIS) Scrubber: For vehicles/engine models not converted to GDI, EPA
projects the use of an AIS scrubber as is now used on some PZEV models. These would reduce
R9 R'?
air induction system emissions by about90 percent. '
Canister honeycomb: This is a lower gasoline working capacity activated carbon device designed
to load and purge very easily and quickly. This device reduces canister bleed emissions by 90
percent but also provides control for the hot soak plus diurnal test. It comes in various sizes and
can be optimized based on the anticipated bleed emission rate.
Fuel system architecture: This includes reducing connections and improve seals and o-rings and
moving parts into the fuel tank: Vapor leaks from connections and the emission rates from these
leaks is exacerbated if poor sealing techniques or low grade seal materials are use in connectors
such as o-rings. Reducing connections in the fuel and evaporative systems and improving
techniques and materials could reduce these emissions by 90 percent. This would reduce hot
soak plus diurnal emissions, improve durability, and help to assure compliance with the leak
standard. Another means to reduce leak-related vapor emissions is to move fuel evaporative
system parts which are external to the fuel tank to the inside. Emissions from these parts would
be completely eliminated. This would reduce hot soak plus diurnal emissions, improve
durability, and help to assure compliance with the leak standard.
OBD evaporative system leak monitoring: Beginning in the 2017 model year, the OBD system
will need to be able to find, confirm, and signal a leak in the evaporative system of 0.020"
cumulative diameter or greater. This is done on most vehicles today as a part of meeting CARB
requirements, but will be mandatory under EPA regulations.
The evaporative emission standards discussed above also apply to gaseous-fueled
vehicles. EPA expects manufacturers to comply through the use of good design practices such as
1-54
-------
those specified in published consensus standards to address issues such as leaks from micro-
cracks and system connections and in a broader sense system mechanical and structural integrity.
1.6.1.2 Technologies expected to be optimized if necessary to achieve further
reductions:
Upgrade canister and optimize purge: This strategy is mostly available for higher powered V-8
engines. A canister with greater working capacity and/or more purge air volume could reduce
hot soak plus diurnal emissions by 80-90 percent and create capacity for the hot soak plus diurnal
and canister bleed emissions. However, it should be noted that the available emission reductions
are not large, because of the amount of purge needed to accomplish these reductions relative to
the reductions achieved from current canisters and purge strategies.84'85
Fuel tank and fuel line upgrades: Fuel vapor permeation contributes to hot soak plus diurnal
emissions. There are upgrades to fuel line materials which could reduce emissions in models
where these best practices are not yet used. In these situations, current permeation emission rates
could be reduced by 90 percent.
Improve fuel tank barrier layer and seam weld manufacturing: Fuel tanks are designed to limit
permeation emissions. Fuel tanks are typically made of high-density polyethylene with an
embedded barrier layer of ethyl vinyl alcohol (EvOH) representing about 1.8 percent of the
average wall thickness for reducing permeation emissions. In some cases manufacturers could
increase the EvOH barrier thickness to about 3 percent of the average wall thickness to provide a
more uniform barrier layer, to provide better protection with ethanol-based fuels, and to improve
permeation resistance generally. Recent developments in production processes have led to
improved barrier coverage around the ends of the tank where the molded plastic is pinch-welded
to form a closed vessel. This technology would likely be coupled with the increase in EvOH in
the overall tank material or other techniques to reduce permeation from these seams. These
changes are expected to decrease emission rates over the diurnal test from about 40 mg per day
to 15 mg per day from the fuel tank assembly. It is likely that this change would be done as part
of a fuel tank design changeover and not out of a normal tooling cycle.
Upgrade fill tube material and connection to fuel tank: The connection of the fill tube to the fuel
tank is the largest connection in the fuel system. Improving the security of the fill tube
connection to the tank could reduce vapor leaks. The fill tube itself has a larger diameter than
other fuel or vapor lines and thus has a relatively large diameter. For higher permeation
resistance the tube can be upgraded to one having an FKM inner layer. We would expect such
changes to occur together.
Table 1-11 presents a summary of EPA projections of the application of the widespread
technologies across the LDV, LDT, MDPV, and HDGV fleets. These projections are based on
consideration of the most effective technologies to achieve the required reductions. In this
context, effective means not only what technologies might provide the largest reductions which
could be used to meet more than one standard, but also which technologies provide these
reductions in the most cost efficient way. The baseline emission rates and percent efficiencies
from the various technologies are both EPA estimates based on review of the literature and
discussion with various manufacturers and vendors. The reductions achieved are larger than the
1-55
-------
difference between the baseline and design target because of the need to cover for the non-fuel
hydrocarbons which are measured in certification but decay in use. Note that the last column in
Table 1-11 identifies which standards the technology will address. In some cases the expected
emission reductions are larger from HDGVs. This is due primarily to the volume of their fuel
tanks and unique aspects of some elements of their fuels systems relative to smaller passenger
cars and light trucks.
1-56
-------
Table 1-11 Application of Evaporative Emission Control Technologies for Tier 3
Vehicle Class
nonfuel (g)
M5ATstd(3d/2d)(g)
T3 std (g)
Canister bleed std (g)
M5AT Baseline (g) *
T3 Target (g)2
Red Needed (g)
Red Achieved (g)
LDV
0.1
0.5/0.65
0.3
0.02
0.47
0.2
0.27
0.39
0.1
0.65/0.85
0.3
0.02
0.47
0.2
0.27
0.39
0.125
0.65/0.85
0.4
0.02
0.42
0.3
0.12
0.39
0.15
0.9/1.15
0.5
0.02
0.72
0.37
0.35
0.39
0.15
0.9/1.15
0.5
0.02
0.72
0.37
0.35
0.39
0.175
1.0/1.25
0.5
0.02
0.72
0.37
0.35
0.39
LHDGV
0.2
1.4/1.75
0.6
0.03
0.96
0.45
0.51
0.55
HHDGV
0.25
1.9/2.3
0.6
0.03
0.96
0.45
0.51
0.52
Technology
Canister
honeycomb
35mmx75mm
(V-4&V-6)
Canister
honeycomb
35mmx50mm
(V-8)
AIS scrubber
PFItoGDI5
Fuel system
(a)reduce
connections &
improve seals/ o-
rings 6
(b) move parts
into tank
OBD software
upgrades7
Current
Emissions
150-200mg3
(use 150mg)
150- 225 mg
(use 150mg
excHDGV)
150-300 mg
(use 225 mg)
excHDGV4
150-300 mg
(use 225mg)
exc HDGV
25- 50m g
(use 25mg)
exc HDGV
75- 100m g
(use75mg)
n/a
Control
Efficiency
90%
90%
90%
90%
90%
100%
n/a
mg red & net red
% applied mg
135_100% 135
0 0
200_40% 80
200_60% 120
22_100% 22
75_50% 37
100%
mg red & net red
%applied mg
135_100% 135
0 0
200_27% 54
200_73% 146
22_100% 22
75_50% 37
100%
mg red & net red
% applied mg
135_76% 103
135_24% 32
200_27% 54
200_73% 146
22_100% 22
75_50% 37
100%
mg red & net red
% applied mg
135_10% 13
135_90% 122
200_27% 54
200_73% 146
22_100% 22
75_50% 37
100%
mg red & net red
%applied mg
0 0
135_100% 135
200_27% 54
200_73% 146
22_100% 22
75_50% 37
100%
mg red & net red
% applied mg
0 0
135_100% 135
200_27% 54
200_73% 146
22_100% 22
75_50% 37
100%
mg red & net red
% applied mg
0 0
200_100% 200
0 0
270_100% 270
45_100% 45
75_50% 37
100%
mg red & net red
%applied mg
0 0
200-100% 200
270_100% 270
0 0
45_100% 45
-
-
Standard
Addressed
T3/ bleed
T3/ bleed
T3
T3
T3/leak
T3/leak
leak
1 based on mean plus one standard deviation for 2013 MY 2-day cert results on Tier 2 fuel
2100 mg or 25% below T3 std whichever is greater
2 (365 day per year)(l gal per 5.6 lbs)(l Ib per 454g)(mg reduction/day)(1 g/1000 mg)(0.9 energy density effect); needs to be further multiplied by(15yr) (avg surv fraction for fleet)(gas price)
3 based SAE 2001-01-0733
4 based on SAE 2005-01-0113 and US patent 6464761
5 % conversion from PFItoGDI based on RIA for 2017-2025 EPA GHG Final Rule
6 reduce fuel/ evap system connections, improve seal material (FKM) in engine & fuel/ evap systems, and employ o-rings as needed
7 Most manufacturers meet the 0.020" evaporative system leak monitoring provision now; no new hardware expected
1-57
-------
This page intentionally left blank.
1-58
-------
Table 1-12 presents the information for the technologies expected to see more limited
use, but does not project application rates. It does show however, that these technologies would
in most cases provide smaller reductions than those identified in Table 1-11. In the case of the
upgraded canister and purge optimization, to the degree it is used it is more likely to be a
replacement for the canister honeycomb on V-8 engines than as an additional technology by
itself.
Table 1-12 Technologies Which May Be Optimized If Necessary to Achieve Further
Reductions in Tier 3
Upgrade
canister &
improve
purge (mostly
V-8s)
Improve fuel
tank barrier
layer
thickness and
reduce pinch
seam gaps
Filler neck
connection
and materials
ugrade1
120-150mg
80-90% ~100mg
75mg
70%
"50mg
60mg
80%
"50mg
40m g
90%
"35mg
Fuel line
material
upgrade
1 Kawasaki.M., et al, Low Gasoline Permeable Fuel
Filler Hose, SAETechnical Pa per Series 971080, 1997.
1.6.2 Assessment of the Current EPA Certification Emissions
EPA's current evaporative emission standards vary by vehicle category. Table 1-13
shows the currently applicable hot-soak plus diurnal emission standards and Table 1-14 shows
the Tier 3 standards.
1-59
-------
Table 1-13 Existing EPA Evaporative Emission Standards
Vehicle Category
LDV
LDT1/LDT2
LDT3/LDT4
MDPV
HDGV<14,0001bs
GVWR
HDGV > 14,000 Ibs
GVWR
Hot-soak plus Diurnal (2-day)
0.65 g/test
0.85 g/test
1.1 5 g/test
1.25 g/test
1.75 g/test
2.3 g/test
Hot-soak plus Diurnal (3-day)
0.50 g/test
0.65 g/test
0.90 g/test
1.00 g/test
1.4 g/test
1.9 g/test
Table 1-14 Final Tier 3 Evaporative Emission Standards
Vehicle Category/ Averaging Sets
LDV, LDT1
LDT2
LDT3, LDT4, MDPV
HDGVs
Highest Hot Soak + Diurnal Level
(over both 2-day and 3 -day diurnal tests)
0.300 g/test
0.400 g/test
0.5 00 g/test
0.600 g/test
Based on MY2013 certification data, EPA analyzed the certification hot-soak plus diurnal
emission levels for all vehicle categories that will be subject to the Tier 3 standards. The
following figure shows the hot-soak plus diurnal certification levels (based on the 2-day diurnal
test) for each vehicle category ordered from the lowest to the highest emission levels. (While not
presented in this analysis, the data based on the 3-day diurnal tests shows a similar trend.)
Figure 1-10 also shows the existing and Tier 3 evaporative emission standards.
1-60
-------
2.5
0.5
HDGV
.HDGV
>GV
•Cert. Levels
•Tier 3 Stds
Existing Stds
Families Certified
Figure 1-10: MY2013 PZEV & Federal Hot-soak Plus Diurnal (2-Day) Emission
Certification Levels6
It should be noted that the current evaporative emission data is based on a different
certification test fuel than what is being implemented for the Tier 3 program. While both the
current and Tier 3 certification fuels have a Reid vapor pressure of nominally 9.0 psi, EPA's
current certification test fuel contains no ethanol, whereas the Tier 3 certification fuel contains
10 percent ethanol. Nevertheless, EPA believes this information is still useful in gaging the level
of effort needed by manufacturers to comply with the Tier 3 standards. It is generally
understood that ethanol can impact permeation emissions from the fuel tank and fuel lines to
some degree, but the bulk of evaporative emissions are from diurnal emissions which are
primarily a function of the Reid vapor pressure of the fuel which will be maintained at 9.0 psi
and therefore should not be impacted by the presence of ethanol in the certification fuel.
As can be seen from the figure, there are many families certified below the Tier 3 hot-
soak plus diurnal standards. Of the nearly 450 evaporative families included in the analysis, 40
percent had certification levels below the Tier 3 standards. Some of these families (-50) are
certified to the more stringent PZEV standards, upon which the Tier 3 evaporative emission
standards are based, but most of the families are not. However, the Tier 3 evaporative emission
standards include a new canister bleed test that is not required under the current EPA regulations.
(The families certified to the PZEV requirements are subject to a similar requirement and would
likely meet that new canister bleed test requirement and longer useful life period without further
modification.) Therefore, even though many families are certified below the Tier 3 evaporative
emission standards, manufacturers would still need to make additional changes with many of the
evaporative control systems to ensure compliance with the standards. We expect that
G Note that LHDGVs are vehicles rated 8,501-14,000 Ibs GVWR; HHDGVs are vehicles rated greater than 14,000
Ibs GVWR.
1-61
-------
manufacturers would use the technologies discussed above and use allowances and ABT to
minimize costs and assist in an orderly phase-in of compliant products. In 2013, the baseline
certification values used for our analysis were 0.41 g HC/test for LDVs, LDTl/2s, 0.5 for
LDT3/4 and MDPVs, and 0.63 for HDGVs.
1.6.3 Assessment of California-certified PZEVs
Based on the California Air Resources Board's MY2013 certification list, EPA identified
the vehicles certified by manufacturers to the PZEV requirements. As noted earlier, the Tier 3
evaporative emission standards are very similar to the PZEV evaporative emission requirements
and, as allowed with one of the options for MY2017, manufacturers could sell their evaporative
emission compliant PZEV vehicles nationwide in MY2017. Manufacturers have certified over
50 models of passenger cars and light-duty trucks to the PZEV requirements. EPA believes that
manufacturer's experience with PZEV technologies will assist them as they work to apply
similar technologies across their fleets to comply with the Tier 3 evaporative emission standards.
As described in more detail above, EPA expects manufacturers will employ a number of
technologies to meet the Tier 3 standards. The anticipated control technologies to comply with
the emission standards have already been included on many of the PZEVs. Table 1-15 shows the
12 manufacturers and over 50 models certified to the PZEV standards in MY2013. Two other
manufacturers certified PZEVs in previous model years as well.
1-62
-------
Table 1-15 List of MY2013 Models Certified to CARB's PZEV Requirements
Manufacturer
AUDI/VOLKSWAGEN
BMW
CHRYSLER
FORD
GENERAL MOTORS
HONDA
HYUNDAI
KIA
MAZDA
MERCEDES
SUBARU
TOYOTA
Models
Audi A3, Volkswagen GTI, Volkswagen Jetta, Volkswagen Golf,
Volkswagen Jetta Sportwagen, Volkswagen Jetta, Volkswagen
Jetta GLI, Volkswagen GTI, Volkswagen CC, Volkswagen Beetle,
Volkswagen Passsat
BMW 1281, BMW 3281, BMW 328C1
Chrysler 200, Dodge Avenger
Ford Escape Hybrid, C-MAX Hybrid,
Ford Focus, Ford Fusion Hybrid
Chevy Volt, Chevy Malibu Hybrid, Chevy Cruze, Chevy Sonic,
Buick LaCrosse, Buick Regal, Cadillac ATS, Chevy Equinox,
Chevy Impala, GMC Terrain
Honda Civic GX (CNG), Honda Civic Hybrid,
Honda CR-Z Hybrid, Honda Insight, Honda Insight Hybrid, Honda
Accord
Elantra, Tucson, Sonata, Sonata Hybrid
Kia Spoilage, Kia Forte, Kia Forte ECO, Kia Forte KOUP,
Kia Optima Hybrid
Mazda 3
Mercedes S400 Hybrid, Mercedes C300/
C350/E350/GLK350/E400 Hybrid
Subaru Legacy, Subaru Outback Wagon, Subaru Forester , Subaru
Impreza, Subaru XV Crosstrek
Toyota Prius, Toyota Camry, Toyota Camry Hybrid, Toyota Prius
Hybrid
1.7 ORVR for Complete HDGVs over 10,000 Ibs GVWR
This final rule includes onboard refueling vapor recovery (ORVR) requirements
for complete HDGVs over 10,000 Ibs but equal to or less than 14,000 Ibs GVWR beginning in
the 2018 model year. Due to the similarity of the vehicle chassis and fuel systems and the
commonality of chassis production lines, manufacturers have all implemented ORVR hardware
on complete Class 3 HDGVs (10,001-14,000 Ibs GVWR) since the 2006 MY when the ORVR
phase-in covering Class 2b vehicles (8,501- 10,000 Ibs GVWR) ended. Today, about XX percent
of Class 3 vehicles are incomplete chassis. EPA is including this requirement in Tier 3 to ensure
no backsliding and to give states the opportunity to claim the ORVR reductions for Tier 3
vehicles in their SIPs. This is especially important to states removing Stage II vapor recovery.
Furthermore, EPA is including ORVR requirements for any complete HDGVs over 14,000 Ibs
GVWR effective in the 2022 model year when the Tier 3 evaporative emissions phase-in ends.
H See http://driveclean.ca.gov/searchresults by smog.php?smog slider value=9&x=12&y=12. downloaded on
December 6, 2013.
1-63
-------
While sales of HDGVs over 14,000 Ibs are relatively low, today all would be classified as
incomplete chassis for the purposes of the ORVR requirement.
EPA is not extending the ORVR requirement to incomplete HDGV chassis in this
rule. There is no question of the basic technical feasibility of the requirement. However,
manufacturers have stated that ORVR for incomplete HDGVs presents some system design and
integration issues with regard to the larger fuel tanks and vehicles with two tanks but more
importantly with regard to the activities of secondary manufacturers. Close coordination and
installation instructions are needed to ensure that integrated ORVR/evaporative control systems
are installed in an effective and safe manner. EPA estimates that incomplete HDGV sales are
about (85,000) per year, but with their low fuel economy (-15 mpg) control of refueling
emissions through ORVR may become important as Stage II vapor recovery is removed in ozone
nonattainment areas.
The ORVR requirements discussed above also apply to gaseous-fueled vehicles. EPA
expects manufacturers to comply through the use of good design practices such as those
specified in published consensus standards to address issues such as refueling connections and
system integrity.
1.8 Onboard Diagnostics for Vehicles less than 14,000 Ibs GVWR
As part of the Tier 3 final rule, EPA is incorporating by reference the July 31, 2013
version of the California ARB OBD II regulations for vehicles equal to or less than 14,000 Ibs
GVWR. These requirements apply in the 2017 MY, at least two years after they must be met in
California. As permitted in EPA regulations, manufacturers generally receive an Executive Order
for OBD compliance from the CARB for each test group and EPA will accept that Executive
Order as evidence that the vehicles covered by the test group meet CARB requirements and
therefore meet the identical EPA requirements. Thus, in the case of the 2017 model year
requirements for EPA, we expect manufacturers will already comply with these requirements
before 2017 for their LEV III vehicles and have Executive Orders available.
EPA is adding two requirements related to the leak standard. The first is a requirement
that manufacturers demonstrate before production that their vehicle test groups' OBD-based
evaporative system monitor can detect the presence of a leak with an effective leak diameter at or
above 0.020 inches, illuminate the MIL, and store the appropriate a confirmed diagnostic trouble
codes. Such activity is normally done as part of the evaporative system leak monitor
development and is demonstrated in the Production Vehicle Evaluation Testing program
prescribed in 13 CCR 1968.2(j). However, if the OBD-based evaporative system leak monitor is
to be used in IUVP, its performance needs to be certified before production begins instead of
afterwards. Since this requirement is compatible with CARBs current regulations for OBD-based
evaporative system leak-based monitoring and the Production Vehicle Evaluation Testing
program and is phasing in with the leak standard, there should be no feasibility or lead time
issues.
EPA is also implementing a requirement that OBD systems revise the software so that a
scan readable record is created which indicates if the OBD-based evaporative system leak
monitor has run within the previous 750 miles and if so what was the result. The means by
1-64
-------
which this record will be created and stored is being done in a manner compatible with SAE
J1979 as suggested by the commenters. Since this requirement is phasing in with the leak
standard, there should be no feasibility or lead time issues.
1-65
-------
References
1 Beck, D.D., Sommers, J. W., DiMaggio, C.L. (1994). Impact of sulfur on model palladium-only catalysts under
simulated three-way operation. Applied Catalysis B: Environmental 3, 205-227.
2 Beck, D.D., Sommers, J.W. (1995). Impact of sulfur on the performance of vehicle aged palladium monoliths."
Applied Catalysis B: Environmental 6, 185-200.
3 Beck, D.D., Sommers, J.W., DiMaggio, C.I. (1997). Axial characterization of oxygen storage capacity inclose
coupled lightoff and underfloor catalytic converters and impact of sulfur. Applied Catalysis B: Environmental 11,
273-290.
4 Waqif, M, Bazin, P., Saur, O. Lavalley, J.C., Blanchard, G., Touret, O. (1997), Study of ceria sulfation. Applied
CatalysisB: Environmental 11, 193-205.
5 Bazin, P., Saur, O. Lavalley, J.C., Blanchard, G., Visciglio, V., Touret, O. (1997). "Influence of platinum on ceria
sulfation." Applied Catalysis B: Environmental 13, 265-274.
6 Takei, Y., Kungasa, Y., Okada, M., Tanaka, T. Fujimoto, Y. (2000). Fuel Property Requirement for Advanced
Technology Engines. SAE Technical Paper 2000-01-2019.
7 Takei, Y., Kungasa, Y., Okada, M., Tanaka, T. Fujimoto, Y. (2001). "Fuel properties for advanced engines."
Automotive Engineering International 109 12, 117-120.
8 Kubsh, J.E., Anthony, J.W. (2007). The Potential for Achieving Low Hydrocarbon and NOX Exhaust Emissions
from Large Light-Duty Gasoline Vehicles. SAE Technical Paper 2007-01-1261.
9 Shen, Y., Shuai, S., Wang, J. Xiao, J. (2008). Effects of Gasoline Fuel Properties on Engine Performance. SAE
Technical Paper 2008-01-0628.
10 Ball, D., Clark, D., Moser, D. (2011). Effects of Fuel Sulfur on FTP NOX Emissions from a PZEV 4 Cylinder
Application. SAE Technical Paper 2011-01-0300.
11 Heck, R.M., Farrauto, R.J. (2002). Chapter 5: Catalyst Deactivation in Catalytic Air Pollution Control, 2nd
Edition. John Wiley and Sons, Inc.
12 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
13 Takei, Y., Kungasa, Y., Okada, M., Tanaka, T. Fujimoto, Y. (2000). Fuel Property Requirement for Advanced
Technology Engines. SAE Technical Paper 2000-01-2019.
14 Ball, D., Clark, D., Moser, D. (2011). Effects of Fuel Sulfur onFTP NOX Emissions from a PZEV 4 Cylinder
Application. SAE Technical Paper 2011-01-0300.
15 Coordinating Research Council. 2003. "The Effect of Fuel Sulfur on NH3 and Other Emissions from 2000-2001
Model Year Vehicles." CRC Project No. E-60 Final Report. Accessed on the Internet on 12/4/2013 at the following
URL: http://www.crcao.com/reports/recentstudies2003/E-60%20Final%20Report.pdf.
16 Alliance of Automobile Manufacturers. 2001. "AAM-AIAM Industry Low Sulfur Test Program."
17 Manufacturers of Emission Controls Association. 2013. "The Impact of Gasoline Fuel Sulfur on Catalytic
Emission Control Systems."
18 American Petroleum Institute. 2013. Supplemental Comments of the American Petroleum Institute.
1-66
-------
19 Vehicles that meet the cleanest emission standards by demonstrating very low cold start NMOG and NOx
emissions and zero or approaching zero running NMOG and NOx emissions.
20 Heck, R.M., Farrauto, RJ. (2002). Chapter 5: Catalyst Deactivation in Catalytic Air Pollution Control, 2nd
Edition. John Wiley and Sons, Inc.
21 Luo, T. Gorte, RJ. (2003). A Mechanistic Study of Sulfur Poisoning of the Water-Gas-Shift Reaction Over
Pd/Ceria." Catalysis Letters, 85, Issues 3-4, pg. 139-146.
22 Li-Dun, A., Quan, D.Y. (1990). "Mechanism of sulfur poisoning of supported Pd(Pt)/Al2O3 catalysts for H2-O2
reaction." Applied Catalysis 61, Issue 1, pg. 219-234.
23 Waqif, M., Bazin, P., Saur, O. Lavalley, J.C., Blanchard, G., Touret, O. "Study of ceria sulfation." Applied
CatalysisB: Environmental 11 (1997) 193-205.
24 Bazin, P., Saur, O. Lavalley, J.C., Blanchard, G., Visciglio, V., Touret, O. "Influence of platinum on ceria
sulfation." Applied Catalysis B: Environmental 13 (1997) 265-274.
25 Heck, R.M., Farrauto, RJ. (2002). Chapter 6: Automotive Catalyst in Catalytic Air Pollution Control, 2nd Edition.
John Wiley and Sons, Inc.
26 Luo, T. Gorte, RJ. (2003) A Mechanistic Study of Sulfur Poisoning of the Water-Gas-Shift Reaction Over
Pd/Ceria. Catalysis Letters, 85, Issues 3-4, pg. 139-146.
27 Beck, D.D., Sommers, J.W. (1995) Impact of sulfur on the performance of vehicle aged palladium monoliths.
Applied CatalysisB: Environmental 6, 185-200.
28 Beck, D.D., Sommers, J.W. (1995) Impact of sulfur on the performance of vehicle aged palladium monoliths.
Applied Catalysis B: Environmental 6, 185-200.
29 Maricq, M. M. Chace, R.E., Xu, N., Podsiadlik, D.H. (2002). The Effects of the Catalytic Converter and Fuel
Sulfur Level on Motor Vehicle Paniculate Matter Emissions: Gasoline Vehicles." Environmental Science and
Technology, 36, No. 2 pg. 276-282.
30 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
31 See Preamble Section IVA.6.C and Chapter 1 of the RIA (Section 1.2.3.2) for more details on this study and its
results.
32 See §86.1811-17 (LD) within the Tier3 regulations. Tier3 restrictions to commanded enrichment are also
discussed in further detail within Section IV.A.4.C of this preamble.
33 Tier 2 Regulatory Impact Analysis, EPA 420-R-99-023, December 22, 1999, last accessed on the Internet on
12/04/2013 at the following URL: http://epa.gov/tier2
34 Chapter 6 of the Regulatory Impact Analysis for the Control of Hazardous Air Pollutants from Mobile Sources
Final Rule, EPA 420-R-07-002, February 2007, last accessed on the Internet on 12/04/2013 at the following URL:
http://nepis.epa. gov/Exe/ZvPDF.cgi?Dockev=P1004LNN.PDF.
35 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
36 Takei, Y., Kungasa, Y., Okada, M., Tanaka, T. Fujimoto, Y. (2000). Fuel Property Requirement for Advanced
Technology Engines. SAE Technical Paper 2000-01-2019.
1-67
-------
37 Ball, D., Clark, D., Moser, D. (2011). Effects of Fuel Sulfur on FTP NOX Emissions from a PZEV 4 Cylinder
Application. SAE Technical Paper 2011-01-0300.
38 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
39 Shapiro, E. (2009). National Clean Gasoline - An Investigation of Costs and Benefits. Published by the Alliance
of Automobile Manufacturers.
40 Chapter 6 of the Regulatory Impact Analysis for the Control of Hazardous Air Pollutants from Mobile Sources
Final Rule, EPA 420-R-07-002, February 2007, last accessed on the Internet on 12/04/2013 at the following URL:
http://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P1004LNN.PDF
41 Ball, D., Clark, D., Moser, D. (2011). Effects of Fuel Sulfur on FTP NOX Emissions from a PZEV 4 Cylinder
Application. SAE Technical Paper 2011-01-0300
42 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
43 The NPRM modeling was based on analysis of 81 passenger cars and trucks. Since the NPRM, twelve additional
Tier 2 vehicles were tested and included in the statistical analysis described in the docketed final report, examining
the effect of sulfur on emissions from Tier 2 vehicles. The analysis based on the complete set of 93 Tier 2 vehicles
is reflected in the results presented in this Section and the emissions modeling for FRM.
44 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
45 Peer Review of the Effects of Fuel Sulfur Level on Emissions from the In-Use Tier 2 Vehicles, EPA-HQ-OAR-
2011-0135-1847.
46 EPA In-Use Sulfur Report - Response to Peer-Review Comments, EPA-HQ-OAR-2011-0135-1848.
47 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
48 Ball, D., Clark, D., Moser, D. (2011). Effects of Fuel Sulfur on FTP NOX Emissions from a PZEV 4 Cylinder
Application. SAE Technical Paper 2011-01-0300.
49 American Petroleum Institute. 2013. Supplemental Comments of the American Petroleum Institute. Available in
the docket for this final rule, docket no. EPA-HQ-OAR-2011-0135
50 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
51 The make and model of the tested vehicles are Honda Crosstour, Chevrolet Malibu, Chevrolet Silverado, Ford
Focus and Subaru Outback.
52 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
53 Ford Motor Company. 2013. "Quality Changes Needed to Meet Tier 3 Emission Standards and Future
Greenhouse Gas Requirements." Attachment 2: "Tier 3 Sulfur Test Program - Ford Motor Company Summary
Report." Available within EPA Docket for this final rule, EPA-HQ-2011-0135.
1-68
-------
54 Dominic DiCicco, Ford Motor Company. 2013. "Additional data as requested. RE: Ford Supplemental Comments
on Tier 3." Available within EPA Docket for this final rule, EPA-HQ-2011-0135.
55 See 77 FR 62840-62862, October 15, 2012; and Joint Technical Support Document: Final Rulemaking for 2017-
2025 Light-Duty Vehicle Greenhouse Gas Emission Standards and Corporate Average Fuel Economy Standards
(EPA-420-R-12-901), August 2012, Chapter 3.4.1.7 - 3.4.1.8 (pages 3-88 - 3-95).
56 Ford Motor Company, 2012. "Sustainability 2011/2012 - Improving Fuel Economy." Accessed on the Internet on
11/21/2013 at: http://corporate.ford.com/microsites/sustainabilitv-report-2011-12/environment-products-plan-
economy. Available within EPA Docket for this final rule, EPA-HQ-2011-0135.
57 Ford used the same tert-butyl sulfide fuel sulfur additives used within the EPA testing in IV.A.6.C and d.
58Emissions at 30 ppm sulfur estimated assuming approximately linear emissions effects between 10, 26.5 and 30
ppm gasoline sulfur levels.
59 Ball, D., Clark, D., Moser, D. (2011). Effects of Fuel Sulfur onFTP NOX Emissions from a PZEV 4 Cylinder
Application. SAE Technical Paper 2011-01-0300. Available in the docket for this final rule.
60 American Petroleum Institute. 2013. Supplemental Comments of the American Petroleum Institute. Available in
the docket for this final rule, docket no. EPA-HQ-OAR-2011-0135
61 Add citation to document number and docket number for Tier 3 RIA
62 U.S. Code of Federal Regulations, Title 40, §86.1823-08 "Durability demonstration procedures for exhaust
emissions."
63 Heck, R.M., Farrauto, RJ. (2002). Chapter 6: Automotive Catalyst in Catalytic Air Pollution Control, 2nd Edition.
John Wiley and Sons, Inc.
64 Samenfink, W., Albrodt, H., Frank, M., Gesk, M., Melsheimer, A., Thurso, J., Matt, M. "Strategies to Reduce
HC-Emissions During the Cold Starting of a Port Fuel Injected Engine." SAE Technical Paper 2003-01-0627.
65 Samenfink, W., Albrodt, H., Frank, M., Gesk, M., Melsheimer, A., Thurso, J., Matt, M. "Strategies to Reduce
HC-Emissions During the Cold Starting of a Port Fuel Injected Engine." SAE Technical Paper 2003-01-0627.
66 Yi, J., Wooldridge, S., Coulson, G., Hilditch, J., Iver, C., Moilanen, P., Papaioannou, G., Reiche, D., Shelby, M.,
VanDerWege, B., Weaver, C., Xu, Z., Davis, G., Hinds, B., Schamel, A. "Development and Optimization of the
Ford 3.5L V6 EcoBoost Combustion System." SAE Technical Paper 2009-01-1494.
67 Choi, M., Sun, H., Lee, C., Myung, C., Kim, W., Choi, J. "The Study of HC Emission Characteristics and
Combustion Stability with Spark Timing Retard at Cold Start in Gasoline Engine Vehicle." SAE Technical Paper
2000-01-0182.
68 Eng, James A. "The Effect of Spark Retard on Engine-out Hydrocarbon Emissions." SAE Technical Paper 2005-
01-3867.
69 Hattori, M., Inoue, T., Mashiki, Z., Takenaka, A., Urushihata, H., Morino, S., Inohara, T. "Development of
Variable Valve Timing System Controlled by Electric Motor." SAE Technical Paper 2008-01-1358.
70 Ball, D., Zammit, M., Wuttke, J., Buitrago, C. "Investigation of LEV-HI Aftertreatment Designs." SAE Technical
Paper 2011-01-0301.
1-69
-------
71 Serrano, D., Lavy, J., Kleeman, A., Zinola, S., Dumas, I, Le Mirronet, S., Heitz, D. "Post Oxidation Study During
Secondary Exhaust Air Injection for Fast Catalyst Light-Off." SAE Technical Paper 2009-01-2706.
72 Lee, D., Heywood, J. "Effects of Secondary Air Injection During Cold Start of SI Engines." SAE Technical Paper
2010-01-2124.
73 Sano, K., Kawai, T., Yoshizaki, S., Iwamoto, Y. "HC Adsorber System for SULEVs of Large Volume
Displacement." SAE Technical Paper 2007-01-0929.
74 McDonald, J.F., Schenk, C., Sanchez, L.J., Nelson, BJ. "Testing of Catalytic Exhaust Emission Control Systems
Under Simulated Locomotive Exhaust Conditions." SAE Technical Paper No. 2011-01-1313.
75 U.S. Code of Federal Regulations, Title 40, § 86.1823-08 Durability demonstration procedures for exhaust
emissions.
76 Christiansen, Michael G. "Impact of Lubricating Oil Condition on Exhaust Paniculate Matter Emissions from
Light Duty Vehicles" SAE Technical Paper No. 2010-01-1560.
77 NIOSH Reference Method 5040 - Elemental Carbon (Diesel Paniculate Matter). NIOSH Manual of Analytical
Methods (NMAM), Fourth Edition, 2003.
78 Sobotowski, R. (February, 2013). Test Program to Establish LDV Full Useful Life PM Performance.
Memorandum to the docket.
79 Zhao, F. "Technologies for Near-Zero Gasoline-Powered Vehicles." Society of Automotive Engineers, 2007 and
"advanced Developments in Ultra-Clean Gasoline Powered Vehicles". PT-104. Society of Automotive Engineers.
2004.
80 "Evaporative Emission Control Technologies for Gasoline Powered Vehicles", Manufacturers of Emission
Controls Association, December, 2010.
81 See Chapter 2.6 of the RIA for the rule, 2017 and Later Model Year Light-Duty Vehicle Greenhouse Gas
Emissions and Corporate Average Fuel Economy Standards, .Fleet wide projections in Table 1-11 are based on
weightings of the portions of V-4, V-6, and V-8s expected to be converted to GDI.
82 Leffel,J, Abdolhosseini, R.,(2005) SAE Paper 2005-01-1104, "Requirements Setting, Optimization, and "Best Fit"
Application of AIS Hydrocarbon Adsorption Devices for Engine Evaporative Emissions Breathing Loss Control".
83 Lebowitz, J., Lovette,!, Chan,C., and Frich,D., (2005) SAE Paper 2005-01-0113, "Activated Carbon Coated
Polymeric Foam for Hydrocarbon Vapor Adsorption".
84 Williams,R. Clontz, C., (2001) SAE Paper 2001-01-0733, "Impact and Control of Canister Bleed Emissions".
85 Clontz.,R.,Elum,.M.,McRare,P.,Williams,R., (2007) SAE Paper 2007-01-1929, "Effects of Low Purge Vehicle
Applications and Ethanol Containing Fuels on Evaporative Emissions Canister Performance".
1-70
-------
Chapter 2 Vehicle Program Costs
2.1 Changes to Vehicle Costs between Proposed and Final Rules
We have made several changes to vehicle program costs since the proposal, but two
changes have significant impacts on the final rule costs and help to explain the large reduction in
cost estimates between proposal and final rule. The first of these significant changes involves
the catalyst platinum group metal (PGM) loading costs. As commenters pointed out, the cost
estimates in the proposal have become dated, as they were based largely on four-year-old
estimates of the CARB LEVIII program. For this final rule, we have developed a more robust
catalyst loading cost estimate using a methodology suggested by one commenter.1 This more
robust estimate results in lower costs than estimated in the proposal.
The second significant contributer to reduced final rule cost estimates is the use of the
MY 2017-2025 fleet mix projected to result from the most recent GHG and fuel economy rules.
That projected fleet mix shows a large percentage of four-cylinder engines, which are less costly
to modify to achieve Tier 3 compliance than the proposal's projected MY 2012-2016 fleet mix,
which included many more V-configuration six-cylinder (V6) and eight-cylinder (V8) engines.
We mentioned in the preamble to the proposal our intention to use the projected MY 2017-2025
fleet for our final rule cost analysis (see 78 FR 29970).
We have made many other updates to the analysis for this final rule. For example, we
reviewed the MY2013 certification database to evaluate the certified emissions levels of the
fleet. We found that many vehicles are already being certified with emissions levels that would
meet final Tier 3 standards. Further, many vehicles have certified emission levels that are 70%
of the 0.30 g/mi NOx+NMOG standard, meaning that sufficient compliance margin exists for
those vehicle to comply with Tier 3 without any additional costs. Our final rule estimates no
exhaust emission-related Tier 3 costs for these vehicles (they still incur evaporative emission-
related costs, discussed below).
We have also concluded that active HC adsorbers, which we projected for use on some
vehicles in the proposal, are not likely to be used. Instead, as we discuss in Section 2.3.6, those
vehicles will probably use a passive HC adsorber. The passive HC adsorber is considerably less
costly. We have also decreased our evaporative emission control costs, in part because of the
high penetration of gasoline engines with direct injection projected by the MY 2017-2025 GHG
and fuel economy rules. Direct injection removes a large source of evaporative emissions and,
thus, means fewer vehicles need to add certain evaporative control technologies. We have also
decreased the penetration rates of secondary air injection in the later years of the program, for
reasons described below. Lastly, we have modified very slightly our indirect cost markups to
account for the fact that most of the research and development efforts required of auto makers
are in response to CARB's LEVIII rule and need not be conducted again for Tier 3 compliance.
We have made some changes that have increased costs, although these are smaller than
those that have decreased costs so, on net, estimated vehicle-level costs are lower than in the
proposal. One such change was to double the engine calibration costs (from roughly $2/vehicle
to $5/vehicle), to cover expected calibration efforts associated with PM control on direct injected
2-1
-------
gasoline engines. We discuss this in more detail in Section 2.3.5, below. We have also
increased the penetration rates of the technology we term "optimized thermal management" for
some vehicle categories. We discuss our rationale for this change in Section 2.5, below. Another
change was to update all costs from 2010 dollars to 2011 dollars.A
With respect to total program costs, the significant change since proposal was to exclude
costs incurred on vehicles sold in all states (California and elsewhere) that have adopted the
California LEVIII program. As a result, our estimated costs per vehicle are applied to millions
fewer vehicles in the final rule, thus making the total program costs considerably lower. And
finally, we have included operating savings (fuel savings) associated with avoiding the loss of
fuel that would have otherwise evaporated absent the new Tier 3 controls. The otherwise
evaporated fuel is ultimately used to propel the vehicle, thus providing a savings to the
consumer. We discuss operating costs in Section 2.6, below.
2.2 General Methodology
Although the increase in standard stringency is substantial for the vehicles affected by
this final rule, we do not expect that the associated vehicle costs will be high. Our analysis
shows that the federal fleet is already demonstrating actual emissions performance that is much
cleaner than the level to which it is currently being certified. In fact, many MY2013 vehicles
were certified to levels below the 30 mg combined NMOG+NOx standard, some of which were
certified below 70 percent of the 30 mg standard, an important level since it provides necessary
compliance margin. Although the vehicles that make up the federal light-duty fleet are capable
of meeting lower standards there is no impetus for vehicle manufacturers to certify their
respective fleets to anything lower than the current requirements. In addition, we anticipate that
not every technology will be required on all vehicles to meet the standards. While catalyst
loading and engine calibration changes will most likely be applied on all vehicles, only the most
difficult powertrain applications will require very expensive emissions control solutions such as
secondary air injection. We expect that manufacturers will implement emission control solutions
as a function of increasing cost and will avoid implementing very expensive designs whenever
possible.
To determine the cost for vehicles, we first determined which technologies were most
likely to be applied by vehicle manufacturers to meet the standards. These technologies are then
combined into technology packages which reflect vehicle design attributes that directly
contribute to a vehicle's emissions performance. The attributes considered include vehicle type:
car or truck, number of cylinders, engine displacement and the type of fuel used, either gas or
diesel. We also created separate packages for light-duty and heavy-duty trucks and vans.
Once we know the individual technologies that will likely be used, our next step is to
estimate direct manufacturing costs (DMC) for those technologies. As part of this process, we
A We have updated 2010 dollars to 2011 dollars using the Gross Domestic Product (GDP) price deflator as reported
by the Bureau of Economic Analysis on May 30, 2013. The factor used, taken from Line 1 of Table 1.1.4 Price
Indexes for Gross Domestic Product was 1.035 to convert from 2009$ and 1.021 to convert from 2010$. For
example, to convert from 2010$ to 2011$, we calculated the (value in 2010$)xl.021=(value in 2011$).
2-2
-------
determine the model year for which our estimated cost is deemed valid—i.e., if a widget is
estimated to cost $100, is that valid today when none have been sold or after a few years when
thousands or millions have been sold? This "cost basis" serves as the point in time where
learning effects—discussed below—are set to neutral. In other words, beyond that cost basis,
learning effects serve to decrease the DMC of the technology and, in contrast, prior to that cost
basis, the lack of any learning effects serves to increase the DMC of the technology.
The next step is to apply an indirect cost multiplier (ICM) to estimate the indirect cost of
the technology to the auto maker. This is the same process used in the recent MY 2012-2016
and MY 2017-2025 GHG rules, and the proposal for this rule. The cost calculation approach
presumes that the Tier 3 technologies will be purchased by original equipment manufacturers
(OEMs) from Tier 1 suppliers. So, while the DMC estimates include the indirect costs and
profits incurred by the supplier, the ICMs we apply are meant to cover the indirect costs incurred
by OEMs to incorporate the new technologies into their vehicles and to cover the profits that the
OEM must earn to remain viable. We discuss ICMs and indirect costs in more detail in Section
2.2.2.
We have also estimated costs associated with construction of new PM testing facilities.
We have included these costs separately, rather than as part of the ICMs, since the work
conducted to derive our ICMs (details below) did not include new facility construction by
OEMs. We could have included a new factor within the ICM, but believed a separate analysis of
these costs would be more transparent and allows an easier presentation of them as a line item
cost in our analysis. We present the facility costs in Section 2.7.
The next step is to determine the penetration rate of each of the technologies. As noted
above, we do not believe that each of the Tier 3 technologies will be applied to all
engines/vehicles across the board. An obvious example of this would be the evaporative
emission control technologies that will be added to gasoline vehicles but not to diesel vehicles.6
We expect many of the technologies to be used on only a portion of the Tier 3 fleet. Further, the
Tier 3 standards are not implemented 100% in MY2017 and, instead, increase in stringency from
MY2017 through MY2025. Additionally, and new for this final rule, we know that many
vehicles are already being certified with emissions below the Tier 3 30 mg NMOG+NOx level,
even if we give due consideration to compliance margin. Penetration rates and the resultant
technology costs (i.e., inclusive of the penetration rates) are presented below in Section 2.5
where we also sum these costs to arrive at vehicle package costs.
We have also estimated operating costs associated with the evaporative emission
standards and present them in Section 2.6.
The final step is to calculate the vehicle program costs to arrive at annual costs of the Tier
3 vehicle program. We present the vehicle program costs in Section 2.7.
B Diesel fuel has very low volatility so the fuel does not vaporize the way gasoline does.
2-3
-------
2.2.1 Direct Manufacturing Costs
In making our estimates for both direct manufacturing cost (DMC) and application of
technology, we have relied on our own technology assessments. These assessments include
publicly available information, such as that developed by the California Air Resources Board, as
well as confidential information supplied by individual manufacturers and suppliers.2 We have
also considered the results of our own in-house testing.3 The technology packages that we
developed represent what we consider to be the most likely average emissions control solution
for each vehicle type.
In general, we expect that the majority of vehicles will be able to comply with the Tier 3
standards through refinements of current emissions control components and systems. Some
vehicles, for example large trucks with large displacement engines, in particular LDT3s and
LDT4s, may require additional emission controls such as HC adsorbers. Overall, smaller lighter-
weight vehicles will require less extensive improvements than larger vehicles and trucks.
Specifically, we anticipate a combination of technology upgrades for reducing exhaust emissions
as described below.
2.2.2 Indirect Costs
We are using an approach to estimating indirect costs that is consistent with that used in
our recent 2017-2025 Greenhouse Gas (GHG) final rule.4 Rather than a traditional retail price
equivalent markup (RPE), as described below we are marking up DMCs using an indirect cost
multiplier (ICM). Furthermore, we are applying the ICMs in a manner that differs from the
traditional RPE approach in which the DMC would be multiplied by the RPE factor in any given
year. As such, as the DMC decreased with learning, the product of the RPE factor and the DMC
decreased along with it. However, we have more recently decided that learning impacts
(discussed below) should be applied only to the DMC and not to the indirect costs. Our
approach with ICMs, consistent with the recent 2017-2025 GHG final rule, is to determine the
indirect costs based on the initial value of direct costs and then hold that constant until the long-
term ICM is applied. This is done for all ICM factors except warranties, which are influenced by
the learned value of direct costs.
The ICMs used in this final rule are the same as those used in the proposal with one
exception. For this final rule, we have adjusted the R&D portion of the indirect costs to account
for the fact that the research for Tier 3 compliance and a good portion of the development have
been done or are being done in response to the California LEV III rule. Because that research
and development work is attributable to the LEV III rule, we believe it is double counting to also
consider it in the final Tier 3 costs. Below, we discuss this change in greater detail along with
providing a comparison between ICMs in the proposal and this final rule.
To produce a unit of output, auto manufacturers incur direct and indirect costs. Direct
costs include the cost of materials and labor costs. Indirect costs may be related to production
(such as research and development [R&D]), corporate operations (such as salaries, pensions, and
health care costs for corporate staff), or selling (such as transportation, dealer support, and
marketing). Indirect costs are generally recovered by allocating a share of the costs to each unit
of goods sold. Although it is possible to account for direct costs allocated to each unit of goods
2-4
-------
sold, it is more challenging to account for indirect costs allocated to a unit of goods sold. To
make a cost analysis process more feasible, markup factors, which relate total indirect costs to
total direct costs, have been developed. These factors are often referred to as retail price
equivalent (RPE) multipliers.
Cost analysts and regulatory agencies including EPA have frequently used these
multipliers to estimate the resultant impact on costs associated with manufacturers' responses to
regulatory requirements. The best approach to determining the impact of changes in direct
manufacturing costs on a manufacturer's indirect costs would be to actually estimate the cost
impact on each indirect cost element. However, doing this within the constraints of an agency's
time or budget is not always feasible, and the technical, financial, and accounting information to
carry out such an analysis may simply be unavailable.
RPE multipliers provide, at an aggregate level, the relative shares of revenues (Revenue =
Direct Costs + Indirect Costs + Net Income) to direct manufacturing costs. Using RPE
multipliers implicitly assumes that incremental changes in direct manufacturing costs produce
common incremental changes in all indirect cost contributors as well as net income. A concern
in using the RPE multiplier in cost analysis for new technologies added in response to regulatory
requirements is that the indirect costs of vehicle modifications are not likely to be the same for
different technologies. For example, less complex technologies could require fewer R&D efforts
or less warranty coverage than more complex technologies. In addition, some simple
technological adjustments may, for example, have no effect on the number of corporate
personnel and the indirect costs attributable to those personnel. The use of RPEs, with their
assumption that all technologies have the same proportion of indirect costs, is likely to
overestimate the costs of less complex technologies and underestimate the costs of more
complex technologies.
To address this concern, the agency has developed modified multipliers. These
multipliers are referred to as indirect cost multipliers (ICMs). In contrast to RPE multipliers,
ICMs assign unique incremental changes to each indirect cost contributor
ICM = (direct cost + adjusted indirect cost + profit)/(direct cost)
Developing the ICMs from the RPE multipliers requires developing adjustment factors
based on the complexity of the technology and the time frame under consideration. The ICMs
were developed in a peer-reviewed report from RTI International and were subsequently
discussed in a peer-reviewed journal article.5 Note that the cost of capital (reflected in profit) is
included because of the assumption implicit in ICMs (and RPEs) that capital costs are
proportional to direct costs, and businesses need to be able to earn returns on their investments.
The capital costs are those associated with the incremental costs of the new technologies.
As noted above, for the analysis supporting the Tier 3 proposed rulemaking, EPA used
the ICM approach but made some changes to both the ICM factors and to the method of applying
those factors to arrive at a final cost estimate since the ICM work was originally done by RTI.
Both of these changes make the ICMs used in this analysis consistent with those used in the MY
2017-2025 GHG final rule. The first of these changes was done in response to continued
2-5
-------
thinking about how past ICMs have been developed and what are the most appropriate data
sources to rely upon in determining the appropriate ICMs. We have a detailed discussion of this
change in Chapter 3 of the joint TSD supporting the 2017-2025 GHG rule.6 Because that
discussion is meant to present changes made in the time between the original RTI work (and the
MY 2012-2016 GHG final rule) and the MY 2017-2025 GHG final rule, the full text is not really
relevant in the context of Tier 3. The second change has been done both due to staff concerns
and public feedback suggesting that the agency was inappropriately applying learning effects to
indirect costs via the multiplicative approach to applying the ICMs. This change is detailed
below because it is critical to understanding how indirect costs are calculated in the context of
Tier3.
Table 2-1 shows the ICMs used in the proposal. As noted, these ICMs are consistent
with those used in our recent MY 2017-2025 GHG final rule. Despite the fact that these ICMs
were developed with GHG technologies in mind, we are using them here to estimate indirect
costs associated with criteria emission control technology. We believe the ICMs are applicable
here because, as with the GHG requirements, the technologies considered in Tier 3 are or can be
provided to the auto maker by suppliers and their integration into the end vehicle involves the
same sorts of methods and demands as integrating GHG improving technologies.
Table 2-1 Indirect Cost Multipliers Used in the Proposal
Complexity
Low
Medium
Highl
High2
Near term
1.24
1.39
1.56
1.77
Long term
1.19
1.29
1.35
1.50
The second change noted above made to the ICMs has to do with the way in which they
are applied. In the MY 2012-2016 GHG final rule, we applied the ICMs, as done in any analysis
that relied on RPEs, as a pure multiplicative factor. This way, a direct manufacturing cost of,
say, $100 would be multiplied by an ICM of 1.24 to arrive at a marked up technology cost of
$124. However, as learning effects (discussed below) are applied to the direct manufacturing
cost, the indirect costs are also reduced accordingly. Therefore, in year two the $100 direct
manufacturing cost might reduce to $97 and the marked up cost would become $120 ($97 x
1.24). As a result, indirect costs would be reduced from $24 to $23. Given that indirect costs
cover many things such as facility-related costs, electricity, etc., it is perhaps not appropriate to
apply the ICM to the learned direct costs, at least not for those indirect cost elements unlikely to
change with learning. EPA believes that it is appropriate to allow only warranty costs to
decrease with learning since warranty costs are tied to direct manufacturing costs (warranty
typically involves replacement of actual parts which should be less costly with learning). The
remaining elements of the indirect costs should remain constant year-over-year, at least until
some of those indirect costs, such as R&D, are no longer attributable to the rulemaking effort
that imposed them.
As a result, the ICM calculation has become more complex than originally devised by
RTI. We must first establish the year in which the direct manufacturing costs are considered
"valid." For example, a cost estimate might be considered valid today, or perhaps not until high
2-6
-------
volume production is reached—which will not occur until MY2015 or later. That year is known
as the base year for the estimated cost. That cost is the cost used to determine the "non-
warranty" portion of the indirect costs. For example, the non-warranty portion of the medium
complexity ICM in the short-term is 0.343 (the warranty versus non-warranty portions of the
ICMs are shown in Table 2-2). Consider a technology with an estimated direct manufacturing
cost of $70 and valid in MY2015. For this technology, the non-warranty portion of the indirect
costs would be $24.01 ($70 x 0.343). This value would be added to the learned direct
manufacturing cost for each year through 2018, the hypothetical last year of short term indirect
costs for this technology. Beginning in 2019, when long-term indirect costs begin, the additive
factor would become $18.13 ($70 x 0.259). Additionally, the $70 cost in MY2015 would
become $67.90 in MY2016 due to learning (assuming a 3% learning-by-doing cost reduction
from MY2015 to MY2016, or $70 x (1-3%)). So, while the warranty portion of the indirect costs
would be $3.15 ($70 x 0.045) in MY2015, the warranty portion would decrease to $3.06 ($67.90
x 0.045) in 2016 as warranty costs decrease with learning. The resultant indirect costs of the
example technology would be $27.16 ($24.01+$3.15) in MY2015 and $27.07 ($24.01+$3.06) in
MY2016, and so on for subsequent years.
Table 2-2 Warranty and Non-Warranty Portions of ICMs used in the Proposal
Complexity
Low
Medium
Highl
High2
Near term
Warranty
0.012
0.045
0.065
0.074
Non-warranty
0.230
0.343
0.499
0.696
Long term
Warranty
0.005
0.031
0.032
0.049
Non-warranty
0.187
0.259
0.314
0.448
With that as background, we have made minor changes relative to the proposal to the
ICMs used in this final rule. We have made this change because we believe it is appropriate that
the Tier 3 rule not incur costs for research and development that is being incurred by OEMs to
comply with California's LEV III. As such, we have considered half the R&D portion of the
ICM to be research and half to be development. Further, we have set the research portion to 0.0
and the development portion to 50% of the proposal level. These changes mean that our final
rule cost estimates consider all research dollars and 50% of all development dollars to have been
spent in complying with LEV III. The R&D portion of the proposal's ICMs ranges from 3.6% to
7% of total technology costs depending on complexity level. The changes described here result
in the R&D portion ranging from 1% to 2% of total technology costs. In other words, a $100
DMC would have resulted in a $124 total cost in the proposal (at low complexity in the near
term, the ICM being 1.24). In the final rule, the $100 DMC will result in a $121 total cost (at
low complexity in the near term, the ICM being 1.21).
Table 2-3 shows the resultant warranty and non-warranty factors used in the final rule.
These values are used in the final rule instead of those shown in Table 2-2.
2-7
-------
Table 2-3 Warranty and Non-Warranty Portions of ICMs used in the Final Rule
Complexity
Low
Medium
Highl
High2
Near term
Warranty
0.012
0.045
0.065
0.074
Non-warranty
0.196
0.282
0.417
0.543
Long term
Warranty
0.005
0.031
0.032
0.049
Non-warranty
0.172
0.222
0.301
0.365
2.2.3 Cost Reduction through Manufacturer Learning
For this final rule, we have not changed our estimates of learning and how learning will
impact costs going forward from what was employed in the proposal. We consider there to be
one learning effect—learning by doing—which results in cost reductions occurring with every
doubling of production. In the past, we have referred to volume-based and time-based learning.
Those terms were meant only to denote where on the volume learning curve a certain technology
was—"volume-based learning" meant the steep portion of the curve where learning effects are
greatest, while "time-based learning" meant the flatter portion of the curve where learning effects
are less pronounced. Unfortunately, that terminology led some to believe that we were
implementing two completely different types of learning—one based on volume of production
and the other based on time in production. We now use new terminology—steep portion of the
curve and flat portion of curve—simply meant to make clear that there is one learning curve and
some technologies can be considered to be on the steep portion while others are well into the
flatter portion of the curve. This updated terminology was described in the recent heavy-duty
GHG final rule (see 76 FR 57320) and is entirely consistent with our approach used in the recent
MY 2017-2025 GHG final rule (see 77 FR 62711). These two portions of the volume learning
curve are shown in Figure 2-1.
2-8
-------
Volume Leaning Curve - Steep & Flat Portions
120%
100%
Steep portion of volume learning curve
Flat portion of volume learning curve
20%
0%
Cumulative Production
Figure 2-1 Steep and Flat Portions of the Volume Learning Curve
For some of the technologies considered in this analysis, manufacturer learning effects
would be expected to play a role in the actual end costs. The "learning curve" or "experience
curve" describes the reduction in unit production costs as a function of accumulated production
volume. In theory, the cost behavior it describes applies to cumulative production volume
measured at the level of an individual manufacturer, although it is often assumed—as both
agencies have done in past regulatory analyses—to apply at the industry-wide level, particularly
in industries like the light duty vehicle production industry that utilize many common
technologies and component supply sources. We believe there are indeed many factors that
cause costs to decrease over time. Research in the costs of manufacturing has consistently
shown that, as manufacturers gain experience in production, they are able to apply innovations to
simplify machining and assembly operations, use lower cost materials, and reduce the number or
complexity of component parts. All of these factors allow manufacturers to lower the per-unit
cost of production. We refer to this phenomenon as the manufacturing learning curve.
EPA included a detailed description of the learning effect in the MY 2012-2016 and
2017-2025 light-duty GHG rules, the more recent heavy-duty GHG rule and the proposal to this
rule.7 In past rulemaking analyses, EPA has used a learning curve algorithm that applied a
learning factor of 20 percent for each doubling of production volume. EPA has simplified the
approach by using an "every two years" based learning progression rather than a pure production
volume progression (i.e., after two years of production it was assumed that production volumes
2-9
-------
would have doubled and, therefore, costs would be reduced by 20 percent).0 We apply learning
effects on the steep portion of the learning curve for those technologies considered to be newer
technologies likely to experience rapid cost reductions through manufacturer learning, and
learning effects on the flat portion learning curve for those technologies considered to be more
mature technologies likely to experience only minor cost reductions through manufacturer
learning. As noted above, the steep portion learning algorithm results in 20 percent lower costs
after two full years of implementation (i.e., the MY2016 costs would be 20 percent lower than
the MYs 2014 and 2015 costs). Once two steep portion learning steps have occurred, flat portion
learning at 3 percent per year becomes effective for 5 years. Beyond 5 years of learning at 3
percent per year, 5 years of learning at 2 percent per year, then 5 at 1 percent per year become
effective.
For this analysis, learning effects are applied to all technologies because, while most are
already widely used, the technologies would undergo changes relative to their Tier 2 level
design, and we believe auto makers will find ways to reduce costs in the years following
introduction. The steep portion learning algorithm has not been applied to any technologies in
this analysis because we believe that the technologies considered in this analysis have already
experienced the large cost reductions due to learning in the early years of use. The learning
algorithm applied to each technology and the applicable timeframes are summarized in Table
2-4.
Table 2-4 Learning Effect Algorithms Applied to Technolo
Technology
Catalyst Loading
Optimized Close-coupled Catalyst
Optimized Thermal Management
Secondary Air Injection
Engine Calibration
Hydrocarbon Adsorber
Evaporative Emissions Controls
Selective Catalytic Reduction Optimization
Steep learning
gies Used in this Analysis
Flat learning
2015-2025
2015-2025
2015-2025
2015-2025
2015-2025
2015-2025
2015-2025
2015-2025
No learning
c To clarify, EPA has simplified the steep portion of the volume learning curve by assuming that production
volumes of a given technology will have doubled within two years time. This has been done largely to allow for a
presentation of estimated costs during the years of implementation, without the need to conduct a feedback loop that
ensures that production volumes have indeed doubled. The assumption that volumes have doubled after two years is
based solely on the assumption that year two sales are of equal or greater number than year one sales and, therefore,
have resulted in a doubling of production. This could be done on a daily basis, a monthly basis, or, as we have done,
a yearly basis.
2-10
-------
2.3 Individual Technology Costs
2.3.1 Catalyst Platinum Group Metal (PGM) Loading
Increased application of precious metals in the catalyst is expected to be one of the
primary means of mitigating NMOG and NOx to meet the Tier 3 standards. Increasing the
catalyst PGM loading results in greater catalyst efficiency. In the proposal, we noted that vehicle
manufacturers and suppliers had supplied Confidential Business Information (CBI) that estimates
the cost of increasing the PGM loading and modifications to increase the surface area within the
catalyst. These costs ranged from $80 to $260 and were estimated as being incremental to an
existing Tier 2 Bin 5 compliant vehicle. We went on to state that we believed that the
incremental costs for PGM loading would be less than those CBI estimates we received, and we
estimated the costs to be $60, $80, and $100 for an 14, V6 and V8, respectively (all in 2009
dollars).
For this final rule, we have updated our PGM loading costs using a more robust approach
tailored after the methodology presented by ICCT in a recent SAE paper.8 In that paper, ICCT
outlines a costing methodology based on PGM loads, swept volume ratio (the ratio of catalyst
volume to engine displacement), and some equations that can be used to estimate catalyst
washcoating and canning costs based on catalyst volume. This approach is actually similar to an
approach used by EPA in past cost analyses that focused heavily on aftertreatment device costs.9
We have made these changes for several reasons, but primarily because some commenters
believed our cost estimates were dated, having relied heavily on the CARB LEVIII analysis now
several years old. We agreed with this assessment and also liked the ICCT methodology since it
allows us to provide more detail behind the estimates and to be transparent with the estimate
allowing others to adjust things in ways they may believe make more sense.
In their recent SAE paper, ICCT estimates the PGM loading of Tier 2 catalysts at 0.1 g/L
Platinum (Pt), 1.6 g/L Palladium (Pd) and 0.1 g/L Rhodium (Rh). Further, they estimate that the
swept volume ratio of Tier 2 catalysts is 1.0 (i.e., the catalyst volume equals the engine
displacement). They also provide 3 equations that can be used to estimate catalyst substrate,
washcoating and canning costs. Those equations are shown in Table 2-5. ICCT also included
labor costs in a manner described as consistent with past EPA work.0 ICCT notes that their
methodology considers the catalyst system as a unit and does not distinguish between close
coupled and underfloor catalysts. This was done in an effort to simplify the approach even
though close coupled and underfloor catalysts may well have different loadings.
D ICCT references the source as EPA's Nonroad Tier 4 Regulatory Impact Analysis (EPA420-R-04-007, May
2004). In that analysis, a labor rate of $30/hour (2003$) was used. Updating that to 2011$ using the GDP price
deflator mentioned earlier results in a labor rate of $36/hour for this analysis.
2-11
-------
Table 2-5 ICCT Equations for Estimating Catalyst Component Costs (2011$)
Component
Catalyst substrate
Catalyst washcoat
Catalyst canning
ICCT Cost equation
$6.0xVol+$1.92
$5.0xVol
$2.4xVol
Note: Vol = catalyst volume
Source: SAE 2013-01-0534
In their comments, ICCT notes recent research by Honda and Johnson Matthey showing
PGM usage could be reduced by 25% with respect to current Tier 2 Bin 5 levels and still provide
LEVIII SULEV30 compliance (i.e., Tier 3 Bin 30). This could be done using an improved
layered catalyst and improved oxygen storage capacity (OSC) via adding zirconia along with
ceria. ICCT also notes research by Umicore showing LEV70 to SULEV30 reductions via a 32%
increase in PGM loading on one vehicle and only 16% increase on another. ICCT notes that
these two vehicles did increase catalyst volume between 40% and 200% which would serve to
increase costs. In that study, Umicore stresses the importance of a combined NMOG+NOx
standard versus separate NMOG and NOx standards noting that the combined standard is much
less demanding.10 This is discussed in more detail in Chapter 1 of this RIA, but the combined
standard provides much more flexibility to auto makers than does separate standards, thus
allowing them to control costs much more effectively. What is important here is that the
Umicore catalyst volume increases and subsequent costs were done assuming separate standards
(the California PZEV standards), not a combined standard. Also important in the Umicore study
was that the PZEV catalysts made no use of Pt for NMOG control, relying only on Pd for
NMOG control and Rh for NOx control. This is important because Pd is, generally, less costly
and exhibits less price volatility than Pt. All told, the increased costs in the Umicore study—
again, for separate, not a combined standard—were on the order of $15-$46, well below our
NPRM costs of $60 to $100 (both in 2009$). ICCT also provided their assessment that a 20%
increase in PGM loading is the most that would be required for Tier 3.
We believe that, in general, catalyst loading will increase in the front most portion of the
catalyst system but not necessarily the entire system. We believe this, in part, because of the
strong cases made by commenters and in the recent studies mentioned. We also believe this
because the Tier 3 standards are, in effect, cold start emission standards, and we believe that the
catalyst loading will be increased for the purpose of controlling cold start emissions. This can be
most effectively done by adding metals to the portion of the catalyst system that will reach
operating temperatures most quickly—i.e., that portion closest to the point where gases are
exhausted from the engine. As such, we have estimated that Pd loading will increase 50% and
Rh loading 20% but only in the front most 50% of the catalyst system. We have also estimated
that Pt will be eliminated in favor of Pd, and that total catalyst volume (swept catalyst volume,
SVR) will increase by 20%.
2-12
-------
Using these metrics, the equations shown in Table 2-5 and the price of PGMs,E we can
calculate the increased cost of any Tier 3 catalyst relative to its Tier 2 counterpart provided we
know the Tier 2 catalyst system's volume. For fleet average incremental costs, we need both the
Tier 2 catalyst volumes for each vehicle in the fleet and the sales of each vehicle to get a proper
sales weighted average catalyst cost. To get these, we have used the baseline file used in support
of our recent GHG/CAFE final rule for MY 2017-2025. That baseline file represented the 2008
model year fleet and has in it the engine displacement (i.e., the catalyst volume since we consider
the SVR of Tier 2 catalysts to be 1.0) and the projected sales of each vehicle model in the light-
duty fleet for MY 2012 through 2025. Using that fleet in conjunction with the projected
MY2013 sales, we were able to calculate the catalyst costs shown in Table 2-6.
Table 2-6 Catalyst Loading Direct Manufacturing Costs for Gasoline Vehicles (2011$)
Sales weighted engine
displacement (L)
Sales weighted catalyst
volume (L)
PGM cost
Substrate cost
Washcoat cost
Canning cost
Labor cost
Catalyst cost per
vehicle
Sales weighted catalyst
volume (L)
PGM cost
Substrate cost
Washcoat cost
Canning cost
Labor cost
Catalyst cost per
vehicle
Sales weighted catalyst
volume (L)
PGM cost
Substrate cost
Washcoat cost
Canning cost
Labor cost
Catalyst cost per
vehicle (2011$)
DMC in our proposal
(2009$)
Standard
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
TierS
TierS
TierS
TierS
TierS
TierS
TierS
Increment
Increment
Increment
Increment
Increment
Increment
Increment
Increment
PC 14 G
2.1
2.1
$94
$14
$10
$5
$9
$132
2.5
$125
$17
$12
$6
$9
$169
0.4
$32
$2
$2
$1
$0
$37
$62
PC V6 G
3.3
3.3
$150
$22
$16
$8
$9
$204
3.9
$200
$26
$20
$9
$9
$264
0.7
$50
$4
$3
$2
$0
$59
$83
PC V8 G
4.9
4.9
$223
$31
$24
$12
$9
$299
5.9
$298
$37
$29
$14
$9
$387
1.0
$75
$6
$5
$2
$0
$88
$104
LT 14 G
2.4
2.4
$111
$17
$12
$6
$9
$155
2.9
$148
$19
$15
$7
$9
$199
0.5
$37
$3
$2
$1
$0
$44
$62
LT V6 G
3.6
3.6
$164
$24
$18
$9
$9
$224
4.3
$220
$28
$22
$10
$9
$289
0.7
$55
$4
$4
$2
$0
$65
$83
LT V8 G
5.2
5.2
$239
$33
$26
$13
$9
$320
6.3
$319
$40
$31
$15
$9
$414
1.0
$80
$6
$5
$3
$0
$94
$104
In their study, Umicore estimated the LEV-III PGM costs for a 2.0L engine ranging from
$81-117. These estimates compare favorably to our estimate of $125 for an 14 passenger car.
E For this analysis, we have used the PGM spot price as of July 16, 2013, reported at 9:30AMinNew York. Those
values were: Pt=$l,426/troy oz.; Pd=$735/troy oz.; Rh=$l,000/troy oz.
2-13
-------
While our final estimates are lower than those in our proposed rule, we consider these final rule
costs to be much more robust, transparent and appropriate. Several years have passed since
generating the catalyst loading costs presented in our proposal—as explained, they were
generated as part of the LEV-III rule and our early work on Tier 3. Several commenters
suggested that our proposed costs were now dated, and CARB also recommended that we revisit
our cost estimates in light of the passage of time and more recent information.11
We consider these incremental costs to be applicable in MY2015 with flat learning
applied thereafter. We also consider catalyst loading to be a low complexity technology with
near term markup factors applied through 2022 and long term thereafter. The resultant DMC,
indirect costs (1C) and total costs (TC) are shown in Table 2-7. Note that the values shown do
not include penetration rates.
Note that we have not changed the catalyst loading costs for heavy-duty vehicles relative
to the proposal, with the exception of updating them to 2011 dollars. We do not show costs for
diesel or >14,000 pound gasoline vehicles since those vehicles are not expected to incur any new
catalyst loading costs.
2-14
-------
Table 2-7 Catalyst Loading Costs for Gasoline Vehicles (2011$)
Vehicle
category
PCM
PCV6
PCV8
LTI4
LTV6
LTV8
Class 2b
Class 3
PCM
PCV6
PCV8
LTI4
LTV6
LTV8
Class 2b
Class 3
PCM
PCV6
PCV8
LTI4
LTV6
LTV8
Class 2b
Class 3
Cost
DMC
DMC
DMC
DMC
DMC
DMC
DMC
DMC
1C
1C
1C
1C
1C
1C
1C
1C
TC
TC
TC
TC
TC
TC
TC
TC
2015
$37
$59
$88
$44
$65
$94
$52
$52
$8
$12
$18
$9
$14
$20
$11
$11
$45
$71
$106
$53
$79
$114
$63
$63
2016
$36
$57
$85
$43
$63
$91
$50
$50
$8
$12
$18
$9
$14
$20
$11
$11
$44
$70
$104
$52
$77
$111
$61
$61
2017
$35
$56
$83
$41
$61
$89
$49
$49
$8
$12
$18
$9
$13
$20
$11
$11
$43
$68
$101
$50
$75
$108
$59
$59
2018
$34
$54
$80
$40
$59
$86
$47
$47
$8
$12
$18
$9
$13
$20
$11
$11
$41
$66
$99
$49
$73
$106
$58
$58
2019
$33
$52
$78
$39
$58
$83
$46
$46
$8
$12
$18
$9
$13
$19
$11
$11
$40
$65
$96
$48
$71
$103
$57
$57
2020
$32
$51
$76
$38
$56
$81
$44
$44
$8
$12
$18
$9
$13
$19
$11
$11
$39
$63
$94
$47
$69
$100
$55
$55
2021
$31
$50
$74
$37
$55
$79
$44
$44
$8
$12
$18
$9
$13
$19
$11
$11
$39
$62
$92
$46
$68
$99
$54
$54
2022
$31
$49
$73
$36
$54
$78
$43
$43
$8
$12
$18
$9
$13
$19
$11
$11
$38
$61
$91
$45
$67
$97
$53
$53
2023
$30
$48
$71
$35
$53
$76
$42
$42
$7
$10
$16
$8
$11
$17
$9
$9
$36
$58
$87
$43
$64
$93
$51
$51
2024
$29
$47
$70
$35
$51
$75
$41
$41
$7
$10
$16
$8
$11
$17
$9
$9
$36
$57
$85
$43
$63
$91
$50
$50
2025
$29
$46
$68
$34
$50
$73
$40
$40
$7
$10
$16
$8
$11
$17
$9
$9
$35
$56
$84
$42
$62
$90
$49
$49
PC=passenger car; LT=light truck; DMC=direct manufacturing cost; IC=indirect cost; TC=total cost; 2015 & 2016
costs shown because the cost basis for the technology is 2015 where the learning factor= 1.0; note that the costs
shown do not include penetration rates so these costs represent technology costs, not package costs for packages of
technologies expected to be applied to Tier 3 vehicles.
2.3.2 Optimized Close-coupled Catalyst
Close-coupled catalyst technologies include improvements to the catalyst system design,
structure, and packaging to reduce light-off time. As catalysts are moved closer to the engine the
temperature of the exhaust gases to which catalysts are exposed under high load operation goes
up substantially. As a result some of the materials used in the catalyst construction, as well as
the precious metals used in close-coupled applications, must be improved to survive in the higher
operating temperatures. In the proposal, we stated that cost estimates for close-coupled catalyst
designs received from vehicle manufacturers ranged from $25 to $50, however, they did not
include all of the considerations identified above. Consistent with the proposal but updated to
2011 dollars, we have estimated the cost for an 14 gasoline engine to be $21, a V6 to be $41, and
a V8 to be $62. As noted, all DMC are in 2011 dollars.
We consider these incremental costs to be applicable in MY2015 with flat learning
applied thereafter. We consider close coupled catalysts to be a low complexity technology with
near term factors applied through 2022 and long term thereafter. The resultant costs are shown
in Table 2-8. Note that the values shown do not include penetration rates.
2-15
-------
We do not show costs for any diesel or heavy-duty vehicles since those vehicles are not
expected to incur new close coupled catalyst costs.
Table 2-8 Optimized Close Coupled Catalyst Costs for Gasoline Vehicles (2011$)
Vehicle
category
PCM
LTI4
PCV6
LTV6
PCV8
LTV8
PCM
LTI4
PCV6
LTV6
PCV8
LTV8
PCM
LTI4
PCV6
LTV6
PCV8
LTV8
Cost
DMC
DMC
DMC
1C
1C
1C
TC
TC
TC
2015
$21
$41
$62
$4
$9
$13
$25
$50
$75
2016
$20
$40
$60
$4
$9
$13
$24
$49
$73
2017
$19
$39
$58
$4
$9
$13
$24
$48
$71
2018
$19
$38
$57
$4
$9
$13
$23
$46
$70
2019
$18
$37
$55
$4
$9
$13
$23
$45
$68
2020
$18
$36
$53
$4
$9
$13
$22
$44
$66
2021
$17
$35
$52
$4
$9
$13
$22
$43
$65
2022
$17
$34
$51
$4
$9
$13
$21
$43
$64
2023
$17
$33
$50
$4
$7
$11
$20
$41
$61
2024
$16
$33
$49
$4
$7
$11
$20
$40
$60
2025
$16
$32
$48
$4
$7
$11
$20
$39
$59
PC=passenger car; LT=light truck; DMC=direct manufacturing cost; IC=indirect cost; TC=total cost; 2015 & 2016
costs shown because the cost basis for the technology is 2015 where the learning factor= 1.0; note that the costs
shown do not include penetration rates so these costs represent technology costs, not package costs for packages of
technologies expected to be applied to Tier 3 vehicles.
2.3.3 Optimized Thermal Management
Overall thermal management of the emissions control system to shorten the time it takes
for the catalyst to light-off will most likely be a primary technology for mitigating NMOG on
gasoline vehicles and NOx on diesel vehicles. This technology includes dual wall exhaust
manifolds and pipe that will help maintain exhaust gas temperatures from the exhaust port of the
engine to the close-coupled catalyst or, in the case of diesel engines, the Selective Catalyst
Reduction (SCR) system. In some cases, the packaging of the exhaust system will be modified
to reduce the wetted area of the exhaust path. This will, in turn, reduce the decrease in exhaust
gas temperatures associated with a longer exhaust path. Consistent with the proposal and based
on CBI submitted by exhaust system suppliers and vehicle manufacturers, we estimate that the
cost of implementing dual wall exhaust designs are approximately $31 (2011$) for all engine
applications.
We consider these incremental costs to be applicable in MY2015 with flat learning
applied thereafter. We consider optimized thermal management to be a low complexity
technology with near term factors applied through 2022 and long term thereafter. The resultant
costs are shown in Table 2-9. Note that the values shown do not include penetration rates.
We do not show costs for > 14,000 pound gasoline vehicles since those vehicles are not
expected to incur new optimized thermal management costs.
2-16
-------
Table 2-9 Optimized Thermal Management Costs for Gasoline and Diesel Vehicles (2011$)
Vehicle
category
PC
LT
Class 2b
Class 3
PC
LT
Class 2b
Class 3
PC
LT
Class 2b
Class 3
Cost
DMC
1C
TC
2015
$31
$6
$38
2016
$30
$6
$37
2017
$29
$6
$36
2018
$28
$6
$35
2019
$27
$6
$34
2020
$27
$6
$33
2021
$26
$6
$33
2022
$26
$6
$32
2023
$25
$5
$31
2024
$25
$5
$30
2025
$24
$5
$30
PC=passenger car; LT=light truck; DMC=direct manufacturing cost; IC=indirect cost; TC=total cost; 2015 & 2016
costs shown because the cost basis for the technology is 2015 where the learning factor= 1.0; note that the costs
shown do not include penetration rates so these costs represent technology costs, not package costs for packages of
technologies expected to be applied to Tier 3 vehicles.
2.3.4 Secondary Air Injection
Secondary air injection is a technology that provides a source of combustion air such that
a portion of the exhaust gases are burned in the exhaust manifold. This technology provides
increased heat in the exhaust system that provides for faster catalyst light-off It is used only
during cold start and requires that the air/fuel mixture is rich such that a small amount of fuel is
available for combustion outside of the combustion chamber. We expect that some gasoline
V6's and V8's will require the application of secondary air injection to reduce NMOG
emissions. The secondary air injection system consists of an air pump (normally electrically
powered), plumbing from the pump to the exhaust manifold, an electrically controlled valve,
control circuitry in the powertrain control module, wiring and calibration. CBI estimates
received from vehicle manufacturers and suppliers ranged from $50 to $310. Consistent with the
proposal, we have estimated that the final direct manufacturing cost for secondary air is $104
(2011$) for any application that may need to add it.
We consider these incremental costs to be applicable in MY2015 with flat learning
applied thereafter. We consider secondary air to be a low complexity technology with near term
factors applied through 2022 and long term thereafter. The resultant costs are shown in Table
2-10. Note that the values shown do not include penetration rates.
We do not show costs for gasoline 14, diesel or heavy-duty vehicles since none of those
vehicles are expected to incur new secondary air injection costs.
2-17
-------
Table 2-10 Secondary Air Injection Costs for Gasoline Vehicles (2011$)
Vehicle
category
PCV6
PCV8
LTV6
LTV8
PCV6
PCV8
LTV6
LTV8
PCV6
PCV8
LTV6
LTV8
Cost
DMC
1C
TC
2015
$104
$22
$125
2016
$100
$22
$122
2017
$97
$21
$119
2018
$94
$21
$116
2019
$92
$21
$113
2020
$89
$21
$110
2021
$87
$21
$108
2022
$85
$21
$107
2023
$84
$18
$102
2024
$82
$18
$100
2025
$80
$18
$99
PC=passenger car; LT=light truck; DMC=direct manufacturing cost; IC=indirect cost; TC=total cost; 2015 & 2016
costs shown because the cost basis for the technology is 2015 where the learning factor= 1.0; note that the costs
shown do not include penetration rates so these costs represent technology costs, not package costs for packages of
technologies expected to be applied to Tier 3 vehicles.
2.3.5 Engine Calibration
Product changes considered for engine calibration include engine control and calibration
modifications to improve air and fuel mixtures, particularly at cold start and/or to control
secondary air. While typically there are no direct manufacturing costs associated with the
calibration itself, we recognize that some additional engineering efforts will be required to
implement the changes described above. As in the proposal, we have estimated the per vehicle
cost at $2 (2011$). For gasoline engines, we have added a new engine calibration cost to address
GDI PM-related concerns. We have estimated these new costs for gasoline engines at an
additional $2 (2011$) per engine. The result being a total engine calibration cost of $4 (2011$)
per gasoline engine and $2 (2011$) per diesel engine.
We consider these incremental costs to be applicable in MY2015 with flat learning
applied thereafter. We consider engine calibration to be a low complexity technology with near
term factors applied through 2022 and long term thereafter. The resultant costs are shown in
Table 2-11. Note that the values shown do not include penetration rates.
We do not show costs for >14,000 pound gasoline vehicles since none of those vehicles
are expected to incur new engine calibration costs.
2-18
-------
Table 2-11 Engine Calibration Costs for Gasoline and Diesel Vehicles (2011$)
Vehicle
category
PC, LT,
Class 2b, 3
Gasoline
PC, LT,
Class 2b, 3
Diesel
PC, LT,
Class 2b, 3
Gasoline
PC, LT,
Class 2b, 3
Diesel
PC, LT,
Class 2b, 3
Gasoline
PC, LT,
Class 2b, 3
Diesel
Cost
DMC
DMC
1C
1C*
TC
TC
2015
$4
$2
$1
$0
$5
$3
2016
$4
$2
$1
$0
$5
$2
2017
$4
$2
$1
$0
$5
$2
2018
$4
$2
$1
$0
$5
$2
2019
$4
$2
$1
$0
$5
$2
2020
$4
$2
$1
$0
$4
$2
2021
$3
$2
$1
$0
$4
$2
2022
$3
$2
$1
$0
$4
$2
2023
$3
$2
$1
$0
$4
$2
2024
$3
$2
$1
$0
$4
$2
2025
$3
$2
$1
$0
$4
$2
PC=passenger car; LT=light truck; DMC=direct manufacturing cost; IC=indirect cost; TC=total cost; 2015 & 2016
costs shown because the cost basis for the technology is 2015 where the learning factor= 1.0; note that the costs
shown do not include penetration rates so these costs represent technology costs, not package costs for packages of
technologies expected to be applied to Tier 3 vehicles.
* Actual values are less than 50 cents and appear here as $0 due to rounding for simplicity of presentation.
2.3.6 Hydrocarbon Adsorber
Hydrocarbon adsorbers trap hydrocarbons during cold start and release the hydrocarbons
after the catalyst lights off. Hydrocarbon adsorbers can be applied in two different manners:
The first is a passive device which traps hydrocarbons at cold start and releases them as the
temperature of the device increases. The catalyst may or may not have lit off at the time of
desorption making a rapid catalyst temperature rise and light off critical. The second is an active
hydrocarbon adsorber. This device controls the adsorber exposure to exhaust gases based on
temperature and is able to trap the hydrocarbons until the catalyst has lit off. The effectiveness
of the active hydrocarbon system is much greater than the passive system. However, the active
system is also much more costly. In the proposal, we anticipated that manufacturers would apply
only active systems due to a perception that passive systems were limited in their ability to
mitigate NMOG. We estimated the cost of active hydrocarbon adsorber systems at $150
(2009$). For the final rule, we have changed our expectations and now expect that any HC
adsorber use will be passive rather than active. We base this on comments from MECA and
ICCT and on CBI provided by Tier 1 suppliers after the proposal.12 We have estimated the DMC
of a passive HC adsorber at $16 (2011$, or $15 in 2009$), and we expect their use on only a
portion of the largest gasoline engines.
We consider these incremental costs to be applicable in MY2015 with flat learning
applied thereafter. We consider passive HC adsorbers to be a medium complexity technology
with near term factors applied through 2022 and long term thereafter. The resultant costs are
shown in Table 2-12. Note that the values shown do not include penetration rates.
2-19
-------
We do not show costs for 14 or V6 gasoline or for diesel vehicles since none of those
vehicles are expected to incur new HC adsorber costs.
Table 2-12 Passive HC Adsorber Costs for Gasoline Vehicles (2011$)
Vehicle
category
PCV8,
LTV8,
Class 2b, 3
PCV8,
LTV8,
Class 2b, 3
PCV8,
LTV8,
Class 2b, 3
Cost
DMC
1C
TC
2015
$16
$5
$21
2016
$15
$5
$20
2017
$15
$5
$20
2018
$14
$5
$19
2019
$14
$5
$19
2020
$13
$5
$18
2021
$13
$5
$18
2022
$13
$5
$18
2023
$13
$4
$16
2024
$12
$4
$16
2025
$12
$4
$16
PC=passenger car; LT=light truck; DMC=direct manufacturing cost; IC=indirect cost; TC=total cost; 2015 & 2016
costs shown because the cost basis for the technology is 2015 where the learning factoi= 1.0; note that the costs
shown do not include penetration rates so these costs represent technology costs, not package costs for packages of
technologies expected to be applied to Tier 3 vehicles.
2.4 Evaporative Emission, Canister Bleed, and Leak Controls
2.4.1 DMCs of Technologies Expected to See Widespread Implementation
Chapter 1 identified six different technologies EPA expected to see widespread use in
achieving the emission reductions needed to meet the Tier 3 evaporative emission standard (hot
soak plus diurnal, canister bleed, and leak). These technologies are in-use to varying degrees in
the fleet today. The technologies of interest are identified in Table 1-11 of Chapter 1. Thattable
is replicated here, but with different entries to address the fleet penetration and costs of the
various technologies by vehicle category. The DMCs presented below are EPA estimates based
on discussions with manufacturers and vendors and review of the comments on the CARB LEV
III evaporative emissions rule. See Table 2-13. All DMC are in 2011 dollars. Each of these is
discussed below.
2-20
-------
Table 2-13 Tier 3 Evaporative, Leak and OBD DMCs
Vehicle Class
nonfuel (g)
MSATstd(3d/2d)(g)
T3std(g)
Canister bleed std(g)
MSATBaseline(g)1
TBTarget(g)'
Red Needed(g)
RedAchieved(g)
LDV
0.1
0.5/0.65
0.3
0.02
0.47
0.2
0.27
0.39
0.1
0.65/0.85
0.3
0.02
0.47
0.2
0.27
0.39
0.125
0.65/0.85
0.4
0.02
0.42
0.3
0.12
0.39
LDT3
0.15
0.9/1.15
0.5
0.02
0.72
0.37
0.35
0.39
0.15
0.9/1.15
0.5
0.02
0.72
0.37
0.35
0.39
0.175
1.0/1.25
0.5
0.02
0.72
0.37
0.35
0.39
LHDGV
0.2
1.4/1.75
0.6
0.03
0.96
0.45
0.51
0.55
HHDGV
0.25
1.9/2.3
0.6
0.03
0.96
0.45
0.51
0.52
Control
Technology DMC Efficiency
Canister $8.50 90%
honeycomb 35mmx
75mm
Canister $7.50forV- 90%
honeycomb 8s 35mm x
50mm
AIS scrubber $6 90%
PFItoGDI $0 90%
Fuel system
architecture
(a) reduce $2 90%
connections &
improve seals/ o-
rings as needed
(b) move parts $0.50 100%
into tank
OBD software 0
upgrades
Tota DMC
Cost&
%Application DMC
$8.50_100% $8.50
0 0
$6_40% $2.40
0 0
$2_100% $2
0.5_$0.50 $0.25
0 0
$13.15
Cost&
%Application DMC
$8.50_100% $8.50
0 0
$6_27% $1.62
0 0
$2_100% $2
0.5_$0.50 $0.25
0 0
$12.37
Cost%
%Application DMC
$8.50_76% $6.46
$7.5_24% $1.82
$6_27% $1.62
0 0
$2_100% $2
0.5_$0.50 $0.25
0 0
$12.15
Cost&
%Application DMC
$8.50_10% $0.85
$7.5_90% $6.75
$6_27% $1.62
0 0
$2_100% $2
0.5_$0.50 $0.25
0
$11.47
Cost&
%Application DMC
0 0
$7.5_100% $7.50
$6_27% $1.62
0 0
$2_100% $2
0.5_$0.50 $0.25
0 0
$11.37
Cost&
%Application DMC
0 0
$7.5_100% $7.50
$6_27% $1.62
0 0
$2_100% $2
0.5_$0.50 $0.25
0 0
$11.37
Cost&
%Application DMC
0 0
$7.5_100% $7.50
0 0
0 0
$2_100% $2
0.5_$0.50 $0.25
0 0
$9.75
Cost&
%Application DMC
0 0
$7.5_100% $7.50
$6_100% $6
0 0
$2_100% $2
0 0
0 0
$15.50
0.051
0.051
fuel savings gal/veh-yr 0.051 0.051
1 based on mean pi us one standard deviation for 2013 MY 2-day cert results on Tier 2fuel
1100 mg or 25% below T3 std whichever is greater
3 (365dayperyear)(lgal per 5.6 lbs)(llb per 454g)(mg red uction/day)(lg/1000mg)(0.9 energy density effect); needsto be further multiplied by(15yr)(avgsurvfractionforfleet)(gas price^
0.051
0.051
0.059
0.053
2-21
-------
This page intentionally left blank.
2-22
-------
Canister honeycomb: Current evaporative canisters use high working-capacity activated
carbon, usually with multiple compartments, to optimize vapor loading and purging behavior.
These canisters sometimes employ carbons of different working capacities within each chamber.
Manufacturers may adjust the shapes and sizes of internal compartments, including design
variations to include different grades of carbon in different areas to best manage rapid purge
following engine starting, back purge during overnight parking, vapor loading at different
loading rates, and vapor redistribution and migration while the vehicle is not operating. The
biggest expected change to evaporative emission canisters is the addition of a secondary canister
element, either attached to the canister body, or integral to it, in which a carbon with very low
working capacity is available to capture diffusion emissions (also known as bleed emissions).
This is commonly referred to as a canister scrubber. While this carbon element can hold only a
few grams of hydrocarbon, it back purges easily and purges readily with a short amount of
driving, so it is always ready to capture the small amount of hydrocarbon that escapes the body
of the evaporative canister as a result of diffusion from vapor migration within the carbon bed.
For purposes of this analysis, we expect that all vehicles covered by the rule use a canister
scrubber. Slightly larger scrubbers are expected on vehicles with V-4 and V-6 engines, since
they are expected to have less available purge for the primary canister than will V-8 engines. The
scrubbers will vary in size, but a typical unit DMC is $7.50-8.50. We expect that in most cases
these will be built as an integral part of the current canister to avoid extra packaging costs. In
some cases, dual tank HDGVs may employ two evaporative emission canisters.
Engine/fuel system conversion: To the extent that manufacturers use direct injection,
there should be very little fuel vapor coming from the intake system. Any unburned fuel coming
from the injectors would be preserved in the cylinder or released to the exhaust system and the
catalyst. A small amount of crankcase vapor might remain, but this would likely not be enough
to justify adding carbon to the intake system. As projected in our final rule RIA for the 2017-
2025 GHG emissions, EPA projects a significant movement from port fuel injection (PFI)
engines to gasoline direct injection (GDI) engines. This ranges from 60-100 percent of products
for all categories except gasoline-powered trucks over 14,000 Ibs GVWR. This reduces air
induction systems emissions by 90 percent since the GDI uses a different fuel injection timing
strategy than the PFI This would not involve any additional cost for control of Tier 3
evaporative emissions.
Air intake scrubbers: Manufacturers have identified the engine's intake system as
another source of evaporative emissions. These result from crankcase vapors and from unburned
fuel from injectors, or sometimes from an injection event that occurred shortly before engine
shutdown. One way to prevent these emissions is to add a device containing activated carbon to
the air intake downstream of the air filter, typically in the form of reticulated foam coated with
activated carbon. This device would have only a few grams of working capacity and would be
designed to purge easily to ensure that the vapor storage is available any time the engine shuts
down. This carbon insert would almost completely eliminate vapor emissions from the air intake
system. This analysis projects that vehicles/engines not converting to GDI will use the air intake
scrubber to address this source of emissions. The percentages by vehicle category are shown in
Table 2-13 above. The intake scrubber DMC is approximately $6.00 per vehicle.
2-23
-------
Fuel-system architecture: As discussed below, there may be the opportunity to reduce
permeation emissions from some fuel lines. However, the bigger area of expected development
with respect to fuel lines is to re-engineer whole systems to reduce the number of connections
between fuel-system components and other fuel-line segments. While manufacturers have
already made some changes in this direction, these systems may still involve more than the
optimum number of connections and segments due to assembly and production considerations or
other factors. Designing the fuel system more carefully to minimize connection points will limit
possible paths for fuel vapors to escape. This would reduce emission rates and it should also
improve system durability by eliminating potential failure points. A broader approach to
addressing this source of emissions is to integrate designs and to move fuel-system components
inside the fuel tank, which eliminates the concern for vapor emissions and permeation from those
components and connections. Most of the costs associated with these upgrades lie in
development and tooling. There may be some additional part costs, but the overall trend should
ultimately allow for reduced costs from reducing the number of components and reducing
assembly time. To the extent that fuel-system components are moved inside the fuel tank, there
may be further cost savings since those components would no longer need to be made from low-
permeation materials. DMC for these actions is about $2.50
Onboard Diagnostics (OBD): EPA and CARB have similar but not identical OBD
requirements for LDVs, LDTs, MDPVs, and HDGVs up to 14,000 Ibs GVWR. Within the past
five years CARB has revised their implementation scheme and upgraded requirements to
improve the effectiveness of their systems in addressing potential exhaust and evaporative
system performance issues in use. EPA regulations permit manufacturers to meet CARB's most
recent requirements and to seek a Federal certificate based on meeting CARB's requirements.
Certification based on meeting CARB's requirements and application of those OBD systems
nationwide is common practice in the industry with only a few exceptions. EPA is adopting
current CARB OBD certification, verification, and monitoring requirements. As part of our rule,
we are including two new elements; (1) certification that the OBD evaporative system leak
monitor is able to find a 0.020 inch leak and (2) a requirement that the OBD computer store
information on when the full OBD leak monitoring protocol was last run successfully and the
result of that assessment. Since current CARB OBD requirements are being met by
manufacturers, additional costs are attributable to certification to the 0.020 inch leak detection
requirement and software modification to retain information on the last successful run of the
OBD evaporative system leak monitor. EPA estimates these two items to cost on average
approximately $0.10 (2011$) per vehicle. These are reflected in indirect costs discussed below.
Leak standard testing: As part of the Tier 3 evaporative emission requirements EPA is
proposing a vapor leak emission standard. EPA expects that many of the technologies and
approaches for reducing evaporative emissions described above will assist in addressing
potential vapor leak problems and that in most cases no specific additional measures would be
needed. Nevertheless, there might be two additional cost areas. First would be certification
testing. However, EPA is allowing certification requirements for the vapor leak emission
standard to be met by written attestation rather than by testing since the certification vehicle
would fail the hot soak plus diurnal evaporative emissions standard if it had a 0.02 inch leak.
Manufacturers agree this is appropriate. Second, EPA is proposing to include assessment of the
vapor leak emission standard within the in-use verification testing program (IUVP). However,
we have structured the program to minimize additional costs. Testing will be required on all
2-24
-------
vehicles otherwise procured for exhaust emissions. All vehicles tested for exhaust emissions
must also be tested for the leak emission standard. Thus, we generally expect multiple leak test
results per group but in no case may there be no fewer than one test group representatives for
each evaporative/refueling/leak family. Unless there are performance problems, no additional
vehicle procurement costs are expected. Also, we are proposing to permit the manufacturer to
use its current evaporative system leak monitoring OBD hardware to screen vehicles from IUVP
testing for leaks and/or to use as an option to the proposed EPA test procedure if testing is
needed. The additional costs for leak emission testing for IUVP (approximately $0.10 (2011$)
per vehicle) are included in the indirect costs discussed below.
Taken together, these technologies applied to the fleet to the degree described in the
paragraphs above result in an estimated DMC of $10-15 per vehicle in 2011 dollars.
2.4.2 DMCs of technologies which may be optimized if necessary to achieve further reductions
Chapter 1 of the RIA also identified four technology approaches which are in widespread
use today but with some refinement and optimization could provide additional reductions. These
are discussed below, but not included in the overall cost analysis. These technologies are likely
to see limited application because in comparison to the technologies in Table 1-11 of Chapter 1
the emission reductions available (see Table 1-12 of Chapter 1) are small relative to the costs. If
implemented, as can be seen in Table 2-13, the five approaches discussed above would provide
more than enough reductions to meet the emission targets for the hot soak plus diurnal and
canister bleed standards at certification even accounting for the non-fuel hydrocarbon effects.
Thus, we are not projecting penetration rates for these technologies.
Upgrade canister and improve purge: Recent and projected engine design changes are
increasing the challenge to maintain manifold vacuum for drawing purge air through the
evaporative canister. Several different technology options would help to address this increasing
challenge. Different grades of carbon and canister configurations can lead to a more effective
canister purge for a given volume of air flowing over the canister. If employed, such strategies
would cost $2-4 per vehicle.
Improve fuel tank barrier layer thickness and reduce pinch seam gaps: Fuel tanks are
already designed to limit permeation emissions. Fuel tanks are typically made of high-density
polyethylene with an embedded barrier layer of ethyl vinyl alcohol (EvOH) representing about
1.8 percent of the average wall thickness. The EvOH layer is effective for reducing permeation
emissions. Recent developments in production processes have led to improved barrier coverage
around the ends of the tank where the molded plastic is pinch-welded to form a closed vessel,
which is an important step in eliminating a permeation path through the wall of the fuel tank.
Manufacturers could increase the EvOH barrier thickness to about 3 percent of the average wall
thickness to provide a more uniform barrier layer, to provide better protection with ethanol-based
fuels, and to improve permeation resistance generally. The incremental material cost for this
thicker layer of EvOH comes to about $3.50.
Filler neck connection and materials upgrade: Another area of potential evaporative
emissions is the connection between the fill neck and the fuel tank. The challenge is to design a
low-cost solution that is easily assembled and works for the demanding performance needs
2-25
-------
related to stiffness and flexibility. The best approach is likely either to use mating parts made
from low-permeation materials, or to use conventional materials but cover this joint with
material that acts as a barrier layer. Final designs to address this might vary widely. Technology
to improve the permeation resistance of the fuel filler tube and the security of the connection to
the fuel tank would cost $4-6 per vehicle.
Fuel line permeation: Fuel lines in use today also are designed for low permeation rates.
The biggest portion of fuel and vapor lines are made of metal, but that may still leave several feet
of nonmetal fuel line on a vehicle. There may be development of new materials to further reduce
permeation rates, but it is more likely that manufacturers will adjust the mix of existing types of
plastic fuel lines, and perhaps use more metal fuel lines, to achieve the desired performance at
the lowest possible price. This would likely vary significantly among vehicle models. As an
industry average figure, we estimate upgrades involving $1.60 of additional cost for materials
with greater permeation resistance.
2.4.3 ORVR for Complete HDGVs
Onboard Refueling Vapor Recovery (ORVR): Current EPA standards require vehicle-
based control of refueling emissions for all complete LDVs and LDTs up to 10,000 Ibs GVWR.
We are extending EPA's refueling emission standard to complete heavy-duty gasoline vehicles
(HDGVs) up to 14,000 Ibs GVWR starting with the 2018 model year. Today these HDGVs are
produced by three OEMs. Their chassis and fuel system configurations are very similar to their
slightly lighter GVWR LDT counterparts, which are now covered by the refueling emission
standard. Because annual sales of these 10,001-14,000 Ib GVWR HDGVs is small relative to
their similar lighter GVWR LDT counterparts, for uniformity of production and other cost
savings reasons, manufacturers have installed ORVR on these vehicles since about 2006.
However, they have not been certified since there were no emission control requirements to
certify them against. We are including refueling emission control requirements for these vehicles
but expect no additional costs beyond current practice. Beyond, this the refueling emision
standards apply to all complete HDGVs regardless of their GVWR by the 2022MY. There are no
complete HDGVs above 14,000 Ibs GVWR today, but there have been in the past and if a future
product emerges, this will be a requirement for the 2022 model year.
Table 2-14 presents the evaporative system costs discussed above along with how those
have been weighted to arrive at evaporative system costs for the vehicle categories used
throughout this cost analysis.
2-26
-------
Table 2-14 Evaporative Emission Control System DMC for Gasoline Vehicles (2011$)
Vehicle
Type
LDV
LDT1
LDT2
LDT3
LDT4
MDPV
LHDGV
HHDGV
DMC
$13.15
$12.37
$12.15
$11.47
$11.37
$11.27
$9.75
$15.50
Sales
fraction
100%
17%
57%
17%
8%
1%
100%
100%
Tier3
Cost analysis
Vehicle category
Passenger car
Light truck
Class 2b & 3
>14,000 pound GVWR
Tier3
Cost analysis
DMC
$13.15
$12.00
$9.75
$15.50
We consider these incremental costs to be applicable in MY2015 with flat learning
applied thereafter. We consider evaporative emission controls to be a low complexity
technology with near term factors applied through 2022 and long term thereafter. The resultant
costs are shown in Table 2-15. Note that the values shown do not include phase-in rates. We do
not show costs for diesel vehicles since none of those vehicles are expected to incur new
evaporative emission control costs.
Table 2-15 Evaporative Emission Control System Costs for Gasoline Vehicles (2011$)
Vehicle
category
PC
LT
Class 2b, 3
>14KHD
PC
LT
Class 2b, 3
>14KHD
PC
LT
Class 2b, 3
>14KHD
Cost
DMC
DMC
DMC
DMC
1C
1C
1C
1C
TC
TC
TC
TC
2015
$13
$12
$10
$15
$3
$2
$2
$3
$16
$14
$12
$18
2016
$13
$11
$9
$15
$o
3
$2
$2
$o
3
$15
$14
$11
$18
2017
$12
$11
$9
$14
$3
$2
$2
$3
$15
$14
$11
$17
2018
$12
$11
$9
$14
$3
$2
$2
$3
$14
$13
$11
$17
2019
$11
$10
$8
$13
$o
3
$2
$2
$o
3
$14
$13
$10
$17
2020
$11
$10
$8
$13
$3
$2
$2
$3
$14
$13
$10
$16
2021
$11
$10
$8
$13
$o
3
$2
$2
$o
3
$14
$12
$10
$16
2022
$11
$10
$8
$13
$3
$2
$2
$3
$13
$12
$10
$16
2023
$10
$10
$8
$12
$2
$2
$2
$3
$13
$12
$9
$15
2024
$10
$9
$8
$12
$2
$2
$2
$o
3
$13
$11
$9
$15
2025
$10
$9
$7
$12
$2
$2
$2
$3
$12
$11
$9
$15
PC=passenger car; LT=light truck; DMC=direct manufacturing cost; IC=indirect cost; TC=total cost; 2015 & 2016
costs shown because the cost basis for the technology is 2015 where the learning factoi= 1.0; note that the costs
shown do not include penetration rates so these costs represent technology costs, not package costs for packages of
technologies expected to be applied to Tier 3 vehicles.
2.5 Vehicle Package Costs
The total costs (TC) of a given technology are the direct manufacturing costs (DMC) plus
the indirect costs (1C). These costs change over time due to learning effects and different levels
of indirect costs as discussed above. Here we present our estimated application or penetration
rates for each technology and the subsequent average technology cost estimates by year for each
2-27
-------
technology inclusive of those penetration rates. We then present our approach to developing
package costs—a package being a group of individual technologies added to a given vehicle.
As stated above, we have developed our costs with respect to a given vehicle type and the
type of engine with which it is equipped. Although the cost of achieving the Tier 3 standards
will increase with both the size of the vehicle and the displacement of the engine we have
concluded that the cost by engine type is consistent. The final cost per vehicle is the result of not
only the cost per technology but also the application rate of that technology for each vehicle
type. For example, while the cost of secondary air injection is the same, $119 (2011$) in
MY2017, for both a V6 and V8 application, we anticipate that a lower percentage of V6
applications will require the technology compared to V8 applications. This technology
penetration rate, or application rate, is the first step in developing our vehicle package costs.
Table 2-16 presents our estimates of application rates of each enabling technology by
engine type to meet the Tier 3 standards. These rates are identical, with two exceptions, to the
rates used in the proposal. The changes from the proposal are to the secondary air and the
optimized thermal management technologies. For secondary air, we have used the same starting
rate as used in the proposal, but are now ramping that rate downward in the later years of
implementation. The secondary air application rates are shown in Table 2-17. We are using
these application rates because we believe, based on comments from ICCT and post-proposal
Tier 1 suppliers, that secondary air will follow a similar implementation schedule to past uses of
that technology. In the past, secondary air has been added in the early years of implementation
because it is a very effective and relatively easy to employ technology. As experience is gained,
secondary air is often removed because it is a relatively expensive technology.
Table 2-16 Technology Application Rates for MY2017 and later Passenger Cars
and MY2018 and later Light Trucks and HP Vehicles
Technology
Catalyst Loading
Optimized Close-coupled
Catalyst
Hydrocarbon Adsorber
Evaporative Emissions
Controls
Engine Calibration
Optimized Thermal
Management
SCR Optimization
Gasoline
14
100%
50%
0%
100%
100%
50%
0%
V6
100%
60%
0%
100%
100%
40%
0%
V8
100%
75%
15%
100%
100%
25%
0%
Class 2b,
O
100%
0%
0%
100%
100%
25%
0%
>14K
HD
0%
0%
0%
100%
0%
0%
0%
Diesel
All
0%
0%
0%
0%
100%
25%
100%
Note: 0% entries reflect the fact that the technology is not considered to be an enabler for compliance with
the standards.
2-28
-------
Table 2-17 Technology Application Rates for Secondary Air Injection on Gasoline Vehicles
Vehicle
category
PCV6
PC V8
LTV6
LTV8
2017
25%
75%
0%
0%
2018
25%
75%
25%
75%
2019
25%
75%
25%
75%
2020
15%
65%
15%
65%
2021
15%
65%
15%
65%
2022
5%
55%
5%
55%
2023
5%
55%
5%
55%
2024
5%
45%
5%
45%
2025
5%
45%
5%
45%
PC=passenger car; LT=light truck
MDPVs were included in the light-duty fleet as part of Tier 2. Given their current
certification requirements for criteria pollutants, we have included the costs for MDPVs to meet
the Tier 3 standards with the LT V8 cost estimates. We do not expect that the technologies
required to meet the Tier 3 standards for MDPVs will be very different from those applied to LT
V8s as in many cases there are identical powertrains and chassis between the LT and MDPV
platforms.
The next step in developing vehicle package costs is to consider the phase-in rate of the
standards. For example, the Tier 3 standards do not reach maximum stringency until the 2025
MY, ramping down from a presumed Tier 2 Bin 5 level in MY2016 to the final levels in
MY2025. Manufacturers will be required to start the phase-in of Tier 3 standards on passenger
cars in MY2017 and light trucks in MY2018. Based on the declining fleet averages for cars and
trucks, we have apportioned our estimates for full compliance across of the phase-in years as a
percentage of the final standard. Manufacturers will be required to move from a Tier 2 Bin 5
fleet average in MY2016 (for vehicles <6,000 Ibs GVW) to the Tier 3 standards. This results in
a significant step in stringency in MY2017. It is also important to note that manufacturers will
have the opportunity in MYs 2015 and 2016 to earn Tier 3 credits by producing a fleet that is
cleaner than the current Tier 2 requirements. While we expect that most manufacturers will earn
credits, either by selling California vehicles as 50 state vehicles or by certifying existing vehicles
to lower Tier 2 bins, we have not reflected these credits in our cost analysis.
The ramp down in standards can also be expressed as an increasing percentage of the
fleet meeting the Tier 3 standards, moving from 0 percent compliance in MY2016 to 100 percent
compliance in MY2025 (see Section IV of the preamble, which presents the standards and how
they change by MY). This changing percentage of vehicles complying is treated as being equal
in this analysis to the percentage of costs being incurred. Table 2-18 shows the percentage of
vehicles complying with the new standards. Note that Table 2-18 is identical in content to the
ramp down in standards used in the proposal.
2-29
-------
Table 2-18 Percentage of Vehicles Phasing-in Compliance with the Tier 3 Standards
Vehicle
Category
PC
LT
Class 2b
Class 3
PC
LT
HD
Standards
Exhaust
Exhaust
Exhaust
Exhaust
Evap
Evap
Evap
2017
57%
0%
0%
0%
40%
0%
0%
2018
62%
52%
54%
47%
60%
60%
60%
2019
68%
59%
65%
60%
60%
60%
60%
2020
73%
66%
77%
73%
80%
80%
80%
2021
78%
73%
88%
87%
80%
80%
80%
2022
84%
80%
100%
100%
100%
100%
100%
2023
89%
87%
100%
100%
100%
100%
100%
2024
95%
94%
100%
100%
100%
100%
100%
2025
100%
100%
100%
100%
100%
100%
100%
PC=passenger car; LT=light truck; HD=Class 2b and 3 vehicles, and >14,000 pound gasoline vehicles.
The third step, and new for the final rule, is to consider the compliance rate in the
reference case fleet—i.e., if some vehicles already comply with the Tier 3 standards, we do not
need to add technology and costs to those vehicles. To estimate the reference case compliance
rate, we looked at MY2013 compliance (the most recent full set of compliance data). Filtering
these data to include only those certifications for federal Tier 2 compliance, and combining their
certified NMOG and NOx emissions—Tier 2 vehicles are certified to separate NMOG and NOx
standards, but we combined them as though they were combined standards as are the Tier 3
standards—we were able to determine that 14% of MY2013 passenger car and light truck
certifications had combined NMOG+NOx emissions below the 30 mg level (i.e., the Tier 3 Bin
30 level). That percentage included 33% of passenger car 14 gasoline vehicles and 10% of
passenger car V6 gasoline vehicles. These results are shown in Table 2-19.
Table 2-19 Percentage of MY2013 Certifications with Certified NMOG+NOx Emissions at
the Indicated Tier 3 Levels
Vehicle
Category
PCM
PC V6
PC V8
PCM
PCV6
PC V8
LTM
LTV6
LTV8
LTM
LTV6
LTV8
Total
Fuel
gasoline
gasoline
gasoline
Diesel
Diesel
Diesel
gasoline
gasoline
gasoline
Diesel
Diesel
Diesel
Bin20
25%
5%
2%
0%
0%
0%
4%
3%
0%
0%
0%
0%
10%
Bin30
7%
4%
2%
0%
0%
0%
4%
4%
1%
0%
0%
0%
4%
BinSO
22%
22%
26%
0%
100%
0%
18%
38%
23%
0%
33%
0%
25%
Bin70
37%
46%
41%
100%
0%
0%
36%
31%
34%
100%
67%
0%
38%
Binl25
8%
22%
26%
0%
0%
0%
39%
24%
16%
0%
0%
0%
17%
Binl60
0%
0%
3%
0%
0%
0%
0%
0%
18%
0%
0%
0%
4%
>Binl60
0%
0%
0%
0%
0%
0%
0%
0%
7%
0%
0%
0%
2%
% below
Bin30
33%
10%
3%
0%
0%
0%
7%
7%
1%
0%
0%
0%
14%
Importantly, the percentages shown in Table 2-19 represent certified engine families, not
vehicle sales, and they represent certified emission levels absent any compliance margin. We
have chosen to address these two issues in the following ways. As regards the absence of
2-30
-------
compliance margin—the amount below the standard to which vehicles are typically designed and
certified as insurance against failing the standards in-use—we went the next step and considered
only those certifications that met 70% of the Tier 3 Bin 30 combined NMOG+NOx levels.
Doing this resulted in a total of 11% of the certifications with emissions below 70% of the Bin
30 levels.F This change also resulted in a reduction to 28% of passenger car 14 families and 6%
of passenger car V6 families at 70% of the Bin 30 levels.
As for reconciling certified families with actual sales, we had no way of matching these
certifications to actual sales since the full MY2013 sales were not yet available. We could
simply assume that the percentage of engine families equates to the percentage of sales but, in an
effort to be conservative in our cost estimates, we have chosen instead to assume only a 50%
relationship. In other words, to be conservative, we have chosen to estimate that sales are
represented by only half of the certified engine families. So, the 28% of passenger car 14
families with certified emissions below 70% of the Bin 30 level is taken to represent 14% of
actual passenger car 14 sales. The resultant reference case sales percentages estimated to already
comply with the Tier 3 Bin 30 average are shown in Table 2-20. Note that we have assumed that
no HD vehicles are already at compliant emission levels in the reference case.
Table 2-20 Reference Case Engine Families and Estimated Sales at or below 70% of the
Bin 30 Standard
Vehicle
Category
PCM
PCV6
PC V8
PCM
PC V6
PC V8
LTM
LTV6
LTV8
LTM
LTV6
LTV8
Total
Fuel
gasoline
gasoline
gasoline
diesel
diesel
diesel
gasoline
gasoline
gasoline
diesel
diesel
diesel
% of certified
families
below
Bin30
33%
10%
3%
0%
0%
0%
7%
7%
1%
0%
0%
0%
14%
% of certified
families
below
70% of Bin30
28%
6%
2%
0%
0%
0%
4%
3%
0%
0%
0%
0%
11%
Estimated %
of sales below
70% of Bin30
14%
3%
1%
0%
0%
0%
2%
1%
0%
0%
0%
0%
5%
With each of these percentages—the technology application rate percentage; the phase-in
rate of the standard; and the reference case sales percentage meeting Bin 30—we can then
F A compliance margin of just 70% of the actual standard could be considered too high by traditional measures
where each criteria pollutant has a unique standard level. However, the combined nature of the Tier 3 NMOG+NOx
standard makes traditional compliance margin goals too large and, some have argued, that even 70% in the context
of Tier 3 may be too large. We prefer to be conservative and use 70% of the standard.
2-31
-------
determine the contribution of each individual technology to the resultant package cost for each
vehicle category. This is done by multiplying the total cost of each individual technology in a
given year by it technology application rate for that year, then multiplying this product by the
phase-in rate less reference case sales percentage. An example calculation is shown in Table
2-21 for optimized close coupled catalyst costs on 14 gasoline passenger cars.
Table 2-21 Example Calculation: Contribution of Optimized Close Coupled Catalyst to the
Package Cost for 14 Gasoline Passenger Car (dollar values in 2011$)
Item
Optimized close coupled catalyst Total Cost (TC)
forMY2017
Application rate to meet Tier 3
Standard phase-in percentage for MY2017
Reference case sales in compliance
Contribution to package cost
Value
$24
50%
57%
14%
$5
= ($24)(50%)(57%-14%)
Source
Table 2-8
Table 2- 16
Table 2- 18
Table 2-20
Table 2-22
Table 2-22 through Table 2-30 use this calculation approach to present the contribution
of each technology cost to the resultant package cost. Table 2-31 and Table 2-32 present the
final package costs for gasoline and diesel vehicles, respectively, which simply sum the
appropriate costs shown in Table 2-22 through Table 2-30.
2-32
-------
Table 2-22 Contribution of Individual Technologies to Vehicle Package Costs - Gasoline
Passenger Cars (2011$)
Technology
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls
Engine Calibration
Optimized Thermal
Management
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls
Engine Calibration
Optimized Thermal
Management
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls
Engine Calibration
Optimized Thermal
Management
Engine
14
14
14
14
14
14
14
V6
V6
V6
V6
V6
V6
V6
V8
V8
V8
V8
V8
V8
V8
2017
$18
$5
$0
$0
$4
$2
$8
$37
$15
$16
$0
$5
$3
$8
$57
$30
$50
$2
$6
$3
$5
2018
$20
$6
$0
$0
$7
$2
$8
$39
$16
$17
$0
$8
$o
3
$8
$61
$32
$53
$2
$9
$3
$5
2019
$22
$6
$0
$0
$6
$2
$9
$42
$18
$18
$0
$8
$3
$9
$64
$34
$57
$2
$8
$3
$6
2020
$23
$7
$0
$0
$9
$o
3
$10
$44
$19
$12
$0
$11
$o
3
$9
$68
$36
$52
$2
$11
$3
$6
2021
$25
$7
$0
$0
$9
$3
$10
$47
$20
$12
$0
$10
$3
$10
$72
$38
$55
$2
$11
$3
$6
2022
$27
$7
$0
$0
$11
$o
3
$11
$49
$21
$4
$0
$13
$o
3
$10
$75
$40
$49
$2
$13
$4
$7
2023
$27
$8
$0
$0
$11
$3
$11
$50
$21
$4
$0
$12
$4
$11
$77
$41
$50
$2
$13
$4
$7
2024
$29
$8
$0
$0
$11
$o
3
$12
$52
$22
$5
$0
$12
$4
$11
$80
$42
$42
$2
$12
$4
$7
2025
$30
$8
$0
$0
$11
$o
3
$13
$55
$23
$5
$0
$12
$4
$11
$83
$44
$44
$2
$12
$4
$7
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-16) and/or compliance rate (see
Table 2-18).
2-33
-------
Table 2-23 Contribution of Individual Technologies to Vehicle Package Costs - Gasoline
Light-duty Trucks (2011$)
Technology
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls
Engine Calibration
Optimized Thermal
Management
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls
Engine Calibration
Optimized Thermal
Management
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls
Engine Calibration
Optimized Thermal
Management
Engine
14
14
14
14
14
14
14
V6
V6
V6
V6
V6
V6
V6
V8
V8
V8
V8
V8
V8
V8
2017
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
$0
2018
$25
$6
$0
$0
$8
$2
$9
$37
$14
$15
$0
$8
$2
$7
$55
$27
$45
$2
$8
$2
$5
2019
$28
$6
$0
$0
$7
$3
$10
$41
$16
$16
$0
$8
$3
$8
$61
$30
$50
$2
$8
$3
$5
2020
$30
$7
$0
$0
$10
$3
$11
$45
$17
$11
$0
$10
$o
3
$9
$66
$33
$47
$2
$10
$3
$5
2021
$33
$8
$0
$0
$10
$3
$12
$49
$19
$12
$0
$10
$3
$9
$72
$36
$52
$2
$10
$3
$6
2022
$35
$8
$0
$0
$12
$3
$13
$53
$20
$4
$0
$12
$o
3
$10
$78
$38
$47
$2
$12
$o
3
$6
2023
$37
$9
$0
$0
$11
$3
$13
$55
$21
$4
$0
$11
$3
$10
$81
$40
$49
$2
$12
$4
$7
2024
$39
$9
$0
$0
$11
$4
$14
$58
$22
$5
$0
$11
$4
$11
$86
$42
$42
$2
$11
$4
$7
2025
$41
$10
$0
$0
$11
$4
$15
$61
$23
$5
$0
$11
$4
$12
$90
$44
$44
$2
$11
$4
$7
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-16) and/or compliance rate (see
Table 2-18).
Table 2-24 Contribution of Individual Technologies to Vehicle Package Costs - Gasoline
Heavy-duty Class 2b Trucks (2011$)
Technology
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls
Engine Calibration
Optimized Thermal
Management
Engine
All
All
All
All
All
All
All
2017
$0
$0
$0
$0
$0
$0
$0
2018
$31
$0
$0
$0
$6
$2
$5
2019
$37
$0
$0
$0
$6
$3
$6
2020
$42
$0
$0
$0
$8
$3
$6
2021
$48
$0
$0
$0
$8
$4
$7
2022
$53
$0
$0
$0
$10
$4
$8
2023
$51
$0
$0
$0
$9
$4
$8
2024
$50
$0
$0
$0
$9
$4
$8
2025
$49
$0
$0
$0
$9
$4
$7
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-16) and/or compliance rate (see
Table 2-18).
2-34
-------
Table 2-25 Contribution of Individual Technologies to Vehicle Package Costs - Gasoline
Heavy-duty Class 3 Trucks (2011$)
Technology
Catalyst Loading
Optimized Close-coupled Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions Controls
Engine Calibration
Optimized Thermal Management
Engine
All
All
All
All
All
All
All
2017
$0
$0
$0
$0
$0
$0
$0
2018
$27
$0
$0
$0
$6
$2
$4
2019
$34
$0
$0
$0
$6
$3
$5
2020
$40
$0
$0
$0
$8
$3
$6
2021
$47
$0
$0
$0
$8
$4
$7
2022
$53
$0
$0
$0
$10
$4
$8
2023
$51
$0
$0
$0
$9
$4
$8
2024
$50
$0
$0
$0
$9
$4
$8
2025
$49
$0
$0
$0
$9
$4
$7
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-16) and/or compliance rate (see
Table 2-18).
Table 2-26 Contribution of Individual Technologies to Vehicle Package Costs - >14,000
Pound HD Gasoline Trucks (2011$)
Technology
Catalyst Loading
Optimized Close-coupled
Catalyst
Secondary Air Injection
Hydrocarbon Adsorber
Evaporative Emissions
Controls
Engine Calibration
Optimized Thermal
Management
Engine
All
All
All
All
All
All
All
2017
$0
$0
$0
$0
$0
$0
$0
2018
$0
$0
$0
$0
$10
$0
$0
2019
$0
$0
$0
$0
$10
$0
$0
2020
$0
$0
$0
$0
$13
$0
$0
2021
$0
$0
$0
$0
$13
$0
$0
2022
$0
$0
$0
$0
$16
$0
$0
2023
$0
$0
$0
$0
$15
$0
$0
2024
$0
$0
$0
$0
$15
$0
$0
2025
$0
$0
$0
$0
$15
$0
$0
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-16) and/or compliance rate (see
Table 2-18).
Table 2-27 Contribution of Individual Technologies to Vehicle Package Costs - Diesel
Passenger Cars (2011$)
Technology
Engine Calibration
Optimized Thermal Management
SCR Optimization
Engine
All
All
All
2017
$1
$5
$34
2018
$1
$5
$36
2019
$2
$6
$38
2020
$2
$6
$40
2021
$2
$6
$43
2022
$2
$7
$45
2023
$2
$7
$45
2024
$2
$7
$47
2025
$2
$7
$49
Table 2-28 Contribution of Individual Technologies to Vehicle Package Costs- Diesel
Light-duty Trucks (2011$)
Technology
Engine Calibration
Optimized Thermal Management
SCR Optimization
Engine
All
All
All
2017
$0
$0
$0
2018
$1
$5
$30
2019
$1
$5
$33
2020
$1
$5
$36
2021
$2
$6
$40
2022
$2
$6
$43
2023
$2
$7
$44
2024
$2
$7
$47
2025
$2
$7
$49
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-16) and/or compliance rate
(see Table 2-18).
2-35
-------
Table 2-29 Contribution of Individual Technologies to Vehicle Package Costs- Diesel
Heavy-duty Class 2b Trucks (2011$)
Technology
Engine Calibration
Optimized Thermal Management
SCR Optimization
Engine
All
All
All
2017
$0
$0
$0
2018
$1
$5
$31
2019
$1
$6
$37
2020
$2
$6
$42
2021
$2
$7
$48
2022
$2
$8
$53
2023
$2
$8
$51
2024
$2
$8
$50
2025
$2
$7
$49
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-16) and/or compliance rate
(see Table 2-18).
Table 2-30 Contribution of Individual Technologies to Vehicle Package Costs- Diesel
Heavy-duty Class 3 Trucks (2011$)
Technology
Engine Calibration
Optimized Thermal Management
SCR Optimization
Engine
All
All
All
2017
$0
$0
$0
2018
$1
$4
$27
2019
$1
$5
$34
2020
$2
$6
$40
2021
$2
$7
$47
2022
$2
$8
$53
2023
$2
$8
$51
2024
$2
$8
$50
2025
$2
$7
$49
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-16) and/or compliance rate
(see Table 2-18).
The final package costs are simply the sum of the costs shown in each of Table 2-22
through Table 2-30. These results are shown in Table 2-31 for gasoline vehicles and Table 2-32
for diesel vehicles.
Table 2-31 Vehicle Package Costs by Year for All Gasoline Vehicles (2011$)
Vehicle category
Passenger car
Passenger car
Passenger car
Light-duty truck
Light-duty truck
Light-duty truck
Class 2b
Class 3
>14,000 pound HD
Engine
14
V6
V8
14
V6
V8
2017
$37
$84
$152
$0
$0
$0
$0
$0
$0
2018
$43
$92
$165
$50
$83
$144
$45
$40
$10
2019
$46
$97
$174
$54
$91
$158
$52
$48
$10
2020
$51
$97
$177
$61
$94
$167
$60
$58
$13
2021
$54
$102
$187
$65
$101
$180
$67
$66
$13
2022
$60
$101
$189
$72
$102
$187
$75
$75
$16
2023
$60
$102
$192
$73
$105
$193
$72
$72
$15
2024
$63
$106
$190
$77
$111
$195
$71
$71
$15
2025
$65
$110
$197
$80
$116
$203
$70
$70
$15
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-16) and/or compliance rate (see
Table 2-18).
Table 2-32 Vehicle Package Costs by Year for All Diesel Vehicles (2011$)
Vehicle category
Passenger car
Passenger car
Passenger car
Light-duty truck
Light-duty truck
Light-duty truck
Class 2b
Class 3
Engine
14
V6
V8
14
V6
V8
V8
V8
2017
$40
$40
$40
$0
$0
$0
$0
$0
2018
$43
$43
$43
$36
$36
$36
$37
$32
2019
$46
$46
$46
$40
$40
$40
$44
$40
2020
$48
$48
$48
$43
$43
$43
$50
$48
2021
$51
$51
$51
$47
$47
$47
$57
$56
2022
$53
$53
$53
$51
$51
$51
$63
$63
2023
$54
$54
$54
$53
$53
$53
$61
$61
2024
$56
$56
$56
$56
$56
$56
$60
$60
2025
$59
$59
$59
$59
$59
$59
$59
$59
Note: $0 entries denote zero costs due to a 0 percent application rate (see Table 2-16) and/or compliance rate (see
Table 2-18).
2-36
-------
2.6 Operating Costs
New for the final rule are estimates of operating costs (fuel savings) associated with the
new evaporative emission standards. The fuel that would have evaporated absent the new
standards will ultimately be used to propel the vehicle, thus providing a savings to the consumer.
This savings is very small but nonetheless real. We also considered other operating costs, such
as maintenance costs and repair costs, but concluded that the nature of the Tier 3 compliance
technologies will not result in any increases or decreases in existing operating costs.
In Chapter 1, we discussed at the length the new evaporative standards and the estimated
reductions; hydrocarbon reductions achieved are converted to the fuel saved by each vehicle
category using the equation in footnote 3 of Table 2-13. Those fuel savings in
gallons/vehicle/year are shown in Table 2-33.
Table 2-33 Fuel Savings per Vehicle per Year Associated with the New Evaporative
Emission Standards
Vehicle
type
LDV
LDT1
LDT2
LDT3
LDT4
MDPV
LHDGV
HHDGV
Gallons
saved/vehicle/year
0.051
0.051
0.049
0.051
0.051
0.051
0.059
0.053
Sales
fraction
n/a
17%
57%
17%
8%
1%
n/a
n/a
Vehicle category
Passenger car
Light truck
Class 2b & 3
> 14,000 pound
GVWR
Tier 3 Cost analysis
Gallons
saved/vehicle/year
0.051
0.050
0.059
0.053
Tier 3 compliant vehicles will be expected to realize these fuel savings throughout their
lifetimes. To estimate the lifetime fuel savings, we used the survival fractions shown in Table
2-34, the fuel prices shown in Table 2-35, and the evaporative emission standard phase-in rates
shown in Table 2-18.
2-37
-------
Table 2-34 Vehicle Survival Fractions used in Operating Cost Estimates
Vehicle
age
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
PC
1.000
0.988
0.977
0.961
0.945
0.930
0.911
0.891
0.869
0.840
0.800
0.756
0.706
0.653
0.595
0.531
0.458
0.383
0.308
0.241
0.183
0.139
0.107
0.082
0.063
0.051
0.042
0.034
0.028
0.024
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
LT
1.000
0.978
0.963
0.943
0.931
0.915
0.893
0.870
0.841
0.796
0.742
0.692
0.641
0.583
0.535
0.486
0.442
0.398
0.352
0.309
0.267
0.228
0.202
0.175
0.158
0.145
0.139
0.125
0.111
0.103
0.093
0.083
0.073
0.062
0.050
0.038
0.027
0.000
0.000
0.000
HD
1.000
0.978
0.963
0.943
0.931
0.915
0.893
0.870
0.841
0.796
0.742
0.692
0.641
0.583
0.535
0.486
0.442
0.398
0.352
0.309
0.267
0.228
0.202
0.175
0.158
0.145
0.139
0.125
0.111
0.103
0.093
0.083
0.073
0.062
0.050
0.038
0.027
0.000
0.000
0.000
PC=passenger car; LT=light track;
HD=all Heavy-duty, including >14K
pounds
Source: EPA's MOVES model. For
more information regarding the
MOVES model, see Chapter 7 of this
RIA.
2-38
-------
Table 2-35 Gasoline Prices (2011$)
Calendar
Year
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
Taxed price
($/gal)
$3.15
$3.19
$3.25
$3.32
$3.38
$3.43
$3.45
$3.48
$3.49
$3.52
$3.55
$3.58
$3.63
$3.67
$3.71
$3.75
$3.82
$3.87
$3.94
$4.01
$4.08
$4.15
$4.23
$4.32
Untaxed price
($/gal)
$2.74
$2.79
$2.85
$2.92
$2.98
$3.03
$3.06
$3.09
$3.10
$3.14
$3.17
$3.20
$3.25
$3.29
$3.33
$3.38
$3.44
$3.50
$3.57
$3.64
$3.71
$3.79
$3.87
$3.96
Source: Energy Information Administration, Annual Energy Outlook 2013, Table 59.
Note that the 2040 prices were used for years beyond 2040 in the analysis.
Looking first at the gallons saved during the lifetime of Tier 3 vehicles, the results are
shown in Table 2-36. Clearly, the fuel savings are small on a per vehicle basis, approaching 1
gallon per vehicle during the entire lifetime. Note that the fuel savings reach their maximum in
MY2022 when the phase-in hits 100% and they remain at those levels thereafter. Importantly,
the maximum lifetime savings would actually be realized by any vehicle meeting the new Tier 3
evaporative standards. The lower savings shown in the early years of implementation are the
result of the phase-in. In other words, 40% of the MY2017 passenger cars would realize 0.783
gallons saved during their lifetimes while the remaining 60% would realize none since they
would not be compliant with the new evaporative standards. The resultant savings for the
average MY2017 passenger car would then be (40%)(0.783)+(60%)(0.0)=0.313 gallons.
Nonetheless, while the savings per vehicle are small, they are real and when realized by millions
of vehicles the total gallons saved becomes more meaningful.
2-39
-------
Table 2-36 Gallons of Gasoline Saved during the Lifetimes of Gasoline Vehicles sold in the
indicated MY
Vehicle
category
PC
LT
Class 2b
Class 3
>14KHD
MY
2017
0.313
0.000
0.000
0.000
0.000
MY
2018
0.470
0.491
0.585
0.585
0.524
MY
2019
0.470
0.491
0.585
0.585
0.524
MY
2020
0.626
0.655
0.779
0.779
0.698
MY
2021
0.626
0.655
0.779
0.779
0.698
MY
2022
0.783
0.819
0.974
0.974
0.873
MY
2023
0.783
0.819
0.974
0.974
0.873
MY
2024
0.783
0.819
0.974
0.974
0.873
MY
2025
0.783
0.819
0.974
0.974
0.873
PC=passenger car; LT=light truck; HD=heavy-duty.
We can also look at these gasoline savings on a calendar year basis where, in calendar
year 2017 only the MY2017 vehicles meeting the new evaporative standards will realize any fuel
savings. As shown in Table 2-18 and Table 2-33, this means that only 40% of passenger cars
will realize 0.051 gallons of fuel saved in the 2017 calendar year. But, with gasoline passenger
car sales estimated at 6,184,804 in MY2017, the total gallons saved becomes a more meaningful
number. These results are shown in Table 2-37.
Table 2-37 Annual Gallons of Gasoline Saved
Calendar
Year
2017
2018
2019
2020
2021
2022
2023
2024
2025
2030
PC
124,985
310,678
498,358
751,009
1,002,378
1,315,316
1,625,192
1,933,324
2,239,897
3,661,158
LT
0
107,820
213,383
352,549
487,226
652,291
812,105
967,512
1,118,389
1,800,289
Class 2b
0
14,663
29,055
48,038
66,546
89,471
112,044
134,426
156,504
255,597
Class 3
0
324
642
1,062
1,471
1,978
2,478
2,972
3,461
5,652
>14K HD
0
1,679
3,341
5,557
7,728
10,430
13,089
15,719
18,318
30,125
Total
124,985
435,164
744,779
1,158,215
1,565,350
2,069,486
2,564,908
3,053,954
3,536,569
5,752,821
PC=passenger car; LT=light truck; HD=heavy-duty.
Using the gasoline prices shown in Table 2-35, we can determine the monetary savings
associated with these fuel savings. These results are shown in Table 2-38 (using a 3% discount
rate) and in Table 2-39 (using a 7% discount rate) for the lifetimes of all vehicles sold in each
model year. Table 2-40 presents the annual monetized fuel savings.
Table 2-38 Monetized Lifetime Fuel Savings of all Vehicles Sold in Each Model Year,
Discounted at 3% to the 1st year of the Model Year (Millions of 2011$)
Vehicle
category
PC
LT
Class 2b
Class 3
>14K HD
Sum
MY
2017
$4.66
$0
$0
$0
$0
$4.66
MY
2018
$7.09
$4.22
$0.574
$0.013
$0.066
$12.0
MY
2019
$7.35
$4.29
$0.584
$0.013
$0.068
$12.3
MY
2020
$10.1
$5.76
$0.786
$0.017
$0.092
$16.8
MY
2021
$10.3
$5.77
$0.793
$0.018
$0.093
$17.0
MY
2022
$13.1
$7.17
$0.99
$0.022
$0.117
$21.4
MY
2023
$13.3
$7.16
$1.01
$0.022
$0.119
$21.6
MY
2024
$13.7
$7.19
$1.03
$0.023
$0.121
$22.1
MY
2025
$14.0
$7.24
$1.05
$0.023
$0.124
$22.4
Sum
$93.6
$48.8
$6.82
$0.151
$0.799
$150
Fuel savings calculated with untaxed gasoline prices; PC=passenger car; LT=light truck; HD=heavy-duty.
2-40
-------
Table 2-39 Monetized Lifetime Fuel Savings of all Vehicles Sold in Each Model Year,
Discounted at 7% to the 1st year of the Model Year (Millions of 2011$)
Vehicle
category
PC
LT
Class 2b
Class 3
>14KHD
Sum
MY
2017
$3.47
$0.000
$0.000
$0.000
$0.000
$3.47
MY
2018
$5.27
$3.05
$0.415
$0.009
$0.048
$8.79
MY
2019
$5.47
$3.10
$0.422
$0.009
$0.049
$9.05
MY
2020
$7.51
$4.17
$0.568
$0.013
$0.066
$12.3
MY
2021
$7.69
$4.17
$0.573
$0.013
$0.067
$12.5
MY
2022
$9.73
$5.18
$0.719
$0.016
$0.085
$15.7
MY
2023
$9.9
$5.18
$0.730
$0.016
$0.086
$15.9
MY
2024
$10.2
$5.20
$0.747
$0.017
$0.088
$16.3
MY
2025
$10.4
$5.24
$0.763
$0.017
$0.090
$16.5
Sum
$69.7
$35.3
$4.94
$0.109
$0.578
$111
Fuel savings calculated with untaxed gasoline prices; PC=passenger car; LT=light truck; HD=heavy-duty.
Table 2-40 Annual Monetized Fuel Savings (Millions of 2011$)
Calendar
Year
2017
2018
2019
2020
2021
2022
2023
2024
2025
2030
PC
$0.342
$0.866
$1.42
$2.19
$2.99
$3.99
$4.97
$5.97
$6.95
$12.0
LT
$0.000
$0.301
$0.608
$1.030
$1.45
$1.98
$2.48
$2.99
$3.47
$5.92
Class 2b
$0.000
$0.041
$0.083
$0.140
$0.198
$0.271
$0.343
$0.415
$0.486
$0.841
Class 3
$0.000
$0.001
$0.002
$0.003
$0.004
$0.006
$0.008
$0.009
$0.011
$0.019
>14K HD
$0.000
$0.005
$0.010
$0.016
$0.023
$0.032
$0.040
$0.049
$0.057
$0.099
Total
$0.342
$1.21
$2.12
$3.38
$4.67
$6.28
$7.84
$9.43
$11.0
$18.9
Fuel savings calculated with untaxed gasoline prices; PC=passenger
car; LT=light truck; HD=heavy-duty.
2.7 Vehicle Program Costs
With the package costs presented in Table 2-31 and Table 2-32 and the operating costs
presented in Table 2-40, we can begin to develop vehicle program costs associated with the new
Tier 3 standards. The vehicle program costs multiply package costs by appropriate vehicle sales
per year to estimate the annual technology costs of the program. We then subtract from those
annual technology costs the annual operating savings associated with the evaporative standards.
We also include the annual PM facility costs as discussed below.
The first step to this is determining the projected sales of each vehicle category, or
package, as presented in Table 2-31 and Table 2-32. To do this, we have started with the latest
sales projections from our MOVES database which provides projected sales by passenger car,
light truck, etc., and gasoline versus diesel. However, MOVES does not provide sales
projections to the 14 versus V6 versus V8 level of granularity which we need for Tier 3 vehicle
program costs.
2-41
-------
For a fleet mix breakout at the level needed, we are using the fleet mix projections
stemming from OMEGA runs done in support of our recent GHG final rulemakings.13 We began
with the baseline database developed in support of the MY 2017-2025 GHG final rule.14 That
baseline database provides the fleet sales mix in the years 2017-2025 for each of the vehicle
category/engine/fuel combinations listed in Table 2-31 and Table 2-32. However, that baseline
database is not reflective of the MY 2012 to 2016 or the MY 2017-2025 GHG final rules which
are expected to have an impact on the sales mix of the vehicle category/engine/fuel combinations
largely due to an expectation that engines will be turbocharged and downsized to achieve better
GHG performance while also maintaining vehicle performance and utility. This downsizing is
expected to provide downward effects on overall Tier 3 costs since vehicles with smaller engines
are expected to incur lower costs than vehicles with larger engines. Therefore, using the baseline
database and the technology penetration rates expected from the MY 2017-2025 GHG final rule,
we have developed a Tier 3 reference case fleet. This reference fleet is the fleet we have used in
developing Tier 3 vehicle program costs. Note that the Tier 3 control case fleet and the reference
case fleet are, in effect, one in the same since Tier 3 itself is not expected to have any impact on
the car/truck fleet mix or the I4/V6/V8 fleet mix.
Note that this reference case fleet differs considerably from the reference case fleet used
in the proposal. The proposal used a fleet mix representing a future fleet meeting the MY2016
GHG standards. That fleet mix had considerably less turbocharging and downsizing of engines
since the MY2016 GHG standards were less stringent than the MY 2017-2025 standards
represented in the final rule's reference case fleet.
Table 2-41 shows the baseline fleet mix—representing the best estimates of the future
fleet absent any GHG rules—and Table 2-42 shows the Tier 3 reference fleet—representing the
future fleet in the presence of the MY 2012-2016 and MY 2017-2025 GHG final rules. Table
2-43 shows projected sales of light-duty and heavy-duty vehicles excluding sales in California
and other states that have adopted LEVIII.0
G Vehicle sales in California and other states that have adopted LEVIII are estimated at 36% of the nationwide total.
2-42
-------
Table 2-41 Baseline Light-Duty Fleet Mix
Vehicle
category
PC
LT
EV
PC
LT
a
f
w
14
14
V6
V6
V8
V8
14
14
V6
V6
V8
V8
All
All
"u
Ł
G
D
G
D
G
D
G
D
G
D
G
D
All
All
2017
37.8%
0.0%
21.4%
0.0%
3.2%
0.0%
4.4%
0.0%
22.1%
0.1%
10.7%
0.0%
0.2%
62.6%
37.4%
2018
38.1%
0.0%
21.7%
0.0%
3.1%
0.0%
4.0%
0.0%
23.4%
0.1%
9.3%
0.0%
0.2%
63.2%
36.8%
2019
38.4%
0.0%
21.8%
0.0%
3.2%
0.0%
3.9%
0.0%
23.4%
0.1%
9.0%
0.0%
0.2%
63.6%
36.4%
2020
38.7%
0.0%
22.0%
0.0%
3.2%
0.0%
3.7%
0.0%
23.4%
0.1%
8.6%
0.0%
0.2%
64.2%
35.8%
2021
38.9%
0.0%
22.4%
0.0%
3.2%
0.0%
3.6%
0.0%
23.0%
0.1%
8.5%
0.0%
0.2%
64.7%
35.3%
2022
39.2%
0.0%
22.3%
0.0%
3.2%
0.0%
3.6%
0.0%
23.2%
0.1%
8.2%
0.0%
0.2%
64.9%
35.1%
2023
39.7%
0.0%
22.3%
0.0%
3.1%
0.0%
3.6%
0.0%
23.1%
0.1%
8.0%
0.0%
0.2%
65.3%
34.7%
2024
40.0%
0.0%
22.5%
0.0%
3.1%
0.0%
3.5%
0.0%
23.0%
0.1%
7.5%
0.0%
0.2%
65.9%
34.1%
2025
40.4%
0.0%
22.7%
0.0%
3.2%
0.0%
3.5%
0.0%
22.7%
0.1%
7.2%
0.0%
0.2%
66.5%
33.5%
PC=passenger car; LT=light truck; EV=electric vehicle; G=gasoline; D=diesel.
Table 2-42 Tier 3 Reference Case Light-Duty Fleet Mix
0 0
•fl Ml
•5 2
> s
PC
LT
EV
PC
LT
a
f
w
14
14
V6
V6
V8
V8
14
14
V6
V6
V8
V8
All
All
1
fe
G
D
G
D
G
D
G
D
G
D
G
D
All
All
2017
49.8%
0.3%
9.1%
0.8%
2.3%
0.2%
19.1%
0.0%
11.1%
0.1%
6.9%
0.1%
0.2%
62.6%
37.4%
2018
51.9%
0.2%
7.7%
0.6%
2.2%
0.1%
20.3%
0.0%
10.0%
0.1%
6.4%
0.1%
0.3%
63.1%
36.9%
2019
53.9%
0.2%
6.2%
0.5%
2.1%
0.1%
21.4%
0.0%
8.9%
0.1%
6.0%
0.1%
0.5%
63.5%
36.5%
2020
56.0%
0.1%
4.8%
0.3%
2.1%
0.1%
22.6%
0.0%
7.7%
0.1%
5.5%
0.1%
0.6%
64.0%
36.0%
2021
58.0%
0.1%
3.4%
0.2%
2.0%
0.0%
23.7%
0.0%
6.6%
0.1%
5.1%
0.0%
0.8%
64.4%
35.6%
2022
60.1%
0.0%
1.9%
0.0%
1.9%
0.0%
24.9%
0.0%
5.5%
0.1%
4.7%
0.0%
0.9%
64.9%
35.1%
2023
60.8%
0.0%
1.8%
0.0%
1.6%
0.0%
26.0%
0.0%
4.9%
0.1%
3.7%
0.0%
1.2%
65.4%
34.6%
2024
61.5%
0.0%
1.6%
0.0%
1.3%
0.0%
27.0%
0.0%
4.3%
0.1%
2.7%
0.0%
1.5%
65.9%
34.1%
2025
62.2%
0.0%
1.5%
0.0%
0.9%
0.0%
28.1%
0.0%
3.7%
0.1%
1.7%
0.0%
1.8%
66.4%
33.5%
PC=passenger car; LT=light truck; EV=electric vehicle; G=gasoline; D=diesel.
One very important piece of information shown in the above tables is the gasoline 14
share of the fleet mix—that share being 63% for passenger cars and 29% for light trucks. So,
OMEGA projects that fully 92% of the light-duty fleet will be gasoline 14 by MY2025. With the
exception of electric vehicles, gasoline 14 engines are the least costly of the vehicle categories at
2-43
-------
achieving Tier 3 emission levels. This helps, in part, to explain the large reduction in program
costs (presented below) in this final rule analysis relative to the proposal.
For heavy-duty Class 2b and 3, we expect no downsizing of engines or other changes to
engines that might influence Tier 3 costs as a result of the MY 2014-2018 Heavy-duty GHG rule.
Therefore, we are using the baseline fleet as the reference fleet for this analysis. However, we
have updated the HD baseline fleet relative to the proposal using more recent MOVES data.
Table 2-43 Projected Tier 3 Sales by Year"
Ł fr
75 o
% g>
g S
> O
PCM
PCM
PCV6
PCV6
PCV8
PCV8
LTI4
LTI4
LTV6
LTV6
LTV8
LTV8
EV
PC
LT
Light-
duty
Class
2b
Class
2b
Class 3
Class 3
Class
2b
Class 3
>14K
HD
Heavy-
duty
LD +
HD
73
&
G
D
G
D
G
D
G
D
G
D
G
D
All
All
All
G
D
G
D
G
D
G
All
All
2017
5,195,481
22,542
767,432
65,042
221,891
14,440
2,030,320
0
1,003,041
11,155
644,502
11,102
32,163
6,318,991
3,700,120
10,019,111
413,426
28,228
9,142
97,119
441,654
106,261
53,101
601,015
10,620,126
2018
5,345,425
16,732
617,934
48,787
212,696
10,718
2,124,486
0
879,974
9,982
594,005
8,301
46,553
6,298,846
3,616,747
9,915,593
411,174
27,710
9,092
95,336
438,884
104,428
52,557
595,869
10,511,462
2019
5,621,877
11,301
482,618
33,640
208,494
7,239
2,269,022
0
777,339
9,041
557,387
5,688
62,080
6,427,249
3,618,477
10,045,726
412,765
27,572
9,127
94,863
440,337
103,990
53,216
597,544
10,643,270
2020
5,863,693
5,685
341,259
17,961
202,729
3,642
2,400,216
0
667,197
8,017
516,100
2,985
77,465
6,512,434
3,594,515
10,106,949
410,660
27,320
9,081
93,995
437,980
103,076
53,508
594,563
10,701,513
2021
6,078,373
0
197,189
2,080
195,923
0
2,520,536
0
552,966
6,947
471,969
247
92,581
6,566,146
3,552,665
10,118,811
409,218
27,253
9,049
93,764
436,470
102,812
53,558
592,841
10,711,652
2022
6,127,715
0
181,270
1,852
161,634
0
2,618,264
0
492,601
6,342
368,953
223
120,496
6,595,482
3,486,383
10,081,865
405,730
27,018
8,972
92,955
432,747
101,926
53,349
588,023
10,669,888
2023
6,217,541
0
166,561
1,637
128,438
0
2,733,236
0
435,536
5,779
268,449
202
149,188
6,668,409
3,443,202
10,111,611
407,101
26,979
9,002
92,821
434,080
101,823
53,505
589,408
10,701,019
2024
6,341,466
0
152,573
1,427
95,552
0
2,864,048
0
380,146
5,241
168,242
181
178,997
6,777,644
3,417,858
10,195,502
411,430
27,101
9,098
93,243
438,532
102,341
53,952
594,824
10,790,326
2025
6,481,560
0
138,652
1,217
62,251
0
3,003,581
0
324,535
4,705
66,502
160
209,761
6,903,720
3,399,482
10,303,202
415,032
27,219
9,177
93,649
442,252
102,826
54,514
599,592
10,902,793
PC=passenger car; LT=light truck; EV=electric vehicle; LD=light-duty; HD=heavy-
a Sales exclude vehicle sales in California and other states that have adopted LEVIII.
continue beyond 2025 but are not presented here.
duty; G=gasoline; D=diesel.
, or roughly 36% of the nationwide total; sales
2-44
-------
Using these projected sales, we can calculate the annual costs of the Tier 3 vehicle
program for each vehicle category/engine/fuel combination. We can then add the passenger car
and light-duty truck results to get the costs for light-duty and add the Class 2b, 3 and > 14,000
pound HD costs to get the costs for heavy-duty. We have done this separately for the exhaust
and evaporative standards and then the combined standards. The results are shown in Table
2-45.
In addition to considering the costs associated with improving the emission control
systems on vehicles, we also expect that manufacturers will be required to improve their
capability to measure particulate matter (PM) at the levels being required. For additional
information on the test procedure changes, see Section IV.H of the preamble.
We are using the same PM facility upgrade cost inputs as used in the proposal, except
that we have updated those costs to 2011 dollars. To determine the appropriate costs for
upgrading test facilities for PM measurement we used two sources of information: The first was
the cost that the EPA incurred in upgrading its own PM measurement equipment, and the second
was information provided by vehicle manufacturers reflecting estimates for upgrading their
internal facilities. The cost estimates ranged from $250,000 to $500,000 per PM test site (both in
2010$). We recognize that the number of sites that a manufacturer will require is dependent on
the number of vehicle models it expects to develop and certify in a given model year. As stated
in Section IV. A, we have limited the number of certifications required per model year to 25
percent of the represented durability groups, thereby potentially reducing the number of test sites
that require upgrade. In addition, costs will vary by manufacturer depending on the state of their
current test facilities.
Our estimated costs for each manufacturer are show in Table 2-44. With a certification
responsibility of 25 percent of its given model year durability groups we believe that
manufacturers with annual sales of 1 million units or less will require 2 facility upgrades at an
average cost of $375,000 (in 2010$, or $383,000 in 2011$). For manufacturers with greater than
1 million units per year annual sales we believe that 4 facility upgrades may be required to meet
the Tier 3 requirements.
Table 2-44: PM Facility Costs Imposed by Tier 3 (2011$)
Annual
Sales
Volume
1 million
Fleet
#ofPM
Sites to be
upgraded
2
4
Cost per site
$383,000
Weigh
Room
Costs
Facility
Cost/Manufacturer
$766,000
$1,530,000
#of
Manufacturers
18
5
23
Total Costs
$13,800,000
$7,66,000
$21,400,000
Note that the number of manufacturers in the 1 million
range rather than two in the 1 million range (Chrysler).
We also anticipate that each manufacturer would hire a new full time employee to cover
additional PM measurement-related work. We have estimated this employee to cost each
manufacturer $153,000 (2011$) per year. With 23 manufacturers, the total cost would be
2-45
-------
$3,520,000 per year every year going forward. In contrast, the PM facility costs shown in Table
2-44 represent one-time costs we expect to be incurred in the year prior to implementation of the
standards. These costs are shown in Table 2-45.
Table 2-45 Undiscounted Annual Costs of the Tier 3 Vehicle Program (Millions of 2011$)
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2030
Exhaust
LD
$0
$268
$519
$555
$571
$598
$605
$606
$620
$635
$632
HD
$0
$0
$20.2
$24.2
$27.8
$31.5
$34.8
$00 o
JJ.J
$00 1
33.1
$32.8
$31.8
All
$0
$268
$539
$579
$599
$630
$640
$639
$653
$668
$664
Evaporative
LD
$0
$25.5
$70.2
$69.2
$94.3
$92.8
$116
$111
$109
$108
$108
HD
$0
$0
$3.24
$3.18
$4.12
$4.04
$4.93
$4.73
$4.70
$4.66
$4.56
All
$0
$25.5
$73.4
$72.4
$98.4
$96.8
$121
$116
$114
$113
$113
Operating
LD
$0
$0
-$1.17
-$2.03
-$3.22
-$4.44
-$5.96
-$7.45
-$8.95
-$10.4
-$18.0
HD
$0
$0
-$0.047
-$0.094
-$0.160
-$0.226
-$0.309
-$0.390
-$0.472
-$0.553
-$0.959
All
$0
$0
-$1.22
-$2.12
$o o o
J.JO
-$4.67
-$6.27
-$7.84
-$9.42
-$11.0
-$19.0
Facility
& staff
$21.4
$3.52
$o c^>
3.52
$o c^
3.52
$o c^>
3.52
$o c^
3.52
$o c^>
3.52
$o c^
3.52
$o c^>
3.52
$o c^
3.52
$o c^>
3.52
Total
$21.4
$297
$615
$653
$697
$725
$758
$751
$761
$773
$761
Note: Costs shown include costs for the Tier 3 standards on vehicles sold outside California and other states that
have adopted LEVIII; operating savings use untaxed gasoline prices.
By then sales weighting the exhaust and evaporative results by sales in each of the
vehicle category/engine/fuel combinations, we can calculate the annual technology costs for
passenger cars, light-duty trucks and heavy-duty trucks. We show these cost per vehicle results
for the exhaust standards in Table 2-46, for the evaporative standards in Table 2-47and for the
combined exhaust and evaporative standards in Table 2-48. The costs shown in these three
tables include all direct and indirect costs for new vehicle hardware (they exclude operating
savings and PM facility costs). They also include the effects of learning, and the expected
penetration rates and phase-ins of the Tier 3 standards.
Table 2-46 Cost per Vehicle for the Tier 3 Exhaust Emission Standards (2011$)
Passenger car
Light-duty truck
All light-duty
Class 2b
Class 3
>14KHD
All heavy-duty
All LD and HD
2017
$42
$0
$27
$0
$0
$0
$0
$25
2018
$45
$65
$52
$38
$32
$0
$34
$51
2019
$47
$70
$55
$45
$41
$0
$40
$54
2020
$48
$72
$56
$52
$48
$0
$47
$56
2021
$50
$76
$59
$59
$56
$0
$53
$59
2022
$52
$76
$60
$65
$64
$0
$59
$60
2023
$52
$75
$60
$63
$61
$0
$57
$60
2024
$54
$75
$61
$62
$60
$0
$56
$61
2025
$55
$75
$62
$60
$59
$0
$55
$61
Note: Costs shown include costs for the Tier 3 standards on vehicles sold outside California and other states that
have adopted LEVIII.
2-46
-------
Table 2-47 Cost per Vehicle for the Tier 3 Evaporative Emission Standards (2011$)
Passenger car
Light-duty truck
All light-duty
Class 2b
Class 3
>14KHD
All heavy-duty
All LD and HD
2017
$4
$0
$3
$0
$0
$0
$0
$2
2018
$7
$8
$7
$6
$1
$10
$5
$7
2019
$7
$8
$7
$6
$1
$10
$5
$7
2020
$9
$10
$9
$8
$1
$13
$7
$9
2021
$9
$10
$9
$8
$1
$13
$7
$9
2022
$11
$12
$12
$9
$1
$16
$8
$11
2023
$11
$11
$11
$9
$1
$15
$8
$11
2024
$11
$11
$11
$9
$1
$15
$8
$11
2025
$10
$11
$11
$9
$1
$15
$8
$10
Note: Costs shown include costs for the Tier 3 standards on vehicles sold outside California and other states that
have adopted LEVIII.
Table 2-48 Cost per Vehicle for the Combined Tier 3 Exhaust and Evaporative Emission
Standards(2011$)
Passenger car
Light-duty truck
All light-duty
Class 2b
Class 3
>14KHD
All heavy-duty
All LD and HD
2017
$46
$0
$29
$0
$0
$0
$0
$28
2018
$51
$73
$59
$44
$33
$10
$39
$58
2019
$53
$78
$62
$51
$41
$10
$46
$61
2020
$57
$82
$66
$60
$49
$13
$54
$65
2021
$59
$86
$68
$66
$57
$13
$60
$68
2022
$63
$88
$72
$75
$65
$16
$68
$71
2023
$63
$87
$71
$71
$62
$15
$65
$71
2024
$64
$87
$72
$70
$61
$15
$64
$71
2025
$65
$86
$72
$69
$60
$15
$62
$72
Note: Costs shown include costs for the Tier 3 standards on vehicles sold outside California and other states that
have adopted LEVIII.
2-47
-------
References
1ICCT Comments in Response to the Tier 3 Proposed Rulemaking, Docket ID No. EPA-HQ-OAR-2011-0135-
4304; Posada, Francisco, et. al., "Estimated Cost of Emission Control Technologies for Light-Duty Vehicles Part 1 -
Gasoline," SAE 2013-01-0534, 4/8/2013.
2 California Air Resources Board Initial Statement of Reasons, Public Hearing to Consider LEV III, December 7,
2011, Workshop Document (http://www.arb.ca.gov/regact/2012/leviiighg2012/leviiighg2012.htm). Document ID#
EPA-HQ-OAR-2011-0135-0438.
3 "The Effects of Fuel Sulfur Level onEmissions from Tier 2 Vehicles in the In-Use Fleet," EPA-420-D-13-003.
Available in docket number EPA-HQ-OAR-2011-0135.
4 The 2012-2016 GHG final rule can be found at 75 FR 25374; the 2017-2025 GHG final rule can be found at 77 FR
62624.
5 RTI International. Automobile Industry Retail Price Equivalent and Indirect Cost Multipliers. February 2009.
http://www.epa.gov/otaq/ld-hwy/420r09003.pdf; Rogozhin, A.,et al., "Using indirect cost multipliers to estimate the
total cost of adding new technology in the automobile industry," International Journal of Production Economics
(2009), doi: 10.1016/j.ijpe.2009.11.031. The peer review for the RTI report is at http://www.epa.gov/otaq/ld-
hwy/420r09004.pdf.
6 "Joint Technical Support Document: Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse Gas
Emission Standards and Corporate Average Fuel Economy Standards," EPA-420-R-12-901, August 2012,
Document ID# EPA-HQ-OAR-2011-0135-4948.
7 75 FR 25324, 76 FR 57106, 77 FR 62624, 78 FR 29816.
8 Posada, Francisco, et. al., "Estimated Cost of Emission Control Technologies for Light-Duty Vehicles Part 1 -
Gasoline," SAE 2013-01-0534, 4/8/2013.
9 "Regulatory Impact Analysis: Heavy-Duty Engine and Vehicle Standards and Highway Diesel Fuel Sulfur Control
Requirements," EPA420-R-00-026, December 2000; "Final Regulatory Analysis: Control of Emissions from
Nonroad Diesel Engines," EPA420-R-04-007, May 2004; "Regulatory Impact Analysis: Control of Emissions of
Air Pollution from Locomotive Engines and Marine Compression Ignition Engines Less than 30 Liters Per
Cylinder," EPA420-R-08-001, March 2008.
10 Ball, Douglas and David Moser, "Cold Start Calibration of Current PZEV Vehicles and the Impact of LEV-III
Emission Regulations," SAE 2012-01-1245, 4/16/2012.
11 See ICCT, Docket ID No. EPA-HQ-OAR-2011-0135-4304, at page 4 of 21; see MECA, Docket ID No. EPA-HQ-
OAR-2011-013 5-4675, at page 3 of 7; see CARB, Docket ID No. EPA-HQ-OAR-2011-0135-4621, at page 23.
12 See ICCT, Docket ID No. EPA-HQ-OAR-2011-0135-4304, at page 11 of 21; see MECA, Docket ID No. EPA-
HQ-OAR-2011-0135-4675, at page 3 of 7.
13 OMEGA is the Optimization Model for reducing Emissions of Greenhouse gases from Automobiles. Information
about OMEGA can be found at http://www.epa.gov/otaq/climate/models.htm.
2-48
-------
14 See Chapter 1 of "Regulatory Impact Analysis: Final Rulemaking for 2017-2025 Light-Duty Vehicle Greenhouse
Gas Emission Standards and Corporate Average Fuel Economy Standards," EPA-420-R-12-016, August 2012.
2-49
-------
Chapter 3 Establishing New Emission Test Fuel Parameters
In-use gasoline has changed considerably since EPA's emission test gasoline
specifications were first set and last revised. Sulfur and benzene content have been reduced and,
perhaps most visibly to consumers, gasoline containing 10 percent ethanol by volume (E10) has
replaced non-oxygenated gasoline (EO) across the country. The relationship between emissions
certification test fuel and in-use fuel is important in recognition of the fact that fuel properties
can affect emission levels. Therefore, in revising specifications for emission test gasoline, it is
important to have a thorough assessment of fuel available to the public.
We primarily used two sources of fuel property information to determine appropriate
specifications for emissions test fuel. One was the Reformulated Gasoline and Anti-Dumping
Batch Report data submitted to EPA (referred to in this section as batch data). Producers and
importers of gasoline and related blendstocks must submit data to EPA for each batch of gasoline
produced or imported. These data include batch volume as well as physical and chemical
properties that can be used to determine whether the fuel is compliant with applicable standards
and regulations. These reports are considered Confidential Business Information and thus only
aggregated data is presented here.
The second data source was the Alliance of Automobile Manufacturers (AAM) North
American Fuel Survey. Each summer and winter, the AAM collects over 300 gasoline samples
from retail stations in 29 major metropolitan areas in 23 states plus the District of Columbia.
Areas currently sampled include: Albuquerque, NM; Atlanta, GA; Billings, MT; Boston, MA;
Chicago, IL; Cleveland, OH; Dallas, Houston, and San Antonio, TX; Denver, CO; Detroit, MI;
Fairbanks, AK; Kansas City and St. Louis, MO; Las Vegas, NV; Los Angeles and San Francisco,
CA; Memphis, TN; Miami, FL; Minneapolis/St. Paul, MN; New Orleans, LA; New York, NY;
Philadelphia and Pittsburgh, PA; Phoenix, AZ; Salt Lake City, UT; Seattle, WA; Washington,
D.C.; and Watertown, SD. Although the AAM North American Fuel Survey does not represent
all U.S. gasoline, it is designed to have good coverage of the U.S. market.
Note that this assessment focuses on fuel properties for summertime, regular grade, E10
gasoline since this is most relevant to the certification testing conditions and fuel specifications.
3.1 Assessment of Current Gasoline Properties
3.1.1 Ethanol Content
According to the Energy Information Administration (EIA), ethanol is now blended into
almost every gallon of U.S. gasoline, bringing the average gasoline ethanol content to 9.7 percent
denatured ethanol by volume (vol%) as shown in Figure 3-1. Denaturant, generally a
hydrocarbon blendstock such as natural gas liquids or low-octane gasoline components, is added
at a rate of approximately 2 volume percent by the ethanol producer before shipping it to fuel
terminals. (This is required to differentiate the product from beverage alcohol.)
3-1
-------
Avg.% Ethanol in Gasoline
by Volume
Year
Figure 3-1. Denatured Ethanol Content in U.S. Gasoline Over Time
The plot shows a rapid increase in ethanol content starting around 2002 and leveling out
after 2010 as it approaches 10 percent volume (this average figure also includes a small amount
of E85 use in Flex Fuel vehicles). While EPA has approved use of E15 in gasoline vehicles of
model year 2001 and later, its use has not yet become widespread.A
Figure 3-2 shows the distribution of ethanol levels across 404 regular grade summer E10
gasoline samples collected by AAM in 2010 and 2011. These data suggest the range of ethanol
blending in E10 is relatively narrow, with most samples falling between 9.3 and 10.0 volume
percent (excluding denaturant). Higher and lower values in the data are likely due to test method
uncertainty. Based on this information, an emission test fuel target of 9.8 volume percent
ethanol (excluding denaturant) is appropriate.
1 The E15 approval decision was published in 76 FR 4662 (January 26, 2011).
3-2
-------
20%
15%
Q.
ra
10
•s 10%
I
01
Q.
5%
0%
8.5 8.7 8.9 9.1 9.3 9.5 9.7 9.9 10.1
Volume Percent Ethanol (ASTM D5599)
10.3 10.5
Figure 3-2. Range of Ethanol in E10 Gasoline in 2010-11 AAM Summer Surveys
3.1.2 Octane
U.S. gasoline must meet a minimum octane rating (also known as (R+M)/2 or anti-knock
index, AKI) of 87 for regular grade in most parts of the country. Denatured fuel ethanol has a
typical octane rating of 115 AKI, making it a high-octane blendstock. However, finished
gasoline has not experienced an increase in octane due to increased ethanol blending. Given this
situation, along with data presented in the next subsection, it is evident that many refiners have
backed off on octane production at the refinery by reducing levels of aromatics and olefms.
Producing these high-octane components at the refinery represents a significant cost to refiners,
so they are able to reduce costs by taking advantage of ethanol's octane value. We estimate that
many refiners are currently producing 84-85 AKI blendstocks for 87 octane regular-grade
gasoline and 88-89 AKI blendstocks for 91 octane premium-grade gasoline, such that the final
E10 blends meet minimum octane requirements.
According to AAM summer fuel surveys, the average octane of finished regular grade
gasoline has remained constant between 87-88 AKI over the past decade (refer to Figure 3-3)
despite the increasing blend level of ethanol. According to EIA's Petroleum Marketing Annual,
regular grade gasoline represents over 85 percent of U.S. sales.2 Accordingly, we believe the
updated 87-88.4 (R+M)/2 test fuel specification is representative of regular grade in-use
gasoline.
3-3
-------
100
90
80
70
I
tt
O
87.1 86.9 87.0 86.9 87.0 87.2 87.2 87.4 88-1 87.4 87.8 87.4
-Avg. Octane (AAM) » Avg. EtOH (EIA) |
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Figure 3-3. Average Summer Regular Grade Octane and Ethanol Levels Over Time
3.1.3 Total Aromatics and Total Olefms
The term olefin describes a hydrocarbon compound containing at least one unsaturated or
double bond. Aromatics are a specific class of olefms that contain the benzene ring. When
crude oil is distilled into various fractions according to boiling range, the fraction going into the
gasoline pool, called straight run naphtha, contains primarily saturated hydrocarbons. Both
olefms and aromatics have higher AKI values relative to saturates, and therefore increasing their
proportions in the finished gasoline is an important method refiners use to meet required AKI
targets.
Ethanol also has a high AKI value, and as it has become more ubiquitous as a blendstock,
refiners are relying on it to an increasing extent to meet octane targets. The average aromatics,
olefin, and ethanol levels by year for all summer gasoline are shown in Figure 3-4. Here we can
see a general trend of aromatic and olefin levels declining as fuel ethanol content increased.
Using 2010-11 AAM survey data, the average aromatics content of conventional regular grade
E10 gasoline (172 samples) was 24.3 vol% and the average olefms content was 7.3 vol%. When
interpreting these aromatics results as a basis for updating test fuel specifications, we considered
3-4
-------
the fact that ASTM D1319 (used for the AAM surveys) gives a numerical value between 1-2
vol% higher than ASTM D5769 (specified for test fuel analysis) for the same fuel sample.6
30
Si
$ 15
o
o
10
•Avg. Corrected Anomalies (AAM)
• Avg. Corrected Olefins (AAM)
Avg.EtOH (EIA)
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
Figure 3-4. Average Summer Gasoline Aromatics, Olefin and Ethanol Levels Over Time
Although total aromatics and olefins have been reduced over the past decade, there
continues to be variation on a batch-by-batch and geographic basis. Our refinery batch data for
summer 2011 shows a range of gasoline aromatics levels from approximately 5 to 50 vol% with
an average concentration of 24 vol% (Figure 3-5), and a range of olefm levels from 0 to 25 vol%
with an average concentration of 11 vol% (Figure 3-6). The 2011 batch data shown in Figures 3-
5 and 3-6 only reflects the effect of ethanol in reformulated gasoline that is match blended to
account for it. It does not account for ethanol that may be blended into conventional gasoline
after it leaves the refinery since existing gasoline regulations do not readily allow refiners to take
advantage of ethanol properties in most compliance calculations. As a result, AAM and other
gasoline surveys may show lower aromatics and olefins than what is suggested by the batch data.
' Information based on analysis of several recent ASTM cross-check datasets.
3-5
-------
60%
0%
20% 40% 60% 80%
Cumulative Gasoline Volume (% Total)
100%
Figure 3-5. Gasoline Aromatics Distribution Based on Summer 2011 Refinery Batch Data
0%
20% 40% 60% 80%
Cumulative Gasoline Volume (% Total)
100%
Figure 3-6. Gasoline Olefins Distribution Based on Summer 2011 Refinery Batch Data
In the summer of 2010, according to the AAM North American Fuel Survey, measured
in-use aromatics levels ranged from 3 to 47 vol% (Figure 3-7) while olefin levels ranged from
3-6
-------
0.6 to 17 vol% (Figure 3-8). California tends to have lower, tighter in-use levels of aromatics
and olefms as a result of their more stringent fuel regulations. As shown below, gasoline
samples taken from Los Angeles and San Francisco had aromatics levels ranging from 10 to 30
vol% and olefm levels ranging from 1 to 8 vol%. Nevertheless, our updated emission test fuel
specifications for aromatics and olefms still overlap with those established by the California Air
Resources Board (CARS) for LEV III test fuel.c
Cities Sampled, Summer 2010
Figure 3-7. Range of Total Aromatics by AAM City Surveyed, Summer 2010
California LEV III emission test procedures, including fuel specifications, are available at 13 CCR 1961.2.
3-7
-------
o
•o
^ & Ł ^ *- o* * ^ * ^° ^ /S&S/&S/
Cities Sampled, Summer 2010
Figure 3-8. Range of Olefins by AAM City Surveyed, Summer 2010
3.1.4 Aromatics Species
Gasoline speciation data performed by EPA and others shows a wide range of aromatic
molecule configurations from benzene (C6, meaning it contains six carbon atoms), and toluene
(C7), to larger more complicated C10+ aromatics. Between 2007 and 2011, EPA performed
aromatic speciation analyses on 52 fuel samples from various locations throughout the country.
Approximately 60 percent were RFG oversight samples (supplied by refiners as part of the RFG
program) and the remainder were audit samples collected mostly from retail outlets as part of the
City Surveys provision in the RFG program. Total aromatics ranged from 6 to 39 percent of
gasoline by volume, but the relative proportions of molecular species by carbon number were
relatively consistent across the samples (Figure 3-9).
3-8
-------
80%
u
i
M-
_O
O
S
Q.
o% -
"< 25 X
--.z5fSz:
JigS
gg
-------
Since aromatics do not appear to be created equally in terms of the potential impact on
vehicle PM emissions, we believe it is prudent that both the amount and distribution of aromatics
in the updated emissions test fuel is representative of in-use fuel. Figure 3-10 shows averages
and ranges of EPA aromatics speciation data with test fuel specification ranges shown in blue.
Proportions adjusted to total aromatics of 23.0 vol%
Groups Included:
Q - Mono-a rom at ics
G = Naphthalenes
H-Napthleneo/Olefi no-
Benzenes
J-Indenes
C6(Q)
C7(Q)
C8(Q)
C9 (OJ)
C10+ (Q,G,H,J)
Figure 3-10. Mean and Standard Deviation of Aromatics Speciation Data with Test Fuel
Specification Ranges Shown in Blue
3.1.5 Distillation Temperatures
As shown below in Figure 3-11, AAM survey data suggests there has not been a large
change in gasoline volatility curves for summer gasoline over the past decade as ethanol
blending has increased. The T50 and T90 temperatures are treated in more detail in subsections
below.
3-10
-------
500
400
-300
o
|
ts
5
a
_c
1 200
ro
13
100
329
407 409 411 408
328 329 329
-* * *-
319
208
«—
136 136 138 138 137
137 137 136
207 205
135 134 133
-Avg.TIO (AAM) —»-Avg.T50(AAM) -A-Avg.T90 (AAM) -*-Avg. FBP (AAM) |
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Figure 3-11. Summer Gasoline Distillation Temperatures Over Time From AAM Surveys
3.1.5.1 T50 analysis
Splash-blending 10 percent ethanol in an EO base typically lowers T50 by several degrees
relative to the base gasoline. Given that much of the refinery batch data for conventional
gasoline in 2011 did not capture the property changes resulting from ethanol blending, we would
expect the curve in Figure 3-12 to be shifted downward, putting it in closer agreement with the
AAM survey data average of around 202°F as shown in Figure 3-11. The AAM survey data in
Figure 3-13 shows that T50 varies widely in in-use fuel, from around 150°F to 220°F. Plotting
T50 by RVP reveals that T50 values span a higher and narrower range in reformulated and
volatility-control-area fuels (on the left, below about 8.5 psi) compared to conventional gasoline
(above about 9 psi). Adopting a wide specification for test fuel to cover this whole range may
have undesirable effects on consistency of vehicle test results between facilities and over time.
Therefore, we have chosen a range of 190-210°F to represent both fuel types and maintain some
overlap with CARB's specification of 205-215°F.
3-11
-------
o
in
260
240
220
200
180
160
140
0%
20% 40% 60% 80%
Cumulative Gasoline Volime (% Total)
100%
Figure 3-12. Gasoline T50 Range Based on Summer 2011 Refinery Batch Data
tz*
M
OJ
^
o
p
~> ~>r\ -
iyu
loU
1 fin -
•
_4JE /mi |.
ti|||s! :! •ti* »** * *
*ilt ** . ,
»*• *•*!» * * *
.::||««*!|.. : .:
: ':« •• * :'»
«•:•: ":; «,.
i f I
7 8 9 10 11
RVP(psi)
Figure 3-13. T50 Range in 2010-11 Summer Gasoline as Reported in AAM Surveys
3-12
-------
3.1.5.2 T90 Analysis
To develop an understanding of the variation of T90 over the pool of gasoline produced
or imported during the 2011 averaging period, we plotted T90 versus cumulative gasoline
volume. Approximately 90% of the T90 data is linearly distributed along the center portion of
the plot. The remaining 10% of the T90 data is comprised of outliers, with the lower and upper
end of the temperature spectrum tailing off slightly. The volume-weighted average T90 was
325°F (see Figure 3-14). T90 is relatively insensitive to ethanol blending at the 10 percent level,
and therefore we see good agreement between the batch and AAM datasets on average value.
Figure 3-15 shows AAM T90 data plotted by RVP, where there appears to be much less
influence difference between reformulated and conventional fuel pools. Based on this
information, we set the test fuel specification range at 315-335°F.
400
380
360
iT 340
o
320
300
280
260
0%
20% 40% 60% 80%
Cumulative Gasoline Volime (% Total)
100%
Figure 3-14. Gasoline T90 Range Based on Summer 2011 Refinery Batch Data
3-13
-------
370
360
340
M
OJ
2. 320
o
H 310
300
290
280
270
Figure 3-15. T90 Range in 2010-11 Summer Gasoline as Reported in AAM Surveys
3.1.6 Sulfur and Benzene
Gasoline sulfur levels have declined significantly over the past decade under the Tier 2
gasoline program.0 The phase-in period of those standards began in 2004 and continued until
2011, when all geographic and small refiner relief provisions ended. According to AAM
summer fuel surveys, average gasoline sulfur has gone from over 150 ppm in 2000 to less than
30 ppm (the Tier 2 average standard) in 2012 (refer to Figure 3-16). Refinery batch reports for
2011 (refer to Figure 3-17) depict a volume-weighted average gasoline sulfur just below 30 ppm.
Again, given that refinery certification data does not include all ethanol blended into
conventional gasoline, the average sulfur content in-use is expected to be slightly lower, which is
consistent with the AAM data.
After the phase-in of the Tier 3 sulfur limits, gasoline sulfur levels are required to fall to
10 ppm. Sulfur naturally occurs in crude oil and most refineries must spend money to install and
operate units that remove it from gasoline. This sulfur byproduct of refining has little market
value itself, so significant overcompliance with this standard is not expected. Accordingly, the
updated test fuel sulfur specification is being set to a range of 8-11 ppm.
D The Tier 2 final rulemaking was published in 65 FR 6698 (February 10, 2000).
3-14
-------
Gasoline benzene levels have also fallen in recent years primarily due to the MSAT2
program, which enacted an annual average standard of 0.62 volume percent benzene across all
gasoline effective January, 1, 2011 .E According to AAM summer fuel surveys, average gasoline
benzene content has declined from almost 1 vol% in 2006 to less than 0.7 vol% in 2011 (refer to
Figure 3-16). This is in general agreement with refinery batch reports where volume-weighted
average benzene was less than 0.6 vol% in summer 2011 (refer to Figure 3-18). Again, given
that refinery certification data does not include all ethanol blended into conventional gasoline, it
is reasonable that the average benzene content shown in the AAM data is a bit lower than
suggested by the batch reports.
Some benzene naturally occurs in crude oil, but the majority that ends up in finished
gasoline is produced during refinery operations intended to increase the total aromatics content
to meet octane requirements. Therefore most refineries must spend money to install and operate
units that remove benzene from certain blendstock streams before the finished gasoline is made.
In some areas of the country (such as the Gulf Coast), benzene has significant value as a
chemical feedstock and may be extracted from gasoline at a rate that is greater than would
otherwise be required to meet fuel regulations. In most areas of the country, however, meeting
the gasoline benzene limits is the sole driver of any reduction process, and therefore due to the
averaging, banking, and trading provisions in the regulations we don't expect significant
overcompliance on a nationwide basis. Therefore we believe an emissions test fuel benzene
specification of 0.5-0.7 vol% is representative of in-use gasoline now and going forward. These
benzene and sulfur specifications are consistent with CARB's LEV III specifications.
'• The MSAT 2 final rulemaking was published in 72 FR 8428 (February 26, 2007).
3-15
-------
180
160
Avg. Sulfur (AAM) —^~Avg. Benzene (AAM)
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Figure 3-16. Average Summer Sulfur and Benzene Levels Over Time from AAM Surveys
3-16
-------
80
70
60
E 50
Q.
Q.
r 40
cr>
30
20
10
0%
20% 40% 60% 80%
Cumulative Gasoline Volime (% Total)
100%
Figure 3-17. Gasoline Sulfur Based on Summer 2011 Refinery Batch Data
5%
4%
o 3%
0)
c
0)
N TO/
CO
1%
0%
0%
20% 40% 60% 80%
Cumulative Gasoline Volume (% Total)
100%
Figure 3-18. Gasoline Benzene Based on 2011 Refinery Batch Data
3-17
-------
3.2 Gasoline Emission Test Fuel Specifications
As explained in Section IV.F of the preamble, we are updating federal emission test fuel
specifications to better match in-use fuel. The revised test fuel specifications apply for exhaust
emissions testing, fuel economy/greenhouse gas testing, and emissions testing for non-exhaust
emissions (evaporative, refueling, and leak detection testing). The revised gasoline
specifications, found at §1065.710 and shown here in Table 3-1, apply to emissions testing of
light-duty cars and trucks as well as heavy-duty gasoline vehicles certified on the chassis test,
those subject to the Tier 3 standards. For information on how we arrived at the revised ASTM
test procedures, refer to Section 3.3. Commercial gasoline or "street fuel" would continue to be
used for service accumulation (durability fuel). This is consistent with CARS's LEV III
approach and should help limit the total number of test fuels that automakers need to manage.
3-18
-------
Table 3-1. Gasoline Emission Test Fuel Specifications
Property
Antiknock Index
(R+M)/2
Sensitivity (R-M)
Dry Vapor Pressure
Equivalent (DVPE)M
Distillation4
10% evaporated
50% evaporated
90% evaporated
Evaporated final
boiling point
Residue
Total Aromatic
Hydrocarbons
C6 Aromatics
(benzene)
C7 Aromatics
(toluene)
C8 Aromatics
C9 Aromatics
C 10+ Aromatics
Olefins5
Ethanol blended
Ethanol confirmatory
Total Content of
Oxygenates Other than
Ethanol6
Sulfur
Lead
Phosphorus
Copper Corrosion
Solvent- Washed Gum
Content
Oxidation Stability
Unit
-
-
kPa (psi)
°C (°F)
°f^ ('°l-i1^
°C (°F)
°C (°F)
milliliter
volume %
volume %
volume %
volume %
volume %
volume %
mass %
volume %
volume %
volume %
mg/kg
g/liter
g/liter
-
mg/100
milliliter
minute
SPECIFICATION
General „ , High Altitude
,,, ,. Temperature %, ,.
Testing _±\. Testing
fe Testing fe
87.0 - 88.42 87.0 Minimum
7.5 Minimum
60.0-63.4 77.2-81.4 52.4-55.2
(8.7-9.2) (11.2-11.8) (7.6-8.0)
49-60 43-54 49-60
(120-140) (110-130) (120-140)
88-99(190-210)
157-168(315-335)
193-216(380-420)
2.0 Maximum
21.0-25.0
0.5-0.7
5.2-6.4
5.2-6.4
5.2-6.4
4.4-5.6
4.0-10.0
9.6-10.0
9.4-10.2
0.1 Maximum
8.0-11.0
0.0026 Maximum
0.0013 Maximum
No. 1 Maximum
3.0 Maximum
1000 Minimum
Reference
Procedure1
ASTM D2699 and
D2700
ASTM D2699 and
D2700
ASTM D5 191
ASTMD86
ASTM D5769
ASTMD6550
See
§1065.710(b)(3)
ASTMD4815or
D5599
ASTMD4815or
D5599
ASTM D2622,
D5453orD7039
ASTMD3237
ASTMD3231
ASTM Dl 30
ASTM D3 81
ASTM D525
ASTM procedures are incorporated by reference in §1065.1010. See § 1065.70l(d) for other allowed procedures.
2Octane specifications apply only for testing related to exhaust emissions. For engines or vehicles that require the
use of premium fuel, as described in paragraph (d) of this section, the adjusted specification for antiknock index is a
minimum value of 91.0; no maximum value applies. All other specifications apply for this high-octane fuel.
3Calculate dry vapor pressure equivalent, DVPE, based on the measured total vapor pressure, />T, using the following
equation: DVPE (kPa) = 0.956-^T - 2.39 (or DVPE (psi) = 0.956-^T - 0.347. DVPE is intended to be equivalent to
Reid Vapor Pressure using a different test method.
^Parenthetical values are shown for informational purposes only.
3-19
-------
The reference procedure prescribes measurement of olefin concentration in mass %. Multiply this result by 0.857
and round to the first decimal place to determine the olefin concentration in volume %.
6The reference procedure prescribes concentration measurements for ethanol and other oxygenates in mass %.
Convert results to volume % as specified in Section 14.3 of ASTM D4815.
Along with updated emission test fuel parameters, we are adding specifications for
distillation residue, total content of oxygenates other than ethanol, copper corrosion, solvent-
washed gum content, and oxidation stability. These parameters, summarized in Table 3-1, are
consistent with ASTM D-4814 gasoline specifications and CARS's LEV III test fuel
requirements.
3.3 Changes to ASTM Test Methods
Many of the test methods specified in 40 CFR 86.113 for gasoline used in exhaust and
evaporative emission testing have been retained in 40 CFR 1065.710 test fuel specification for
ethanol-blended gasoline. However, some test methods have been replaced with methods
deemed more appropriate, easier to use, or more precise. The following paragraphs highlight the
new reference methods.
ASTM D323 "Standard Test Method for Vapor Pressure of Petroleum Products (Reid
Method)" is not applicable to ethanol-blended gasoline. It is being replaced with an automated
ASTM D5191 "Standard Test Method for Vapor Pressure of Petroleum Products (Mini
Method)," which is appropriate for ethanol-blended gasoline.
ASTM D1319 "Standard Test Method for Hydrocarbon Types in Liquid Petroleum
Products by Fluorescent Indicator Adsorption" is required by 40 CFR 86.113 for use in the
measurement of aromatics and olefms. It is being replaced with ASTM D5769 "Standard Test
Method for Determination of Benzene, Toluene, and Total Aromatics in Finished Gasolines by
Gas Chromatography/Mass Spectrometry" and ASTM D6550 "Standard Test Method for
Determination of Olefin Content of Gasolines by Supercritical-Fluid Chromatography." Method
D5769 enables simultaneous determination of the total aromatic hydrocarbon content, carbon
number-specific content, and benzene content and is already being used in reformulated gasoline
applications. ASTM D1319 does not identify aromatics by carbon number, which is now
required for the Tier 3 test fuel in 40 CFR 1065.710. In addition, ASTM D5769 and D6550 are
more precise and less labor-intensive than ASTM D1319.
Measurement of oxygenates, including ethanol, is being updated to allow two methods
that produce equivalent results: ASTM D4815, "Standard Test Method for Determination of
MTBE, ETBE, TAME, DIPE, tertiary-Amyl Alcohol and Cl to C4 Alcohols in Gasoline by Gas
Chromatography" and ASTM D5599, "Standard Test Method for Determination of Oxygenates
in Gasoline by Gas Chromatography and Oxygen Selective Flame lonization Detection".
For sulfur measurements, ASTM D1266 "Standard Test Method for Sulfur in Petroleum
Products (Lamp Method)" is being replaced with three automated methods: ASTM D2622
"Standard Test Method for Sulfur in Petroleum Products by Wavelength Dispersive X-ray
Fluorescence Spectrometry", ASTM D5453 "Standard Test Method for Determination of Total
3-20
-------
Sulfur in Light Hydrocarbons, Spark Ignition Engine Fuel, Diesel Engine Fuel, and Engine Oil
by Ultraviolet Fluorescence" and ASTM D7039 "Standard Test Method for Sulfur in Gasoline
and Diesel Fuel by Monochromatic Wavelength Dispersive X-ray Fluorescence Spectrometry."
These three new methods are significantly less labor-intensive than ASTM D1266 and are
widely used in the measurement of sulfur content in petroleum products.
3-21
-------
References
1 EIA, January 2012 Monthly Energy Review, Table 3.7a, Petroleum Consumption: Residential and Commercial
Sectors; Table 3.7b, Petroleum Consumption: Industrial Sector; Table 3.7c, Petroleum Consumption: Transportation
Sector, and Table 10.3, Fuel Ethanol Overview.
2 EIA, Petroleum Marketing Annual 2009, Table 45, Prime Supplier Sales Volume of Motor Gasoline by Grade,
Formulation, PAD District and State.
3 lizuka, Masashi, Advanced Technology and Research Institute (ATRI) and Japan Petroleum Energy Center
(JPEC), Effect of Fuel Properties on Emissions from Direct Injection Gasoline Vehicle.
4 Jetter, Jeff, Honda R&D America Inc., Development of a Predictive Model for Gasoline Vehicle Particulate Matter
Emissions, SAE 2010-01-2115.
3-22
-------
Chapter 4 Fuel Program Feasibility
4.1 Overview of Refining Operations
Figure 4-1 shows a process flow diagram for a typical complex refinery, capable of
making a wide product slate (shown on the right side of the figure) from crude oil (input on the
left). Following the figure is a brief description of key units and streams focusing more on the
Natural
Gas
Hydrogen Plant
Fuel Gas
LPG
Gasoline
Aromatics
Kerosene
Jet Fuel
Low Sulfur Diesel
Off-road Diesel
Heating Oil
Resid
Coke
Figure 4-1 Process Flow Diagram for a Typical Complex Refinery
Crude Tower
The purpose of the crude tower is to perform a distillation separation of crude oil into
different streams for additional processing in the refinery and for the production of specific
products. Crude oil is shipped to the refinery via pipeline, ship, barge, rail, or truck, whereupon
it is sampled, tested, and approved for processing. The crude oil is heated to between 650 °F and
700 °F and fed to crude distillation tower. Crude components vaporize and flow upward through
the tower. Draw trays are installed at specific locations up the tower from which desired side
cuts or fractions are withdrawn. The first side-cut above the flash zone is usually atmospheric
gasoil (AGO), then diesel and kerosene/jet fuel are the next side-cuts, in that order. The lightest
4-1
-------
components, referred to here as straight run naphtha, remain in the vapor phase until they exit the
tower overhead, following which they are condensed and cooled and sent to the naphtha splitter.
Naphtha Splitter
The purpose of the naphtha splitter is to perform a distillation separation of straight run
naphtha into light straight run naphtha and heavy straight run naphtha. The feed can be split
between the C5's and C6's in order to assure the C6's and heavier are fed to the reformer.
Naphtha Hydrotreater
The purpose of the naphtha hydrotreater is to reduce the sulfur of light and heavy straight
run streams before those streams are refined further by the isomerization and reformer units.
Isomerization Unit
The purpose of the isomerization unit is to convert the light naphtha from straight chain
hydrocarbons to branched chain hydrocarbons, increasing the octane of this stream. The
isomerate is sent to gasoline blending.
Reformer
The purpose of the reformer unit is to convert heavy straight run (C6 to C8 or C9
hydrocarbons) into aromatic and other higher octane compounds (benzene is one of the aromatic
compounds produced), typically necessary to produce gasoline with sufficient octane. To protect
the very expensive, precious metal catalyst used in reformers, heavy straight run naphtha must be
hydrotreated first before it is fed to the reformer. As the reformer converts the feed
hydrocarbons to aromatics, hydrogen and light gases are produced as byproducts. The liquid
product, known as reformate, is sent directly to gasoline blending, or to aromatics extraction.
Aromatics Extraction Unit
The purpose of aromatics extraction is to separate the aromatic compounds from the rest
of the hydrocarbons in reformate using chemical extraction with a solvent to concentrate the
individual aromatic compounds, (mainly xylene and benzene) for sale to the chemicals market.
Vacuum Tower
The purpose of the vacuum distillation tower unit is to enable a refinery to produce more
gasoline and diesel fuel out of a barrel of crude oil. It separates the vacuum gasoil (VGO), which
is fed to the FCC unit, from the vacuum tower bottoms (VTB) which is sent to the coker, or in
other refineries is made into asphalt. Because most sulfur contained in crude oil is contained in
the heaviest part of crude oil, the VGO and VTB are very high in sulfur.
Fluidized Catalytic Cracker
The purpose of the fluidized catalytic cracker is to convert heavy hydrocarbons, which
have very low value, to higher value lighter hydrocarbons. AGO and VGO are the usual feeds to
4-2
-------
a fluid catalytic cracker (FCC). The full boiling range cracked product leaves the reactor and is
sent to a fractionator. The overhead includes propane, propylene, butane, butylene, fuel gas and
FCC naphtha, which contains a significant amount of sulfur. There are two heavy streams; light
cycle oil (LCO), which can be hydrotreated and blended into diesel fuel or hydrocracked into
gasoline; and heavy cycle oil, sometimes called slurry oil, which can be used for refinery fuel.
Very simple refineries do not have FCC units, and therefore, produce gasoline with very low in
sulfur.
FCC Feed Hydrotreater or Mild Hydrocracker "A"
FCC feed hydrotreaters and mild hydrocrackers hydrotreat or mildly hydrocrack the feed
to the FCC unit which provides two distinct benefits. First, by increasing the amount of
hydrogen in the feed to the FCC unit, the FCC unit increases the conversion of the feed to high
value light products, particularly FCC naphtha which increases the gasoline yield. Second,
hydrotreating the feed removes some contaminants in the feed such as nitrogen and sulfur.
Nitrogen in the feed negatively affects the FCC catalyst. Removing the sulfur in the feed helps
in two ways. Some of the sulfur in the feed is released by the cracking process and results in
high SOx emissions that would otherwise have to be controlled by scrubbers - the FCC feed
hydrotreaters may prevent the need to add a scrubber. Also, FCC feed hydrotreaters remove
sulfur which can allow a refinery to comply with gasoline sulfur standards.
FCC Postreat Hydtrotreater "B"
Postreat hydrotreaters solely hydrotreat the naphtha that is produced by the FCC unit to
reduce its sulfur level which enables compliance with gasoline sulfur standards. The FCC
naphtha is high in olefms, which can be saturated by postreat hydrotreaters, resulting in lower
octane of the FCC naphtha. Vendor companies have developed postreat hydrotreating
technologies that minimize this octane loss.
Distillate Hydrotreater
The purpose of the distillate hydrotreater is to reduce the sulfur of distillate, which is also
called diesel fuel.
Gas Plant
The purpose of the gas plant is to use a series of distillation towers to separate various
light hydrocarbons for further processing in the alkylation or polymerization units or for sale.
Alkylation Unit
The purpose of the alkylation unit is to chemically react light hydrocarbons together to
produce a high quality, heavy gasoline product. Alkylation uses sulfuric or hydrofluoric acid as
catalysts to react butylene or propylene together with isobutane. Following the main reaction
and product separation, the finished alkylate is sent to gasoline blending. Alkylate is low in
RVP.
4-3
-------
Polymerization Unit
The purpose of the polymerization unit is to react light hydrocarbons together to form a
gasoline blendstock. A polymerization unit, often referred to as a "cat poly" is somewhat similar
to an alkylation unit, in that both use light olefins to produce gasoline blendstocks. The feed is
generally propylene and/or butylene from the gas plant. The product, called polygas is sent to
gasoline blending.
Coker Unit
The purpose of the coker unit is to process vacuum tower bottoms (VTB) to coke and to
crack a portion to various lighter hydrocarbons. The hydrocarbons produced by the coker
include cracked gases, coker naphtha, coker distillate and gas oil. The gas is fed to the gas plant,
the naphtha to the naphtha hydrotreater after which the heavy coker naphtha is typically fed to
the reformer, and the distillate either to distillate hydrotreating or to the hydrocracker.
Hydrocracker
The purpose of the hydrocracker is to crack and "upgrade" the feedstock into higher
value products. The feedstock to the hydrocracker is usually light cycle oil (LCO) and coker
distillate, poor quality distillate blendstocks, which are upgraded to diesel fuel, or cracked to
gasoline. Heavier hydrocarbons such as AGO and HVGO can be feedstocks as well.
A more complete description for naphtha hydrotreating is contained in Section 4.2.
4.2 Feasibility of Removing Sulfur from Gasoline
The case that it is feasible to comply with the 10 ppm gasoline sulfur standard can be
made in two ways. First, feasibility can be demonstrated because there are available
technologies that are currently available to achieve significant reductions in gasoline sulfur. A
discussion of these currently available technologies is contained below. Second, refiners in
California are already meeting a 10 ppm average, and certain countries or other regions are
currently complying with a 10 ppm gasoline sulfur cap standard. These two cases will be made
below, but first we will review the source of sulfur in gasoline to understand how sulfur levels
can be further reduced.
4.2.1 Source of Gasoline Sulfur
Sulfur is in gasoline because it naturally occurs in crude oil. Crude oil contains anywhere
from fractions of a percent of sulfur, such as less than 500 ppm (0.05 weight percent) to as much
as 30,000 ppm (3 percent). The average amount of sulfur in crude oil refined in the U.S. is about
14,000 ppm. Most of the sulfur in crude oil is in the heaviest part, or in the heaviest petroleum
compounds, of the crude oil (outside of the gasoline boiling range). In the process of refining
crude oil into finished products, such as gasoline, some of the heavy compounds are broken up,
or cracked, into smaller compounds and the embedded sulfur can end up in gasoline. Thus, the
4-4
-------
refinery units which convert the heavy parts of crude oil into gasoline are the units most
responsible for putting sulfur into gasoline.
The fluidized catalytic cracker (FCC) unit is a refinery processing unit that creates FCC
naptha, which is a high sulfur content gasoline blendstock. FCC naphtha contains from hundreds
to several thousand parts per million of sulfur. The FCC unit cracks large carbon molecules into
smaller ones and produces anywhere from 25 to 50 percent of the gasoline in those refineries
with FCC units. Because the FCC unit makes a gasoline blendstock out of the heavier, higher
sulfur-containing compounds, more than 95 percent of sulfur in gasoline blendstocks comes from
streams produced in that unit. For compliance with the 30 ppm Tier 2 gasoline sulfur standard
refiners reduced the sulfur content of the FCC naphtha. The impact of this action is described
below in subsection 4.2.2.
Straight run naphtha is a gasoline blendstock which contains a moderate amount of
sulfur. Straight run naphtha is the part of crude oil, which after distillation in the atmospheric
crude oil tower, falls in the gasoline boiling range. The heaviest portion of straight run, which
would have more sulfur, is normally desulfurized and reformed in the reformer (to improve its
octane), so its contribution to the gasoline pool is virtually nil.A The light straight run naphtha,
which contains the five-carbon hydrocarbons, contains on the order of 100 ppm sulfur and if this
material is not hydrotreated and processed in an isomerization unit, it is blended directly into
gasoline.
Another refinery unit which produces naphtha with a significant amount of sulfur is the
coker unit. These units produce coke from the heaviest part of the crude oil. In the process of
producing coke, a naphtha is produced that contains more than 3,000 ppm sulfur and many very
unstable olefms. Because this stream is highly olefmic and unstable, refiners tend to hydrotreat
coker naphtha. Coker naphtha is normally split into two different streams. The six- to nine-
carbon hydrocarbons are hydrotreated along with the rest of the heavy naphtha and fed to the
reformer. The five-carbon hydrocarbon part of coker naphtha is called light coker naphtha and
usually contains on the order of several hundred parts per million sulfur. Light coker naphtha is
usually hydrotreated along with the light straight run, and refined further in an isomerization unit
if the refinery has one.
Other gasoline blendstocks contain little or no sulfur. Alkylate, which is produced from
isobutene and butylenes that contain a small amount of sulfur, can end up with a small amount of
sulfur. Most refineries have less than 15 ppm sulfur in this pool, however, some refineries which
feed coker naphtha to the alkylate plant can have much more. On average, alkylate probably has
about 10 ppm sulfur. One more gasoline blendstock with either very low or no sulfur is
hydrocrackate, which is the naphtha produced by hydrocrackers. It is low in sulfur because the
hydrocracking process removes the sulfur. Ethanol, which is eventually blended into gasoline
usually has very little or no sulfur. However, the hydrocarbon used as a denaturant and blended
with ethanol at 2 percent is usually natural gasoline, a C5 to C7 naphtha from natural gas
A Sulfur interferes in the function of the precious metal catalyst used in the reforming process. As a result, refiners
historically have desulfurized the heavy straight run naphtha feed to the reformer from several hundred ppm sulfur
down to less than 1 ppm.
4-5
-------
processing, and it contains anywhere from a few parts per million to a couple hundred parts per
million sulfur. After the denaturant is blended in, the denatured ethanol contains somewhere
between 0 and 10 ppm sulfur. To meet current pipeline and California specifications, denatured
ethanol must contain less than 10 ppm sulfur.
4.2.2 Complying with the Current Tier 2 Gasoline Sulfur Standard
It is important to understand the steps that refiners took to comply with the 30 ppm Tier 2
gasoline sulfur standard because those capital investments and operational changes will play a
major role in determining the steps that refiners take to comply with a more stringent gasoline
sulfur standard.
r\
The Tier 2 sulfur standard was promulgated February 10, 2000. The sulfur standard
requires that refiners reduce their annual average gasoline sulfur levels down to 30 ppm and each
gallon of gasoline cannot exceed a per-gallon standard of 80 ppm. The sulfur standards were
phased in from 2004 to 2006. The compliance deadline for western refiners (GPA) and small
refiners were delayed until 2008. Some small refiners also had their gasoline sulfur deadlines
extended through 2010 if they met the compliance deadline for the highway diesel fuel sulfur
rule. As of January 1, 2011, all refineries are complying with the Tier 230 ppm sulfur standard.
A refinery's previous average gasoline sulfur level is an important factor which
determined whether a refiner would need to make a substantial capital investment to meet the
Tier 2 gasoline sulfur standards. We believe that refiners with low gasoline sulfur levels to begin
with (i.e., gasoline sulfur levels lower than about 50 ppm) probably did not invest in expensive
capital. These refineries have very low sulfur levels due to one or more of a number of possible
reasons. For example, some of these refiners may not have certain refining units, such as either a
FCC unit or a coker, which convert heavy boiling stocks to gasoline. As described above, these
units push more sulfur into gasoline and their absence means much less sulfur in gasoline.
Alternatively, these refiners may either use a very low sulfur (sweet) crude oil which can result
in a low sulfur gasoline, or have already installed an FCC feed hydrotreater, which uses a
heavier, higher sulfur (more sour) crude oil, to improve the operations of their refinery. As
described above, this unit removes much of the sulfur from the heaviest portion of the heavy gas
oil before it is converted into gasoline.
Of the refiners that already had low sulfur levels prior to Tier 2, the refineries with
average sulfur levels below 30 ppm may not have had to do anything to meet the Tier 2
standards. On the other hand, refineries with sulfur levels above 30 ppm but below about 50
ppm, probably are meeting the 30 ppm sulfur standard by employing operational changes only
and avoided making capital investments. Most refineries with gasoline sulfur levels below 50
ppm prior to the Tier 2 investments either do not have a FCC unit, or if they do, probably also
have an FCC feed hydrotreating unit.
The vast majority of gasoline that was being produced prior to the inception of the Tier 2
program was by refineries with higher sulfur levels. These refiners had to either adapt some
existing hydrotreating unit or install new capital equipment in these refineries to meet the Tier 2
gasoline sulfur standards. As stated above, the FCC unit is responsible for most of the sulfur in
gasoline. Thus, investments for desulfurizing gasoline involved the FCC unit to maximize the
sulfur reduction, and to minimize the cost of compliance with Tier 2. These desulfurization units
4-6
-------
were installed for treatment of either the gas oil feed to the FCC unit, or the gasoline blendstock
that is produced by the FCC unit. Each method has advantages and disadvantages.
4.2.2.1 Using FCC Feed Pretreat Hydrotreating to Comply with Tier 2
Some refiners installed FCC feed hydrotreaters (also known as pretreaters) at their
refineries to comply with the Tier 2 gasoline sulfur standards. FCC pretreaters treat the vacuum
gas oil, heavy coker gas oil and, in some cases, atmospheric residual feed to the FCC unit using a
hydrotreater or a mild hydrocracker. These units are designed to operate at high pressures and
temperatures to treat a number of contaminants in the feed. Besides sulfur, FCC pretreaters also
reduce nitrogen and certain metals such as vanadium and nickel. These nonsulfur contaminants
adversely affect the FCC catalyst, so the addition of this unit would improve the functioning of
the unit. Also, because hydrotreating which occurs in the FCC pretreater reacts hydrogen in the
feedstock, it increases the yield of the FCC unit, increasing the production of high profit-making
products, such as gasoline and light olefins.3 While FCC pretreaters provide yield benefits that
offset the capital costs of adding this type of desulfurization, the costs are still high enough that
many refiners would have a hard time justifying the installation of this sort of unit. For a
medium to large refinery (i.e., 150,000-200,000 BPCD), the capital costs may exceed $250
million. Because of the higher temperatures and pressures involved, utility costs are expensive
relative to postreat hydrotreating as explained below. Using FCC feed pretreating also allows
refiners to switch to a heavier, more sour crude oil. These crude oils are less expensive per-
barrel and can offset the increased utility cost of the FCC pretreater, providing that the
combination of reduced crude oil costs and higher product revenues justify the switch. Another
benefit for using FCC pretreaters is that the portion of the distillate pool that comes from the
FCC unit would be partially hydrotreated as well. This distillate blendstock, termed light cycle
oil, comprises a relatively small portion of the total distillate produced in the refinery (about 20
percent of on-road diesel comes from light cycle oil), and like FCC naphtha, light cycle oil
contributes a larger portion of the total sulfur which ends up in distillate. Thus, FCC pretreaters
would also help a refiner meet the 15 ppm highway and nonroad diesel fuel standards.
In terms of desulfurization capability, FCC pretreaters have different abilities to remove
sulfur from the gas oil feed depending on the unit pressure. FCC pretreaters can be subdivided
into high pressure units (1400 psi and above), medium pressure units (900 to 1400 psi), and low
pressure units (under 900 psi). High pressure FCC pretreaters typically remove about 90 percent
of the sulfur contained in the gas oil feedstock to the FCC unit, while low and medium pressure
units typically remove 65 to 80 percent of the feed sulfur.4 We are aware of at least 5 refineries
in the U.S. that use high pressure FCC pretreaters. Because there is no postreating at many of
the refineries with FCC pretreaters, control of the feed to these units is a critical determining
factor for how well the FCC pretreater will function as desulfurizers. If the feed becomes too
heavy (due to a higher temperature endpoint), there would be a higher concentration of sulfur
and other contaminants in the feed. To maintain the same sulfur level in the FCC naphtha, the
FCC pretreater unit would have to be operated at a higher temperature which causes the catalyst
to lose its effectiveness more quickly.5
FCC pretreaters improve desulfurization indirectly by improving the desulfurization
performance of the FCC unit itself. When FCC units crack the vacuum gas oil into naphtha,
about 90 percent of the sulfur is typically cracked out of the hydrocarbons converted to FCC
4-7
-------
naphtha (or the FCC naphtha contains only about 10 percent of the sulfur present in the feed) and
is removed as hydrogen sulfide. When FCC pretreaters are used, the amount of sulfur in the
feed, which ends up in the FCC naphtha, is only about 5 percent. This means that about 95
percent of the sulfur in the feed is removed from the FCC feed when it is cracked into FCC
naphtha. This is due to the additional hydrogen in the pretreater, which reacts with the feed
hydrocarbons. With more hydrogen molecules available in the feedstock after hydrotreatment,
the FCC cracking reactions can react more hydrogen with the sulfur contained in the feed to
produce more hydrogen sulfide.
For complying with Tier 2, refiners with existing pretreaters or those that installed high
pressure FCC pretreaters were able to comply with the 30 ppm sulfur standard without the need
to install a FCC naphtha hydrotreater. Refineries that had either a low pressure, or medium
pressure FCC feed hydrotreater were generally less able to comply with the 30 ppm gasoline
sulfur standard with the FCC hydrotreater by itself, and were more likely to also install an FCC
postreater.
4.2.2.2 Using FCC Naphtha Postreat Hydrotreating to Comply with Tier 2
A less capital intensive alternative for reducing FCC naphtha sulfur levels to comply with
Tier 2 is FCC naphtha hydrotreating (also known as postreaters). FCC postreaters only treat the
gasoline blendstock produced by the FCC unit. This unit is much smaller than an FCC pretreater
because only about 50 to 60 percent of the feed to the FCC unit ends up as FCC naphtha, a
gasoline blendstock. The unit is sometimes smaller still because some refiners which choose to
use a fixed bed hydrotreater may only treat the heavier, higher sulfur portion of that stream with
hydrotreating, and then treat the lighter fraction with another lower desulfurization cost
technology. FCC postreaters operate at lower temperatures and pressures than FCC pretreaters,
which further reduces the capital and operating costs associated with this type of desulfurization
equipment. Furthermore, because feed to the FCC unit has corrosive properties, FCC pretreaters
use more corrosion-expensive metallurgy, which is not needed for postreaters.6 For a medium to
large-sized refinery, the capital costs are on the order of $70 million for a conventional FCC
postreater - about a third the cost of an FCC pretreater.
One disadvantage of this desulfurization method is that the octane value and/or some of
the gasoline yield may be lost depending on the process used for desulfurization. Octane loss
occurs by the saturation of high octane olefins which are produced by the FCC unit. Most of the
olefins are contained in the lighter fraction of FCC naphtha.7 Increased olefin saturation usually
means higher hydrogen consumption. There can also be a loss in the gasoline yield caused by
mild cracking that breaks some of the gasoline components into smaller fractions which are too
light for blending into gasoline. If there is octane loss, the octane loss can be made up by
increasing the feed to or the severity of the reformer, the aromatics production unit of the
refinery, producing more alkylate, or purchasing high octane gasoline blendstocks (such as
reformate) which is routinely trading between refineries. Sometimes vendors of FCC pretreater
technologies design octane increasing capability into their designs, which is discussed below in
the section about the individual postreater technologies.
The loss of octane and gasoline yield caused by FCC postreating is lower with
o
technologies that were developed prior to the implementation of the Tier 2 program. These
4-8
-------
processes are termed selective because they achieve the lower sulfur while preserving much of
the octane and gasoline yield (they were designed specifically for treating FCC naphtha). Octane
is preserved because the hydrotreating units and their catalysts are specially designed to avoid
saturating olefms. These selective processes, or parts of these processes, usually operate at less
severe conditions and result in less cracking and thus, preserve yield compared to conventional
hydrotreating processes. The less severe conditions also lower the capital and operating costs for
this process. The lower operating costs arise out of the reduced utility requirements (e.g., lower
pressure). For example, because these processes are less severe, there is less saturation of
olefms, which means that there is less hydrogen used. Less olefin saturation also translates into
less octane loss, which would otherwise have to be made up by octane boosting processing units
in the refinery. The lower capital and operating costs of these newer FCC postreaters are
important incentives for refiners to choose this desulfurization methodology over FCC
pretreaters. For this reason, refiners chose to use the more recently developed FCC postreaters
technologies for meeting the 30 ppm Tier 2 gasoline sulfur standard.
Not saturating the olefms to preserve octane and limit hydrogen consumption provides a
different challenge. During desulfurization, when the hydrogen sulfide is formed and there is a
significant concentration of olefms present, the hydrogen sulfide compounds tend to react with
the olefmic hydrocarbon compounds forming mercaptan sulfur compounds. This reaction is
called "recombination" because the removed sulfur recombines with the olefmic hydrocarbons
contained in the naphtha.9 This is particularly a problem if the light cat naphtha is present in the
hydrotreater because the highest concentration of olefms is in the light cat naphtha. The
recombination reactions occur more readily if the hydrotreater is operated more severely (at a
higher temperature) to increase the sulfur removal, and the feed to the hydrotreater is high in
sulfur. However, while operating this type of hydrotreater more severely can result in the further
removal of the original sulfur present in the hydrocarbons, it also can result in the formation of
more recombination mercaptans that results in a "floor" reached for the amount of sulfur that can
be removed from the hydrocarbons. This cycle of increased sulfur removal and simultaneous
increase in recombination results in the saturation of more olefms and increases the consumption
of hydrogen. There are a number of different vendor-specific technologies that each vendor may
use to avoid or address recombination reactions as discussed below. It is important to note that
the technologies employed to reduce recombination may require the addition of some capital
costs which offsets some or perhaps all the capital cost savings due to the milder operating
conditions of these selective hydrotreater technologies compared to nonselective hydrotreating.
One means to achieve high levels of desulfurization while avoiding much of the problem
with recombination reactions is by using a two-stage hydrodesulfurization methodology. A two-
stage unit has two desulfurization reactors, but instead of just adding additional reactor volume,
the hydrocarbons exiting the first reactor are stripped of gaseous compounds (most importantly,
the hydrogen sulfide is removed), injected with fresh hydrogen, and then hydrodesulfurized
again in the second stage. Both reactors undergo modest desulfurization and hydrogen sulfide
concentrations remain sufficiently low to avoid recombination reactions. The disadvantage of
this approach is that the second stage incurs greater capital costs compared to single-stage
configurations. Because Tier 2 was not too constraining, we believe that refiners installed few, if
any, two-stage desulfurization units to comply with those gasoline sulfur standards.
4-9
-------
Whatever strategy chosen by the refiner to comply with Tier 2, a critical criterion was
that the postreater be capable of cycle lengths that match that of the FCC unit, which typically is
4 years. If the postreater were to require a catalyst changeout before the FCC unit requires a
shutdown, the refiner would either have to shutdown the FCC unit early to mirror that of the
postreater, or store the high sulfur FCC naphtha (this stream would be too high in sulfur to blend
directly to gasoline under the Tier 2 80 ppm cap standard) until the postreater was started up
again and is able to hydrotreat the stored up high sulfur FCC naphtha.
We know of six FCC postreater technologies that refiners used to comply with the Tier 2
gasoline sulfur standards. These are Axens (was IFF) Prime G and Prime G+, Exxon Scanfming,
CDTech's CDHydro and HDS, Sinopec's (was Phillips)S-Zorb and UOP's ISAL and
Selectfming.
Axens Prime G+, Exxon Scanfming and UOP's ISAL and Selectfming are all fixed bed
desulfurization technologies. These processes are called fixed bed because the catalyst resides in
a fixed bed reactor.10 The high sulfur gasoline blendstock is heated to a high temperature (on the
order of 600 degrees Fahrenheit) and pumped to a high pressure to maintain the stream as a
liquid. It is then combined with hydrogen before it enters the reactor. The reactions occur
within the bed of the catalyst. While the petroleum is in contact with the catalyst in the reaction
vessel, the sulfur reacts with hydrogen and is converted to hydrogen sulfide. Also, depending on
the process, some of the olefin compounds that are present in the cracked stream are saturated
which increases the amount of octane lost and hydrogen consumed. After the reactor, the
gaseous compounds, which include unreacted hydrogen, hydrogen sulfide, and any light end
petroleum compounds which may have been produced in the reactor by cracking reactions, are
separated from the liquid compounds by a gas/liquid separator. The hydrogen sulfide must be
stripped out from the other compounds and then converted to elemental sulfur in a separate
sulfur recovery unit. The recovered sulfur is then sold. If enough hydrogen is present, and it is
economical to recover, it is separated from the remaining hydrocarbon stream and recycled.
Otherwise, it is burned with light hydrocarbons as fuel gas.
Each of these fixed bed desulfurization technologies is somewhat different. Axens Prime
G+ desulfurization process largely preserves olefins as its strategy for diminishing octane
loss.11'1213 The Axens process employs a selective hydrogenation unit (SHU) as a first step. The
role of this unit is to saturate the unstable diolefm hydrocarbons in a hydrogen rich environment,
and react the light mercaptan and sulfide hydrocarbons together. The SRU also converts exterior
olefins to interior olefins, which results in a small increase in octane. The mild operating
conditions of the SHU tend to avoid the saturation of monoolefms. After exiting the SRU, the
FCC naphtha is sent to a distillation column which separates the light FCC naphtha (typically
comprising about one fourth of the total cat naphtha) from the heavy naphtha. Because the light
sulfur compounds were reacted together and those compounds no longer fall within the light cat
naphtha boiling range, the light cat naphtha is low in sulfur and can be blended directly into
gasoline. The heavy cat naphtha which is naturally high in sulfur and which also contains the
self-reacted light mercaptans and sulfides from the SHU, is sent to a fixed bed hydrotreater. The
fixed bed hydrotreater contains both cobalt-molybdenum and nickel-molybdenum catalyst. An
important way that Axens avoids recombination reactions is by separating the light sulfur
compounds from the light naphtha and keeping the light naphtha out of the fixed bed
hydrotreater. The desulfurized heavy cat naphtha is blended into the gasoline pool.
4-10
-------
If the feed to the Axens Prime G unit is very low in sulfur, a low capital investment
option was available to the refiner by feeding the entire FCC naphtha stream to the hydrotreating
reactor avoiding the SHU and splitter. This option trades lower capital cost with somewhat
higher octane loss and hydrogen consumption. Because of the low severity of the hydrotreating
reactor (low severity is possible because the lower amount of desulfurization that is occurring),
the amount of octane loss and hydrogen consumption is modest. There are more than 180 Prime
G+ units operating worldwide, and approximately 40 in the U.S.
The first step in Exxon's fixed bed Scanfining process is to mildly heat the full FCC
naphtha and pass it through a small reaction vessel which reacts the diolefins to monoolefins.14 15
161? is 19 rpjie ^j pŁŁ nap^a js ^jien heated further, injected with hydrogen gas and sent to the
fixed bed hydrotreating reactor, which is packed with a catalyst developed jointly between
Exxon and Akzo Nobel (now Albermele). If the degree of desulfurization is relatively modest,
the amount of recombination is low and the FCC naphtha is sent to gasoline blending. If,
however, the degree of desulfurization is higher (due to FCC naphtha with a higher sulfur
content), then there likely would be an excessive number of recombination reactions. In this
case, Exxon recommends either one of two different technologies to address the recombination
reactions. One technology is Zeromer. Zeromer is a fixed bed reactor vessel installed after the
main fixed bed hydrotreater reactor that specifically designed to hydrodesulfurize the mercaptan
sulfur from the FCC naphtha without saturating olefins.2 Another technology Exxon developed,
in conjunction with Merichem, is an extractive mercaptan removal technology named Exomer.
The Exomer technology differs from other sulfur extraction technologies in that it is capable of
91
extracting mercaptans from the entire FCC naphtha pool. Like Zeromer, the Exomer
technology would be an add-on technology installed after the Scanfining fixed bed reactor.
There are 16 Scanfining units operating in the U.S.
UOP has licensed two FCC naphtha hydrotreating technologies. When Tier 2 was being
phased-in, UOP was licensing a technology named ISAL developed by INTEVEP S.A.22 23 The
ISAL process is different from the other FCC naphtha hydrotreaters because instead of avoiding
the saturation of olefins as sulfur is being hydrotreated out of FCC naphtha, the ISAL process
completely saturates the olefins. To avoid a large octane loss, the ISAL process separates the
olefin-rich, light cat naphtha from the heavy cat naphtha. The light cat naphtha is treated by an
extractive desulfurization technology such as Merox which does not saturate olefins. Only the
heavy cat naphtha is sent to the ISAL reactor. To offset the octane loss caused by the saturation
of the olefins in the heavy cat naphtha as it is being desulfurized, the ISAL catalyst isomerizes
and conducts some mild cracking and reforming of the heavy cat naphtha. One downside of the
ISAL process is that, due to the complete saturation of olefins, the hydrogen consumption is
higher relative to the selective hydrodesulfurization technologies that avoid saturating olefins.
UOP has since developed and licensed a FCC naphtha desulfurization technology called
SelectFining.24 SelectFining is a selective hydrodesulfurization technology that seeks to
minimize olefin saturation to minimize both octane loss and hydrogen consumption.
SelectFining treats the full FCC naphtha. The full range FCC naphtha is first sent to a diolefin
saturating reactor before being sent to the SelectFining reactor. SelectFining relies on its catalyst
design to selectively remove sulfur and prevent recombination reactions. UOP recommends a
two-stage reactor setup for high levels of desulfurization.
4-11
-------
The next two FCC naphtha desulfurization technologies, CDTech and S-Zorb do not use
fixed bed reactors, but very different technologies which are also very different from each other.
Each will be discussed separately.
The CDTech process still uses the same type of catalyst used in fixed bed reactors.
7S 9ft 77
However, it also utilizes catalytic distillation. ' ' Catalytic distillation is a technology which
has been applied for a number of different purposes. CDTech is currently licensing the
technology to produce MTBE and selective hydrogenation processes, including FCC naphtha
desulfurization and benzene saturation. As the name implies, distillation and desulfurization, via
catalyst, take place in the same vessel. This design feature saves the need to add a separate
distillation column sometimes used with fixed bed hydrotreating. All refineries have a
distillation column after the FCC unit (called the main fractionation column) that separates the
FCC naphtha from the most volatile components (such as liquid petroleum gases), the distillate
or diesel (light cycle oil), and the heavy ends or residual oil. However, if a refiner only wishes to
treat a portion of the FCC naphtha, then a second distillation column would need to be added
after the main FCC fractionation column to separate out the portion of the FCC naphtha that he
wishes not to treat. With the CDTech process, the refiner can choose to treat either the entire
pool or a portion of the pool, but choosing to treat a part of the pool, thus negating any need for
an additional distillation column.
The most important portion of the CDTech desulfurization process is a set of two
distillation columns loaded with desulfurization catalyst in a packed structure. The first vessel,
called CDHydro, treats the lighter compounds of FCC gasoline and separates the heavier portion
of the FCC naphtha for treatment in the second column. The second column, called CDHDS,
removes the sulfur from the heavier compounds of FCC naphtha. All of the FCC naphtha is fed
to the CDHydro column. The five- and six-carbon petroleum compounds boil off and head up
through the catalyst mounted in the column, along with hydrogen which is also injected in the
bottom of the column. The reactions in this column are unique in that the sulfur in the column is
not hydrotreated to hydrogen sulfide, but they instead are reacted with dienes in the feed to form
thioethers. Their higher boiling temperature causes the thioethers to fall to the bottom of the
column. They join the heavier petroleum compounds at the bottom of the column and are sent to
the CDHDS column. Because the pressure and temperature of the first column is much lower
than conventional hydrotreating, saturation of olefins is reduced to very low levels. The olefin
saturation which does occur is necessary to eliminate diolefins. Thus, little excess hydrogen is
consumed. CDTech offers an option to refiners to put in an additional catalyst section in the
CDHydro column to increase octane. This octane enhancing catalyst isomerizes some of the
olefins, which increases the octane of this stream by about three octane numbers, and few of the
olefins are saturated to degrade this octane gain. The seven-carbon and heavier petroleum
compounds leave the bottom of the CDHydro unit and are fed into the CDHDS column. There,
the heavier compounds head down the column and the lighter compounds head up. Both
sections of the CDHDS column have catalyst loaded into them, which serve as hydrotreating
reaction zones. Similar to how hydrogen is fed to the CDHydro column, hydrogen is fed to the
bottom of the CDHDS column.
The temperature and pressure of the CDTech process columns are lower than fixed bed
hydrotreating processes, particularly in the upper section of the distillation column, which is
where most of the olefins end up. These operating conditions minimize yield and octane loss.
4-12
-------
While the CDTech process is very different from conventional hydrotreating, the catalyst used
for removing the sulfur compounds is the same. One important difference between the CDTech
process and conventional hydrotreating is that CDTech mounts its catalyst in a unique support
system, while conventional catalyst is usually dumped into the fixed bed reactor. CDTech has
13 CDHydro/CDHDS desulfurization units in operation in the U.S.
Phillips Petroleum Co. commercialized and licensed an adsorption desulfurization
technology called S-Zorb, which it sold to SINOPEC in 2007.28 29 S-Zorb uses a chemical
adsorption process, instead of hydrotreating, as the principal methodology for the removal of
sulfur from FCC naphtha. Adsorption has the benefit of operating at much lower pressure and
temperatures, which lowers operating costs. S-Zorb, uses two separate columns and is constantly
moving an adsorption catalyst from the reactor vessel to the regeneration column, and back
again.30 The untreated FCC naphtha and hydrogen are fed to the reaction vessel where the sulfur
is catalytically removed the sulfur from the petroleum compound and facilitated by the hydrogen
present in the reactor. The catalyst, which begins to accumulate the removed sulfur, is
transferred over to the regeneration column on a continual basis where the sulfur is removed
from the catalyst using hydrogen as the scavenging compound. Then the hydrogen disulfude is
converted to sulfur dioxide and sent to the sulfur recovery unit. Because the process still relies
upon catalytic processing in the presence of hydrogen, there is some saturation of olefms, with a
commensurate reduction in octane. Through a literature search, we believe that 7 S-Zorb
desulfurization units were originally licensed for Tier 2. Other sources indicated that only 4
units are actually operating today.
We also conducted a literature search and asked both refiners and vendors to identify the
FCC naphtha desulfurization technology that was installed at each refinery to enable compliance
with Tier 2. A summary of the total number of units by vendor and technology type is
summarized in Table 4-1.
Table 4-1 Estimated Number of FCC Desulfurization Technologies Installed to comply
with Tier 2 by Vendor Company or Technology
Axens
Prime G
40
Exxon
Scanfming
16
CDTech
15
Sinopec S-
Zorb
4
UOP ISAL
UOP Selectfming
2
FCC Feed
HT
17
No FCC
Unit
14
4.2.3 Meeting a 10 ppm Gasoline Sulfur Standard
To meet a 10 ppm average gasoline sulfur standard, we believe that the primary strategy
that refiners would adopt would be to further reduce the sulfur level of FCC naphtha. There are
three primary reasons why we believe this will be the primary strategy and therefore, used it for
analyzing the compliance costs for Tier 3. The first reason is that FCC naphtha is by far the
largest contributor of sulfur to the gasoline pool, by virtue of both its volume and sulfur content,
even after refiner's use of hydrotreating to reduce the sulfur in the FCC naphtha to comply with
Tier 2. Table 4-2 below summarizes the estimated average volumes and average sulfur levels for
the primary blendstocks typically blended into gasoline for the current Tier 2 standards. By
using the refinery-by-refinery model to model today's situation for the typical refinery, we
estimate that the FCC naphtha contains about 75 ppm for the typical refinery complying with the
4-13
-------
30 ppm Tier 2 sulfur standard and that gasoline blendstock typically contributes to about 34
percent of a refiner's gasoline pool. Table 4-2 also summarizes the changes in gasoline
blendstock sulfur levels we believe would occur when complying with the 10 ppm gasoline
sulfur standard. Using the refinery-by-refinery model, we project that a 10 ppm gasoline sulfur
standard can be met by a typical refinery by reducing the sulfur level of FCC naphtha from about
75 ppm to 25 ppm. We believe that virtually all refineries that have an FCC unit would not be
able to comply with the proposed 10 ppm gasoline sulfur standard without further desulfurizing
the FCC naphtha. The second reason is that both vendors and refiners have told us that this is the
gasoline blendstock stream that they intend to address. Both vendors and refiners have explained
to us that, for most refineries, existing FCC naphtha hydrotreaters can be retrofitted with only a
modest capital cost to realize the sulfur reduction needed. Third, further reducing the sulfur of
the FCC naphtha as the means to comply with Tier 3 is supported by other cost studies. When
these studies assessed the costs for further reducing the sulfur levels of gasoline, they also
focused further reducing the sulfur levels of the FCC naphtha. See the subsection at the end of
Chapter 5 discussing these other cost studies.
Table 4-2 Estimated Typical Gasoline Blendstock Volumes and Sulfur Levels after Tier 2
and Complying with a 10 ppm Sulfur Standard
Gasoline
Blendstock
FCC Naphtha
Reformate
Alkylate
Isomerate
Butane
Light Straight Run
Naphtha and
Natural Gas Liquids
Hydrocrackate
Ethanol
Coker Naphtha
Other Gasoline
Blendstocks
Total/Sulfur
Average
30 ppm Tier 2 Gasoline Sulfur
Standard
Volume
(Percent)
37
23
13
O
4
5
O
10
2
1
100
Sulfur (ppm)
75
0.5
10
0.5
5
15
8
5
1
10
30
10 ppm Gasoline Sulfur Standard
Volume
(Percent)
36
22
13
3
4
5
3
12.5
2
1
100
Sulfur (ppm)
25
0.5
10
0.5
5
5
8
5
1
1
10
Reducing FCC naphtha from 75 ppm to 25 ppm would likely be accomplished in
different ways depending on the desulfurizing technology and configuration used for Tier 2, and
whether the current capital employed for lowering gasoline sulfur is severely taxed or not severely
taxed. For purposes of this discussion, we will discuss the likely steps taken to comply with Tier
3 based on whether a refiner solely used an FCC pretreater or FCC postreater to comply with Tier
2. While we provided an example for a typical refinery needing to reduce its FCC naphtha from
75 ppm to 25 ppm to enable compliance with Tier 2, refineries that are not typical would have
starting and ending sulfur levels that are different from this example. Despite these differences,
4-14
-------
we believe that every refinery can physically comply with a 10 ppm gasoline sulfur standard.
This is because there are no technical difficulties removing sulfur from gasoline - the challenge is
to comply while minimizing the cost of doing so, such as by minimizing the associated octane
loss and by taking advantage of the flexibilities provided to comply. The cost analysis, reflecting
the ability for refineries to comply though the use of the averaging banking and trading provisions
is discussed in Chapter 5 of this RIA. Section 4.2.3.5 below also discusses the importance of the
averaging, banking and trading program.
The one exception is the case where a refinery does not have an FCC unit. Refineries in
this situation would likely already be producing gasoline which is 10 ppm or below. If such a
refinery's gasoline is above 10 ppm, then the refiner would need to address one or more of
several different gasoline blendstocks, including light straight run and natural gas liquids. Our
discussion on treatment of other gasoline streams can be found in Section 4.2.3.3.
4.2.3.1 Meeting 10 ppm if Refiners Used an FCC Feed Pretreater to Comply with
Tier 2
Refiners that relied on an FCC pretreater to comply with Tier 2 at a refinery would likely
only be able to achieve 10 ppm sulfur gasoline if its FCC pretreater is a high pressure unit.3132
This is because most refineries that have FCC pretreaters process sour crude oils and if the unit is
a mid or low-pressure unit, the unit pressure would likely be too low to sufficiently desulfurize
the FCC feed. This may be true even if the refiner added reactor volume to its existing low or
medium pressure FCC pretreater, which does cause additional desulfurization. Mid and low
pressure FCC pretreaters just cannot remove enough of the sulfur in the gas oil feed to the FCC
unit to achieve adequately low sulfur levels in the FCC naphtha. If a refinery processes
moderate to low sulfur crude oil and has a low to mid-pressure FCC pretreater, however, it may
be able to achieve an adequate degree of desulfurization in the FCC naphtha to enable the refiner
to reduce its gasoline sulfur down to 10 ppm. If a refinery cannot achieve a sufficient level of
desulfurization with its current or revamped FCC pretreater to comply with a 10 ppm gasoline
sulfur standard, then the refiner will have to install a grassroots FCC postreater. Alternatively,
refiners in this situation would be in the best situation to take advantage of the averaging,
banking and trading program (ABT). Using the ABT provisions to its advantage, the refiner
would achieve the most desulfurization that it can with its existing FCC pretreater (perhaps 20
ppm sulfur gasoline), and then would purchase credits to demonstrate the remainder of its
compliance with the 10 ppm gasoline sulfur standard. Such a refiner would then avoid the need
to install an expensive grassroots FCC postreater.
While they are expensive to install, FCC pretreaters provide important operating cost
advantages over postreaters. An important advantage of FCC pretreating is that it occurs
upstream of the FCC unit and therefore, does not jeopardize the octane value of the olefms
produced in the FCC unit. Another advantage of the FCC pretreater is that it tends to increase
the yield of naphtha from the FCC unit, which improves operating margins for the refinery with
such a unit. Thus, refiners that are able to use FCC pretreaters to comply with the Tier 3 sulfur
standard would likely yield a further return on any investment made, and offset some if not all of
the increased operating costs incurred.
4-15
-------
A downside to solely relying on FCC pretreating to comply with Tier 3 is that a refinery
has less operational flexibility. An FCC feed hydrotreater must be shutdown every one to three
years to change out the catalyst, which is usually more often than when the FCC unit is
shutdown. During the shutdown of the FCC pretreater, if the FCC unit remains operational a
refiner has to figure out what to do with its high sulfur FCC naphtha (the refiner could blend up
gasoline without FCC naphtha). For this reason, refiners that are complying with Tier 2 by
solely relying on an FCC pretreater may choose to either install a grassroots postreater instead to
comply with Tier 3, or purchase credits.
The refiners most likely to rely on FCC pretreater to comply with Tier 3 are those with
high pressure FCC pretreaters. As noted earlier, there are likely only 5 refineries have high
pressure FCC pretreaters in the U.S. More refiners with FCC pretreaters may be able to comply
with the Tier 3 standards using just their FCC pretreater, however, if they undercut the FCC
naphtha into the diesel pool. The sulfur reduction in the FCC naphtha caused by undercutting
would enable refiners to rely on lower pressure FCC pretreaters to comply with Tier 3, while
also increasing diesel supply.
4.2.3.2 Meeting 10 ppm if Refiners Used an FCC Postreater to Comply with Tier 2
If a refiner installed an FCC postreater to comply with the Tier 2 gasoline sulfur standard,
there are several considerations about the current configuration of the postreater which would
affect how a refiner would use this unit to comply with a 10 ppm gasoline sulfur standard. EPA
considered the issue of the degree of desulfurization the postreater is currently facing. In doing
so, EPA analyzed several examples to understand the types of revamps and associated
investments that might occur for such refiners.
For the first example, if the refinery is refining a very sour (high sulfur) crude oil and the
sulfur of the FCC naphtha exiting the FCC unit is 2,400 ppm, the postreater is currently
removing almost 97 percent of the feed sulfur. This assumes that the sulfur level of the FCC
naphtha exiting the postreater is 75 ppm, which is a very high level of desulfurization. When
attempting to achieve further sulfur reduction in the FCC naphtha, the refiner must be concerned
about the increased occurrence of recombination reactions and the potential for much more
octane loss and hydrogen consumption. This refiner would strongly consider adding a second
stage, which may actually reduce the level of recombination reactions and the octane loss
currently experienced by the postreater. Most of the vendors offer a second stage option. In the
case of CDTech, they call the second reactor, added as part of its second stage, a polishing
reactor. We contacted the desulfurization engineer at Sinopec who explained that these units
could be turned up and that no additional capital investments would be needed (though there are
additional operating costs). A Conoco-Phillips hydrotreating specialist we spoke to confirmed
that this would be the strategy for their S-Zorb units. We also considered an additional option of
the refiner is interested in improving its operating margins such as increased gasoline production,
and has ample capital dollars to spend. Such a refiner could add an FCC feed hydrotreater to
increase its yield of FCC naphtha, or a mild hydrocracker to increase its production of low sulfur
distillate.
In contrast, if a refiner is processing a very sweet (low sulfur) crude oil, the sulfur level
exiting the FCC unit may be as low as 300 ppm, and under Tier 2, the level of desulfurization
4-16
-------
necessary to bring that stream down to 70 ppm is about 81 percent which is a very modest level
of desulfurization. Similarly, a refinery processing a moderately sour crude oil with a medium
pressure FCC feed hydrotreater could be in a similar situation. The refineries in this situation
could have a lot more capacity in their existing postreaters to achieve lower sulfur without
additional capital cost investments. However, many refiners in this situation which invested in
an FCC postreater for Tier 2 may have minimized their capital investments. For example, a
refiner may have avoided the capital and operating cost of a splitter with its postreater by
hydrotreating the full range FCC naphtha. Therefore, the increased severity of the postreater
needed to achieve 20 ppm in the FCC naphtha to meet a 10 ppm gasoline sulfur standard might
create a larger octane loss and higher hydrogen consumption than what the refinery could easily
provide without a significant additional capital investment. In this case, the refiner can invest
some capital in the postreater to minimize the increase in octane loss and hydrogen consumption.
For example a refiner with an Axens unit in this situation could add the SHU and a splitter. A
refiner with a Scanfming unit in this situation wishing to minimize the octane loss and hydrogen
consumption could add a Zeromer or an Exomer unit. Alternatively, if the refiner is processing a
moderately sour crude oil and has a moderate pressure FCC feed hydrotreater, the refinery may
choose instead to revamp the FCC feed hydrotreater for its operational benefits rather than
revamp the postreater.
We also considered a third example where a refiner with a postreater has FCC gasoline
exiting the FCC unit at 800 ppm. This is probably most typical of a refinery processing either
crude oil containing an average amount of sulfur, or, perhaps a refinery refining a very sour
crude oil but treating the vacuum gas oil with a low pressure FCC feed hydrotreater. The current
FCC naphtha hydrotreater would be achieving about 90 percent desulfurization when producing
FCC naphtha with 80 ppm sulfur. In looking to reduce the FCC naphtha down to 20 ppm to
comply with a 10 ppm sulfur standard, such a refiner would not likely consider adding a second
stage reactor. This is because avoiding both increased octane loss and hydrogen consumption for
the additional increment of sulfur reduction would probably not justify the capital costs
associated with a second stage reactor. Instead of a second stage reactor, a refiner could revamp
the existing FCC postreater with additional reactor volume, or add capital for addressing
recombination reactions, both likely to be a lot less capital intensive than a second stage. A no
investment option is possible for refiners in this situation, although the increase in octane loss
and hydrogen consumption is likely to be significant.
The most important part of an FCC hydrotreater is likely the catalyst used in the unit.
Due to continuing research, catalysts are constantly being developed which are more active, thus
achieving greater desulfurization at a lower temperature, and minimizing octane loss and
hydrogen consumption due to lower olefm saturation. When the Tier 2 naphtha desulfurizers
were being put into service the most recent catalysts were likely used in those units. These
catalysts can be changed when a postreater is undergoing regular maintenance, and new and
improved catalysts can be used to improve the desulfurization capacity of the unit. If refiners
need to reduce their gasoline sulfur levels to 10 ppm gasoline, they would be expected to
upgrade to the most recent catalyst to minimize their costs. Using the most active catalyst
available would reduce the capital cost that would need to be incurred and reduce the hydrogen
consumption and octane loss that would otherwise occur. We are aware of newer lines of more
active catalysts being marketed by Axens and UOP. It is likely that since the time catalysts were
4-17
-------
loaded into FCC postreaters to comply with Tier 2 all vendors are now offering improved
hydrotreating catalysts.
4.2.3.3 Desulfurizing Other Blendstocks
A more stringent gasoline sulfur standard could require refiners to address other gasoline
streams with high enough in sulfur content to be a concern to the refiners when complying with
Tier 3. This is because without addressing such gasoline streams, the refiner would have to
reduce their FCC naphtha even lower in sulfur resulting in high per gallon costs at the lower
sulfur levels. The gasoline streams that we have identified that could require additional
desulfurization include light straight run naphtha and natural gas liquids.
Light straight run naphtha (LSR) is naturally occurring in the crude oil and is
desulfurized at many refineries before it is sent to an isomerization unit. However, a number of
refineries do not have isomerization units and therefore, some or perhaps many of these
refineries may not be treating this stream today. Natural gas liquids (also termed pentanes plus)
are naphtha streams sourced from natural gas wells, which are purchased by refiners, and
blended into the gasoline pool. Depending on the source of the specific naphtha stream being
purchased, these streams could vary widely in gasoline sulfur, ranging from a few ppm sulfur up
to several hundred ppm sulfur.
Refiners have multiple options for addressing the sulfur levels of these various streams.
The LSR and natural gas liquids can be hydrotreated in either the FCC postreaters or the naphtha
hydrotreaters. Because these naphtha streams do not have any olefms, there is essentially no
octane loss and, therefore, hydrogen consumption is lower compared to hydrotreating FCC
naphtha. Another way of treating these streams would be to use caustic extraction to extract the
mercaptan sulfur from these streams. Since only the mercaptans are removed with the extraction
technology, the final sulfur level would not be as low compared to desulfurization using
hydrotreating. If the crude oil that is being refined by a particular refinery is low in sulfur, the
refiner would likely only need to use extractive desulfurization to ensure that the sulfur in the
LSR is adequately low under Tier 3. Finally, the refiner could choose to simply not purchase the
natural gas liquids and sell the LSR on the open market as opposed to treating these streams. If a
refiner decides to not treat the LSR or natural gas liquids, other refiners with excess capacity in
their FCC postreaters or naphtha hydrotreaters could purchase that volume, treat these streams
and blend the volume into their gasoline pool.
For the NPRM, we did not know whether butane being blended by refiners still has high
sulfur content and if refiners would need to treat it under Tier 3. We therefore assumed that
some refiners might have to treat butane using extractive desulfurization licensed by UOP
(Merox) or Merichem. A vendor we spoke to explained that almost all butane is being treated
today using extractive desulfurization and the final sulfur level is under 5 ppm.
In summary, to comply with a 10 ppm gasoline sulfur standard, refiners have a range of
options available to them that mostly involve reducing the sulfur content of the FCC naphtha. If
a refinery has a high pressure FCC pretreater, the refiner may be able to turn up the hydrotreating
severity of that unit. If a refinery has a low or medium pressure FCC pretreater and no
postreater, the refinery would likely need to either install a grassroots FCC postreater to comply
4-18
-------
with a 10 ppm gasoline sulfur standard, or reduce sulfur as much as possible with its current
capital and rely on the ABT program for the remainder. Refiners with FCC postreaters have
multiple options. If a refinery is short on octane and hydrogen, the refiner is likely to invest in
capital (e.g., a second stage reactor) to avoid as much octane loss and hydrogen consumption as
possible. However, if the refiner has a lot of excess octane and hydrogen, the refiner may choose
to avoid any capital cost investments or only make small capital investments and tolerate the
higher octane loss and hydrogen consumption by simply turning up the severity of its current
FCC postreater. Refineries with postreaters could always invest in an FCC pretreater
(hydrotreater or mild hydrocracker) to improve margins or produce more low sulfur diesel fuel.
Finally, in blending up their gasoline, some refiners may still be blending in some produced or
purchased gasoline blendstocks with high enough sulfur levels to be of concern when faced with
the Tier 3 sulfur standard. Several options exist for addressing the sulfur in these gasoline
blendstocks.
It should be noted that the preceding is EPA's best assessment of the steps refiners would
have to take to comply with Tier 3. Refiners may choose to pursue alternative strategies that
further other business objectives and also enable compliance with Tier 3 (e.g., installation of
hydrocrackers, conversion of FCC feed hydrotreaters to mild hydrocrackers). It is not possible
for EPA to project such alternative strategies on a refinery-by-refinery basis. While such
alternative strategies may be triggered by or timed with actions to comply with Tier 3, they are
not, and should not be, considered to be Tier 3 compliance actions.
4.2.3.4 Demonstrated Compliance with a 10 ppm Gasoline Sulfur Standard
Currently, there are multiple cases of refiners complying with 10 ppm or lower gasoline
sulfur programs. The State of California requires gasoline sold in the State to meet a 15 ppm
gasoline sulfur standard on average and a 20 ppm cap (California gasoline's per-gallon sulfur cap
dropped to 20 ppm on January 1, 2012). Furthermore, refiners can produce gasoline which
varies in composition, provided that the California Predictive Emissions Model (which, like
EPA's Complex Model, estimates vehicle emissions from fuels of varying composition)
confirms that the proposed fuel formulation meets or exceeds the emissions reduction that would
occur based on the default fuel requirements. California refineries are using the flexibility
provided by the Predictive Model to surpass the prescriptive standards for gasoline sulfur and are
producing gasoline which contains around 10 ppm sulfur on average. They are making this very
low sulfur gasoline despite using Californian and Alaskan crude oils which are heavier and more
sour than most other crude oils being used in the U.S. today. Thus, the experience in California
demonstrates that commercial technologies already exist to permit refiners to produce very low
sulfur gasoline.
Japan currently has a 10 ppm gasoline sulfur cap that took effect January 2008. Europe
also has a 10 ppm sulfur cap that has been adopted by the 30 Member States that comprise the
European Union (EU) and the European Free Trade Association (EFTA) as well as Albania and
Bosnia-Herzegovina. Under a 10 ppm cap standard, the gasoline sulfur level likely averages
about 5 ppm. Although gasoline in Japan and Europe is made from different crude oil sources
and much of the heavier ends are cut into diesel fuel, these international fuel programs (along
with California) provide evidence that advanced gasoline desulfurization technologies have been
deployed and are readily available enable compliance with the proposed Tier 3 fuel program.
4-19
-------
4.2.3.5 Improved Feasibility with the ABT Provisions
The averaging, banking and trading (ABT) and small refiner and small volume refinery
aspects of the proposed Tier 3 gasoline sulfur program would ease the feasibility of compliance
with the program. In the absence of the small refiner and small volume refinery provisions, all
refineries would have to comply with the 10 ppm gasoline sulfur standard by January 1, 2017.
Most refiners would have to make capital investments in their refineries to enable compliance
with the 10 ppm gasoline sulfur standard by this date. These investments include revamped FCC
pretreaters and postreaters, and the installation of grassroots FCC postreaters. As described
above, reaching 10 ppm sulfur in the gasoline pool is feasible by each refinery. However,
refiners assess the economic feasibility of their refineries differently depending on past and
expected future economic performance. They therefore have different tolerances for making
capital investments and absorbing increased operating costs. This is particularly true when
gasoline demand is projected to be flat and renewable fuel blending is expected to increase.
Refiners who own small refineries are concerned about the higher per-barrel costs for the capital
installed at those small refineries.
The small refiner and small volume refinery provisions will delay compliance for these
entities until January 1, 2020. Small refiners need more time because they have smaller
engineering staffs that they can dedicate to oversee the necessary refinery changes, thus they are
more likely to complete the necessary changes to their refineries later than large refiners. The
banking provisions of the ABT program effectively phase in the sulfur standard over six years
starting in 2014 through the end of 2019. The phase-in allows refiners to stagger their
investments to their economic advantage. Refineries that are expected to incur the lowest costs
for achieving lower gasoline sulfur levels can comply early and earn sulfur credits. These credits
can then be used to demonstrate compliance starting in 2017 by refineries that are expected to
incur higher costs for reducing their gasoline sulfur levels allowing those refineries to delay
investments for lowering their gasoline sulfur. This phase-in of the gasoline sulfur standard will
help spread out the various aspects of the construction process by the US refining industry
complying with the sulfur standard including: the preliminary design demands on the vendor
companies that license the desulfurization technology to refiners, the detailed design demands on
the engineering companies that provide that service to refiners, the permitting demands on the
states that must provide environmental permits to refiners, and the demands on the fabrication
shops that construct the reactors and other major hardware which must be installed at refineries
to realize the gasoline sulfur reductions. For more on how the proposed ABT provisions are
expected to help with lead time, refer to Section 4.3.
Finally, the averaging provisions of the ABT program will provide additional flexibility
and help to reduce the costs of the gasoline sulfur program. The averaging provisions will allow
refiners to reduce the gasoline sulfur levels to under 10 ppm at their lower cost refineries and
generate credits to sell to refiners who would purchase the credits for higher cost or financially
challenged refineries.
4.2.3.6 Implications of an Average Gasoline Sulfur Standard Less than 10 ppm
Although a more stringent sulfur standard than 10 ppm would increase the emission
benefits of Tier 3, there are practical reasons for finalizing a 10 ppm annual average sulfur
4-20
-------
standard instead of a more stringent standard, e.g., 5 ppm or even a 10 ppm cap imposed in parts
of Europe and in Japan. The lower the sulfur standard, the more costly it is for refiners to
achieve the lower sulfur standard, and a 5 ppm average standard is much more costly than a 10
ppm average standard. We identified several reasons why the costs increase so much for more
deeply desulfurizing the gasoline pool.
First, as desulfurization severity increases, the operating and capital costs associated with
desulfurizing FCC naphtha also increases. FCC naphtha is very rich in high-octane olefins. As
the severity of desulfurization increases, more olefins are saturated, further sacrificing the octane
value of this stream and further increasing hydrogen consumption. Also, as desulfurization
severity increases, there is an increase in the amount of the removed sulfur (in the form of
hydrogen sulfide) which recombines with the olefins in the FCC naphtha, thus offsetting the
principal desulfurization reactions. There are means to deal with the recombination reactions;
however, this probably means either higher hydrogen consumption and octane loss, or greater
capital investments. For example, the most expensive capital investment for an FCC postreater
is a two stage desulfurization unit. A sulfur standard less than 10 ppm would likely require more
refiners to invest in a second stage for their FCC postreater.
Second, as shown in Table 4-2, other refinery streams contain very modest amounts of
sulfur, yet a 5 ppm sulfur standard would likely require desulfurization of some of these streams.
Because refineries have different sulfur levels in their non-FCC streams based on their feedstock
sulfur levels and their configurations, those with higher sulfur levels in other refinery streams
may have to desulfurize additional streams. Each additional individual gasoline stream that
requires desulfurization is incrementally a lot more expensive than addressing the sulfur from the
FCC unit because more volume has to be processed. The amount of sulfur reduction is a lot
lower, and the capital costs are higher on a per-barrel basis for lower volume gasoline blendstock
streams.
Third, further desulfurization of gasoline down to 5 ppm essentially removes the
flexibility offered by the 10 ppm gasoline sulfur standard with the ABT program. Each U.S.
refinery is in a different position today, both technically and financially, relative to the other
refineries. In general, they are configured to handle the different crude oils they process and turn
their crude oil slate into a widely varying product slate to match their available markets. Those
processing heavier, sour crudes would have a more challenging time reducing gasoline sulfur
under the proposed Tier 3 program. Also, U.S. refineries vary greatly in size (atmospheric crude
capacities range from less than 5,000 to more than 500,000 barrels per day) and thus have
different economies of scale for adding capital to their refineries. As such, it is much easier for
some refineries to get their sulfur levels below 10 ppm than for others to reach 10 ppm. This
allows the ABT program to be used to reduce the cost of the proposed gasoline sulfur standard.
If the gasoline sulfur standard were to be 5 ppm, the ability of refiners to average sulfur
reductions across their refineries would likely end and thus, significantly increase the capital and
operating costs while significantly reducing the desulfurization flexibility.
Our cost estimates for a 5 ppm average standard as compared to a 10 ppm average sulfur
standard bears this out. We estimate the average cost for a 10 ppm gasoline sulfur standard
(assuming nationwide credit trading) to be 0.65 eVgal compared to 1.27 eVgal for the 5 ppm
standard. The cost per sulfur reduction (marginal cost) for the 10 ppm average standard is 0.65
4-21
-------
0/gal for the 20 ppm sulfur reduction from Tier 2, which averages 0.045 0/gal for each ppm of
sulfur reduction. The marginal cost for the 5 ppm standard is 0.49 0/gal for the 5 ppm sulfur
difference from the 10 ppm average standard, which averages 0.098 0/gal per each ppm of sulfur
reduction, which is over 2 times higher. Therefore, we believe that an annual average standard
of 10 ppm at the refinery gate with an ABT program is reasonable and maximizes the amount of
sulfur reduction and the associated emission reductions before the costs begin to steeply escalate.
We note that in most European countries and Japan, the gasoline sulfur level is capped at
10 ppm. We, however, are not considering a 10 ppm cap for the U.S. due to the increased cost
and increased challenges of ensuring compliance for every batch of fuel. The cost estimates
described above for 5 ppm do not capture any additional costs refiners might need to incur to
deal with offspec batches of fuel that get produced. Therefore, we are finalizing a 10 ppm
average sulfur standard coupled with 80 and 95 ppm caps at the refinery gate and downstream,
respectively, similar to what currently exists under the Tier 2 program. We believe this is the
most prudent approach for lowering in-use sulfur while maintaining flexibility considering cost
and other factors. These per-gallon caps are important in the context of an average sulfur
standard to provide an upper limit on the sulfur concentration that vehicles must be designed to
tolerate. Since there are many opportunities for sulfur to be introduced into gasoline downstream
of the refinery, these caps also limit downstream sulfur contamination and enable the
enforcement of the gasoline sulfur standard in-use. For more on our consideration of
downstream caps, refer to Section 4.2.4.2.
4.2.4 Challenges with Lowering Today's Sulfur Caps
4.2.4.1 Impacts of Lowering the 80 ppm Refinery Cap
We are maintaining the existing 80 ppm refinery gate cap standard in the final Tier 3 fuel
standard. For the NPRM, we considered lowering this cap to either 50 ppm or 20 ppm. If we
lowered the refinery cap standard to 20 ppm, then refiners would only be able to take advantage
of very little of the averaging aspect of the ABT program. That is because, under a 20 ppm cap
standard, we estimate that the maximum sulfur level that refineries could average is about 14
ppm sulfur. Thus, the compliance scenario if the cap standard were 20 ppm would essentially be
the same as the non-ABT case we analyzed. In this case, refiners would have little of the
flexibility offered by the ABT program.
If the cap standard were to be lowered to 50 ppm, the final compliance scenario under the
Tier 3 fuels program would be somewhere between the ABT scenario that we analyzed and the
non-ABT scenario that we analyzed (probably much closer to the ABT case). Under a 50 ppm
cap standard, we estimate that the maximum average gasoline sulfur level that refineries could
average is 35 ppm. Although EPA batch data shows 40 refineries that averaged between 35 and
80 ppm sulfur during 2011, our cost modeling shows only 8 of those would continue to average
more than 35 ppm under a 10 ppm average standard and an 80 ppm cap. If the 80 ppm cap were
to be reduced to 50 ppm, we project that those 8 refineries that averaging over 35 ppm would be
forced to reduce their sulfur levels below 35 ppm regardless of their compliance costs. . Thus,
the 10 ppm average standard reduced the number of refineries that average greater than 35 ppm
sulfur from forty to eight.
4-22
-------
A more stringent cap would also affect refiners' ability to process high sulfur FCC
naphtha when there is a short term shutdown of the FCC postreater. If the FCC postreater goes
down, the refinery would likely continue operating the FCC unit and store up the high sulfur
FCC naphtha. Since the FCC naphtha is too high in sulfur to blend directly with gasoline, the
refinery would have to either sell the high sulfur FCC naphtha to other refiners, or hydrotreat the
stored up FCC naphtha along with the ongoing hydrotreating of high sulfur FCC naphtha once
the FCC postreater was back online. If a stringent cap were in place, the refiner would have little
room for short term production of higher sulfur gasoline if it was feeding a larger than normal
quantity (stored and new production) of FCC naphtha to the FCC postreater. A stringent sulfur
cap may cause refiners to oversize the FCC postreater and add additional FCC naphtha storage to
ensure that, regardless of the higher feed volume needed to process the stored material, the FCC
naphtha desulfurization unit could continue to desulfurize the FCC naphtha down to the required
sulfur level to comply with the cap standard and ultimately the 10 ppm average standard. If the
cap were to be lowered, a 50 ppm cap standard would continue to provide refiners with some
flexibility while a 20 ppm cap would not. As shown in Section 4.4.2, refiners are currently
taking full advantage of the flexibility offered by the 80 ppm cap standard offered under Tier 2.
4.2.4.2 Downstream Sulfur Caps
The feasibility of complying with a downstream sulfur cap is dependent on the
differential between the refinery/importer gate sulfur cap and the downstream cap. This
differential must provide sufficient flexibility for worst-case situations when the potential
sources of sulfur addition downstream of the refinery/importer compound in a single batch of
gasoline that was introduced into the system at the refinery/importer gate sulfur cap.
We proposed, and are finalizing an 80 ppm refinery gate sulfur cap and 95 ppm
downstream sulfur cap. These requirements are applicable under the current Tier 2 gasoline
program. Therefore, under the Tier 3 program, we are maintaining the same 15 ppm differential
between refinery gate and downstream sulfur caps that currently exists under the Tier 2 program.
This 15 ppm differential has proven to be sufficient to accommodate the unavoidable addition of
sulfur downstream of the refinery gate from contamination during distribution, the use of
additives, and the disposition of transmix generated during distribution.
The downstream sulfur cap applies at all locations downstream of the refinery or importer
gate including the gasoline produced by transmix processors and after the use of additives. The
potential sources of sulfur addition downstream of the refinery/importer gate and issues
associated with the feasibility of meeting the downstream sulfur cap are discussed in the
following subsections.
4.2.4.2.1 Sulfur Addition Downstream of the Refinery and Importer Gate
The sulfur content of gasoline can increase downstream of the refinery/importer due to
contamination during distribution, the use of additives, and the disposition of transmix generated
during distribution.
A small amount of sulfur contamination takes place during distribution as a result of the
shipment of gasoline over long distances by pipeline and other modes due to the sharing of the
4-23
-------
same distribution assets with other higher-sulfur petroleum products, e.g., jet fuel. Steps can be
taken to limit sulfur contamination. However, it is an unavoidable feature of the efficient multi-
product distribution system in the U.S. We estimate that sulfur contamination of gasoline can be
limited to a worst case maximum of 3 or 4 ppm in the future, even for the most involved and
long-distance distribution pathways. Typical levels of sulfur contamination will likely be much
lower given the removal of many sources of sulfur contamination in the fuel distribution system
in recent years.
There were no direct regulatory controls on the sulfur content of gasoline additives under
the Tier 2 program. The contribution to the sulfur content of finished gasoline from gasoline
additives is accommodated in the differential between the refinery gate and downstream sulfur
caps. The functional components of some gasoline additives such as silver corrosion inhibitors
and demulsifiers are inherently high in sulfur content. However, the contribution to the overall
sulfur content of the finished fuel is very limited. For example, silver corrosion inhibitors can
contain as much as 30 percent sulfur but because of very low treatment rates can add only 0.17
ppm to the sulfur content of the finished fuel.33 At seldom used highest treatment rates, the use
of gasoline additives upstream of the consumer has the potential to add ~1 ppm to the sulfur
content of the finished fuel. Aftermarket additives that are added directly into the vehicle fuel
tank also have the potential to increase gasoline sulfur content. One particular aftermarket
performance and anti-wear additive can contribute ~2 ppm sulfur to the treated fuel.B34
Transmix is a necessary byproduct of the multi-product refined product pipeline
distribution system. Batches of different products are shipped in sequence in pipelines without
any physical barrier between the batches. Transmix is produced when the mixture at the
interface between two adjacent products cannot be cut into either batch. Transmix typically
accumulates at the end of pipeline systems far from refineries. There are two methods of
disposing of transmix. Most transmix is sent to transmix processing facilities for separation into
saleable distillate and gasoline products through use of a simple distillation tower.
The other means of transmix disposal is for pipeline operators to blend small quantities
directly into batches of gasoline during shipping. This typically takes place at remote pipeline
locations where small volumes of transmix accumulate that would be difficult to consolidate and
ship to transmix processors. Pipeline operators that blend transmix into the gasoline in their
systems must ensure that the resulting gasoline meets all fuel quality specifications and the
endpoint of the blended gasoline does not exceed 437 °F.C35 This practice currently can add as
much as 3 to 5 ppm to the sulfur content of gasoline although we believe that the contribution is
typically less.
Transmix processing facilities do not handle sufficient volumes to support the installation
of currently-available desulfurization units. Therefore, the sulfur content of the products they
produce is predominantly governed by the sulfur content of the transmix they receive. In many
cases, transmix contains jet fuel which can have a sulfur content as high as 3,000 ppm. Due to
B Aftermarket additives are defined as additives sold to vehicle operators for direct addition to vehicle fuel tanks.
c 437 F is the maximum endpoint allowed for gasoline in the ASTM International specification for gasoline in
ASTMD4814.
4-24
-------
the overlapping distillation characteristics of jet fuel and gasoline, it is unavoidable that some jet
fuel in transmix will be present in the gasoline produced by transmix processors.
Transmix processors produce ~0.1 percent of all gasoline consumed in the U.S. The
small volume of transmix-derived gasoline along with the fact that such gasoline is typically
mixed with other gasoline before delivery to the end user, substantially limits the potential
impact on gasoline sulfur levels. Furthermore, data provided by the largest operator of transmix
processing facilities, shown in Figure 4-2, indicates that relatively few batches of the gasoline
they produce approach 80 ppm sulfur.36 Most batches are approximately 10 ppm above the
current 30 ppm refinery sulfur average. We anticipate that this 10 ppm differential would likely
continue under the 10 ppm refinery average sulfur standard.
90
70
60
10
•
*
: .
A * * A*
;-*- T^
A A . *A4 i ^ •• A*
• • ** > Ł^ A4* A* «F
J .A AA ". 4SV 44 A... A»^ ^ A^
»lndianol3
• St. Louis
A Deepwater
X Greensboro
V Dorsey
Colton
10/14/2009 01/22/2010 05/02/2010 08/10/2010 11/18/2010 02/26/2011 06/06/2011 09/14/2011
Figure 4-2 Kinder Morgan Transmix Gasoline Product Sulfur Levels
4.2.4.2.2 Maintaining the Current 15 ppm Differential Between the Refinery
/Importer Gate and Downstream Sulfur Caps
The 80 ppm refinery gate and 95 ppm downstream sulfur caps finalized today maintains
the current 15 ppm differential between the refinery/importer gate sulfur cap and the downstream
sulfur cap.
4-25
-------
The current 15 ppm differential was established under the Tier 2 program to
accommodate the sulfur contamination during distribution, the sulfur contribution from transmix
blending by pipeline operators, the sulfur contribution from the use of additives, and to enable
compliant gasoline to be produced by transmix processors. Transmix processors need to produce
gasoline sufficiently below the downstream sulfur cap to accommodate the addition of sulfur
from the use of additives and contamination during further distribution. Experience under the
Tier 2 program has shown that a 15 ppm differential is sufficient for downstream parties to
ensure compliance with the downstream sulfur cap
Maintaining the current 95 ppm downstream sulfur cap with an 80 ppm refinery/importer
gate sulfur cap under the Tier 3 program represents no change from current requirements. As a
result, there will be no increased difficulty or additional costs associated with satisfying a 95
ppm downstream sulfur cap beyond those that were already incurred under the Tier 2 program.
Furthermore, the reduction in the refinery average sulfur standard under the Tier 3 requirements
may make it somewhat easier to comply with the downstream sulfur cap given that most gasoline
produced would be at or near 10 ppm sulfur.
4.2.4.2.3 Cap on Sulfur Contribution to Gasoline from the use of a Gasoline
Additive
The Tier 3 rule requires that each gasoline additive may add no more than 3 ppm sulfur to
the sulfur content of gasoline when used at the maximum recommended treatment rate. All
current gasoline additives are currently compliant with this requirement. Therefore,
implementing this requirement will not place an additional burden on gasoline additive
manufactures. We are implementing this requirement to preclude the possibility that high sulfur
blendstocks might be added to gasoline in the guise of a gasoline additive.
4.3 Sulfur Credits
We conducted an analysis of 2012 Tier 2 Gasoline Sulfur Credit Banking and Allotment
Generation Reports submitted to EPA by U.S. refineries to ascertain the availability of sulfur
credits and the fluidity of the sulfur credit trading market. These reports must be submitted by
producers and importers of gasoline destined for sale in the United States. Such facilities submit
Credit Banking and Allotment Generation Reports for sulfur credits it possesses or possessed
over a given year. This data is Confidential Business Information.
Sulfur credits must be used by the refiner that generated them or they can be transferred
up to two times. These credits may be used at the refinery where they were generated, banked by
a refiner for future use or use at another one of its refineries, or sold/transferred to another
refiner. If a transferee does not use credits that they purchased, they may transfer them to
another party; the second transferee must then use or bank the credits (i.e., they cannot be
transferred again). Credits have a five-year life span from the year of generation. As such, 2007
credits expired at the end of 2012.
We performed several analyses of the data including: (a) the use of intercompany
gasoline sulfur credit trading, (b) 2012 gasoline sulfur credit balances, (c) gasoline sulfur credit
usage by age, and (d) the reduction of gasoline sulfur levels over time.
4-26
-------
The analysis revealed the number of gasoline sulfur credits banked by refiners as well as
the extent to which gasoline sulfur credits are traded within a parent company (intra-company)
and between competing parent companies (intercompany). These metrics, coupled with
information on the gasoline sulfur credit usage and the reduction of gasoline sulfur levels over
time under the current Tier 2 program, provides an indication of the ability and flexibility of
refineries to rely on trading of credits for compliance with the Tier 3 standards being finalized
today.
4.3.1 Intra- and Inter-Company Trading
Our analysis found that approximately 44% of sulfur credits transacted in 2012 were
transferred intra-company whereas approximately 56% of gasoline sulfur credits transacted were
transferred between competing parent companies (Figure 4-3), suggesting the existence of a
robust and fluid gasoline sulfur trading market.
2012 Inter-vs. Intracompany
Gasoline Sulfur Credit Transfers
Transfer 44%
56%
Figure 4-3 2012 Inter- vs. Intra-company Gasoline Sulfur Credit Transfers
4.3.2 Tier 2 Sulfur Credit Analysis
Credits under the current Tier 2 program have a five year life, which means they could
still be used in 2017. For the final Tier 3 program, we are allowing Tier 2 credits to be carried
over for use in complying with Tier 3. As a result, we assessed gasoline sulfur credit availability
for 2012. Our analysis of the number of sulfur credits generated and the number of sulfur credits
used and transferred resulted in a difference of just under 400 billion ppm-gallon sulfur credits in
2012. This equates to 2-3 months of compliance with the 10 ppm Tier 3 standard. If refiners
were to simply continue to accrue at this rate until 2017 without taking any additional actions to
comply early with Tier 3, gasoline sulfur credits generated by refineries would afford
approximately a one-year delay in implementation of the standard.
Of the gasoline sulfur credits transacted in 2012, 40% were generated in 2007, 30% were
generated in 2008, 7% were generated in 2009, 9% were generated in 2010, 11% were generated
4-27
-------
in 2011, and 3% were generated in 2012 (Figure 4-4). Thus, a refiner's willingness to trade
gasoline sulfur credits varies with the time remaining until credits expire. Gasoline sulfur credits
set to expire at the end of 2012 (i.e., credits generated in 2007) were traded thirteen times more
often than sulfur credits generated in 2012, which expire at the end of 2017. Thus, having a
finite credit life is helping to stimulate the robustness of the current credit trading market.
p<
T
Ł 50%
irt
±:
T3 /i no/ 4
Percentage Cre
h- > KJ (JJ 4
O O O O C
^v Łv Łv ^v c
NJ
0
srcentageof Gasoline Sulfur Credits
ransacted in 2012 by Creation Year
|L
^^^^
^^
\
>- *~ t^-^ -^
07 2008 2009 2010 2011 2012
Year of Credit Generation
Figure 4-4 Percentage 2012 Gasoline Sulfur Credits Transacted by Year of Generation
4.4 Sulfur Level Analysis
4.4.1 Volume-Weighted Gasoline Sulfur
The analysis also revealed that volume-weighted average gasoline sulfur levels have
decreased steadily over the period for which we analyzed data. In 2007, the volume-weighted
gasoline average sulfur concentration was 39.8 ppm, 35.2 ppm for 2009, 31.2 ppm for 2011, and
26.7 for 2012 (Figure 4-5). The steady decline over time was driven by the continued pressure
of the Tier 2 standards. The last flexibilities in the program for a subset of small refiners ended
in 2011. The additional drop in 2012, however, represents over-compliance with the Tier 2 30
ppm average standard.
4-28
-------
Volume-Weighted Average Gasoline
Sulfur Level by Year
2009 2010
Year
2011
2012
Figure 4-5 Volume-Weighted Average Gasoline Sulfur Level by Year
Sulfur levels by summer cumulative gasoline volume for 2011 are depicted in Figure 4-6.
80
70
_ 60
a. 50
Q.
r 40
^30
" 20
10
0
0
Sulfurvs. Summer Cumulative
Gasoline Volume for 2011
_^-~-
^
_. '
»— -"-^
_^— •-
^-
^*
^
j
/
^
% 20% 40% 60% 80% 100%
Cumulative Gasoline Volime {% Total)
Figure 4-6 Sulfur Levels versus Summer Cumulative Gasoline Volume for 2011
4.4.2 Batch-to-Batch Sulfur Variability
Our analysis also revealed the extent of batch-to-batch variability of gasoline sulfur.
While such variability is suggested when examining the aggregated sulfur levels for all gasoline
batches produced in a year, as in Figure 4-7, the implications of this variability is not clearly
evident until gasoline sulfur is examined at the refiner-level.
4-29
-------
Aggregated Sulfur Levels for All
Gasoline Batches Produced in 2011
Innn
cnn
>*
o
C^nn
QJ
=1
O"
li
i nn
1
0
1
III
III II Hum
10 20 30 40 50 60 70 80 90
Sulfur(ppm)
Figure 4-7 Aggregated Sulfur Levels for All Gasoline Batches Produced in 2011
Average gasoline sulfur varies significantly between refiners as well as between gasoline
batches produced by the same refiner, as can be seen in Figures 4-8a-d. These figures depict
2011 gasoline sulfur levels by batch frequency for four refineries from the analysis. In addition
to providing insight into the volume-weighted average gasoline sulfur and the minimum and
maximum gasoline sulfur, the figures also capture the broad distribution of sulfur levels and their
skewness for gasoline batches produced by refiners.
For instance, the average gasoline sulfur for Refiner A (Figure 4-8a) is relatively high
(58.6 ppm), with a broad range between minimum (8 ppm) and maximum (80 ppm). However,
the data are negatively skewed, with most of the observations centered around 70 ppm - well-
above the volume-weighted average - and, to a lesser extent, along a shallow tail to the left (with
lower sulfur values). These lower values offset the higher observations centered about 70 ppm.
Refiner A
Sulfur (ppm)
Figure 4-8a Sulfur Concentration Versus Batch Frequency for Refiner A
4-30
-------
Average sulfur for Refiner B is considerably lower than Refiner A (19.3 ppm versus 58.
ppm), and the data have a much smaller range (4 ppm minimum and 42 ppm maximum), with
only a single higher observation at 63 ppm observation (Figure 4-8b). Unlike Refiner A, the
gasoline sulfur for Refiner B is distributed bimodally about 15 and 25 ppm and is fairly
systematical about the average of 19.3 ppm. The sulfur observations for Refiner B are not
negatively skewed, as is the case for Refiner A.
Refiner B
V 50 C 60 • 7
Sulfur (ppm)
Figure 4-8b Sulfur Concentration Versus Batch Frequency for Refiner B
Like Refiner A, the volume-weighted average sulfur for Refiner C is relatively high (52.7
ppm), with a similarly broad range between minimum (20 ppm) and maximum (76 ppm) (Figure
4-8c). However, unlike Refiner A - which is negatively skewed - these data are largely
symetrical about the volume-weighted gasoline sulfur average.
'.0
25
20
,15
10
5
0
80
Refiner C
52.7 ppm4
ID
20
30 40
Sulfur (ppm)
60
Figure 4-8c Sulfur Concentration Versus Batch Frequency for Refiner C
4-31
-------
Like Refiner B, the volume-weighted average sulfur for Refiner D is relatively low (21.5
ppm), with a similarly broad range between 4 ppm and 51 ppm, with just a few observations
between 70-75 ppm (Figure 4-8d). And as with Refiner B, the data for Refiner D are largely
distributed symmetrically about the volume-weighted sulfur average. See Table 4-3 for a
summary of these parameters.
Refiner D
Sulfur jpprri
Figure 4-8d Sulfur Concentration Versus Batch Frequency for Refiner D
Table 4-3 Summary of Statistics for Four Typical Refiners
Refiner
Refiner A
Refiner B
Refiner C
Refiner D
Sulfur
Average
58.6
19.3
52.7
21.5
Sulfur
Minimum
8.0
4.0
20.0
4.0
Sulfur
Maximum
80.0
63.0
76.0
75.0
Sulfur
(25th%)
44.0
13.0
45.0
16.0
Sulfur
(75th%)
70.0
25.0
58.0
26.0
Given the significant variability in batch-to-batch gasoline sulfur, the data suggest that a
low sulfur average with a high sulfur cap allows refiners to minimize operating costs by
providing flexibility to those refiners producing gasoline above the average standard while
providing incentive to those refiners producing gasoline below the average standard.
4-32
-------
In addition, the data demonstrate the flexibility provided by a high cap to allow normally
low sulfur gasoline-producing refiners to still market occasional higher sulfur batches of gasoline
during abnormal operating conditions. Finally, the data show that the combination of an average
standard with a higher cap allows refiners considerable batch-to-batch flexibility in producing
individual batches of gasoline while still reducing the overall sulfur level.
4.5 Lead Time Assessment
We received a several comments in support of and against our proposed rule regarding
feasibility assessment and lead time. Commenters in the refining industry generally stated that
the amount of lead time proposed was not sufficient. These commenters noted concerns that the
short lead time proposed would drive up costs as there would be unscheduled shut-downs to
install and/or revamp equipment to meet the Tier 3 sulfur standard, and would not provide
enough time for the permitting process. These commenters requested at least five years of lead
time, and noted that EPA has historically provided at least four years of lead time in previous
fuels rulemakings. Commenters in the auto industry, as well as states and non-governmental
organizations (NGOs), encouraged us to finalize the rule as soon as possible and to retain the
January 1, 2017 start date to harmonize our program with California's LEVIII program and to
enable Tier 3 benefits as soon as possible. As discussed in more detail below, we believe the
amount of lead time provided is sufficient, especially given the flexibilities being provided. A
complete discussion on the comments received with regard to lead time can be found in Chapter
5 of the Summary and Analysis of Comments document.
While evaluating the merits of a national gasoline sulfur program to reduce emissions and
enable future vehicle technologies, we also considered the refining industry's ability to reduce
sulfur to 10 ppm on average by January 1, 2017 and the associated costs (for more on fuel costs,
refer to Chapter 5). Based on information gathered from numerous stakeholder meetings and
discussions with vendor companies that provide the gasoline desulfurization technologies both
before and after the proposal, as well as the results from our refinery-by-refinery modeling, we
believe it is technologically feasible at a reasonable cost for refiners to meet the sulfur standards
in the lead time provided. A summary of our feasibility analysis is presented below.
4.5.1 Employment Constraint Analysis
As in prior rules, we also evaluated the capability of E&C industries to design and build
gasoline hydrotreaters as well as performing routine maintenance. This includes an employment
analysis. Two areas where it is important to consider the impact of the fuel sulfur standards are:
1) refiners' ability to procure design and construction services and 2) refiners' ability to obtain
the capital necessary for the construction of new equipment required to meet the new quality
specification. We evaluated the requirement for engineering design, and construction personnel,
in a manner consistent with the Tier 2 analysis, particularly for three types of workers: front-end
designers, detailed designers and construction workers, needed to implement the refinery
changes. We developed estimates of the maximum number of each of these types of workers
needed throughout the design and construction process and compared those figures to the
number of personnel currently employed in these areas.
4-33
-------
The number of person-hours necessary to design and build individual pieces of refinery
equipment and the person-hours per piece of equipment were taken from Moncrief and
Ragsdale37. Their paper summarizes analyses performed in support of a National Petroleum
Council study of gasoline desulfurization, as well as other potential fuel quality changes. The
design and construction factors for desulfurization equipment are summarized in Table 4-4.
Table 4-4 Design and Construction Factors"
Gasoline Refiners
Number of New Pieces of Equipment per Refinery
Number of Revamped Pieces of Equipment per Refinery
60
15
Job Hours Per Piece of New Equipment"
Front End Design
Detailed Design
Direct and Indirect Construction
300
1,200
9,150
Note:
a Revamped equipment estimated to require half as many hours per piece of equipment
Refinery projects will differ in complexity and scope. Even if all refiners desired to
complete their project by the same date, their projects would inevitably begin over a range of
months. Thus, two projects scheduled to start up at exactly the same time are not likely to
proceed through each step of the design and construction process at the same time. Second, the
design and construction industries will likely provide refiners with economic incentives to avoid
temporary peaks in the demand for personnel.
Applying the above factors, we projected the maximum number of personnel needed in
any given month for each type of job. The results are shown in Table 4-5. In addition to total
personnel required, the percentage of the U.S. workforce in these areas is also shown, assuming
that half of all projects occur in the Gulf Coast in Table 4-5. Refineries are generally expected to
not require the full 24-month period to complete scoping studies, process design, permitting,
detailed engineering, field construction, and start-up/shakedown for revamping an existing FCC
postreater.
4-34
-------
Table 4-5 Maximum Monthly Demand for Personnel
Front-End
Design
Detailed
Engineering
Construction
Tier 3 Gasoline Sulfur Program
Number of Workers
Percentage of Current Workforce51
51
3%
202
2%
1,503
1%
Note:
" Based on current employment in the U.S. Gulf Coast assuming half of all projects occur in the Gulf Coast
To meet the Tier 3 sulfur standards, refiners are expected to invest $2.0 billion between
2014 and 2019 and utilize approximately 250 front-end design and engineering jobs and 1,500
construction jobs. The number of estimated jobs required is small relative to overall number
available in the U.S. job market. As such, we believe that because of the ABT program with its
flexibilities such as generous early credit generation period, that there is adequate lead time for
refineries to obtain necessary permits, secure E&C resources, install new desulfurization
equipment and make all necessary retrofits to meet the sulfur standards.
We conducted a refinery-by-refinery analysis to determine the impacts on refinery E&C
demand of implementing the 10 ppm standard without an ABT program. The analysis suggests
that a greater number of refineries would need to make investments in refinery apparatus and
upgrades without an ABT program than would be required with an ABT program. This would
result in a greater demand on the E&C industry. Moreover, the analysis also indicated that the
demand upon the E&C industry would be spread over a shorter period than with the ABT case.
In particular, our refinery-by-refinery analysis indicates that without an ABT program, 72
refineries would revamp existing pre- and postreaters and 18 would install grassroots postreaters
in order to meet the Tier 3 sulfur standards. The remaining 18 refineries are either already in
compliance with the 10 ppm standard or expected to comply with simple process changes. This
is compared to 66 refineries that would revamp existing pre- and postreaters and one refinery
that would install grassroots postreaters in order to meet the Tier 3 sulfur standards under an
ABT program.
4.5.2 Can Refiners Meet the January 1, 2017 Start Date?
An adequate amount of lead time is required for the implementation of any rulemaking.
Depending on the level of effort required to comply, more or less lead time is also required. In
the case of Tier 3, refiners need time to select the technology and the vendor that will provide the
technology with which they will comply with the fuels standard. Next, they need time to arrange
an engineering and construction (E & C) contractor which will design and oversee the
construction of the refinery unit and the time needed to obtain the necessary permits and procure
the necessary hardware. Next, refiners need time to construct the unit. Finally, the refiner needs
time to make the necessary unit tie-ins of the unit with the rest of the refinery and then startup
the unit.
This section explains that when taking into account the time to revamp existing FCC
postreater units or build grassroots postreater units, tie-in the new or revamped units with the rest
4-35
-------
of the refinery and considering the flexibility offered by the ABT program, refiners will be able
to comply with the Tier 3 program within the lead time provided.
4.5.2.1 Time Required to Install Grassroots Units and Revamp Existing Units
The technologies for complying with Tier 3 are well known and well proven. As
previously explained, refiners which complied with Tier 2 using FCC naphtha desulfurization
technologies installed the following units, all of which were grassroots installations: Ax ens
Prime G+, CDTech's CDHydro and CDHDS, UOP's ISAL, Sinopec's S-Zorb and Exxon's
Scanfming. Prior to choosing a technology, refiners needed to evaluate each these different
technologies and choose among them, all of which were largely untested at the time, which
required us to provide more lead time for Tier 2. Since it has been 9 years since the Tier 2 sulfur
standard began to be phased in, refiners now have direct experience with the installation and
operation of these technologies and the vendor companies that license them and continue to
support their installations onsite. This fact will allow refiners to reach a decision very quickly
when complying with Tier 3, particularly, because in most cases the refiners will solely be
revamping the units installed for Tier 2 when complying with Tier 3.
Based on our discussions with refiners, construction companies, vendor companies and
from published literature, we estimated the time it takes to revamp existing postreaters and install
grassroots postreaters. Revamping an existing postreater is expected to require up to two years.
Installing a grassroots postreater is estimated to require three years. Figure 4-6 reflects these
project completion times showing the various major intermediate steps for completing the
projects.0
D The timeline shows overlap between steps which reflect actions that can be taken to set up the following step
while the previous step is being completed. For example, refines can establish the contract for the detailed
engineering while the process design is being completed. Refiners can begin site preparation to prepare for
construction before the detailed design is completed. Finally, refiners can test individual pieces of equipment as
they are installed and while construction is ongoing to find problems that would streamline unit start-up and avoid
delays
4-36
-------
Figure 4.6 Estimated Project Lead Time for Revamps and Grassroots Units
We believe that the revamping of postreaters could take less time than what we estimate
in Figure 4-6 because many of the Tier 3 revamps are expected to be very modest (e.g., change
out a reboiler or heat exchanger). Since EPA held discussions with many refiners in 2011 about
EPA's plan to pursue additional sulfur control post-Tier 2 (Tier 3), refiners began the process of
assessing how they would comply. The Tier 3 proposal was delayed for about a year and it is
our understanding from recent discussions with vendor companies and some refiners that, during
this time, many refiners began assessing how they intended to comply with Tier 3. Thus, many
refiners likely have completed the scoping studies, which involves technology selection, and in
the case of grassroots units, vendor selection as well (refiners with a particular postreater
technology in most cases are expected to simply revamp the same vendor's technology, so there
is no need to select a vendor). If refiners have already completed their scoping studies, we
estimate that installation of the revamps or grassroots units would be about 3 months shorter than
the 2 and 3 years, respectively, than we estimate in Figure 4-6.
These project timelines are reasonable in light of past industry experiences that show
FCC postreaters being installed in refineries in less time than what we estimate. At the Motiva
refinery in Port Arthur, TX, a grassroots CDTech postreater was designed, constructed and
TO
started up in less than 2 years. At two refineries in Germany, two Prime G+ units were
designed, constructed and started up - one of them in two years, and the other in 18 months.39
4-37
-------
As an extreme example, the $3.6 billion dollar, 180 kbbl/day crude oil expansion at Marathon's
Garyville, LA refinery was designed, constructed and started up in less than 4 years.40 This
single project involved the construction of 10 major refinery units, and permitting required only
9 months. Since these may be best case examples, we continue to believe the projections
provided above in Figure 4-6 are reasonable.
4.5.2.2 Program Flexibility that Extends Lead Time
The final Tier 3 program includes an ABT program that would significantly help refiners
comply with the January 1, 2017 start date. There are three provisions of the ABT program
which helps with respect to leadtime.
The ABT program allows for ongoing intra-company and inter-company trading
nationwide. This will allow some refineries to over-comply with the 10 ppm gasoline sulfur
standard (in our analysis, we modeled these refineries bringing their gasoline down to 5 ppm),
allowing other refineries that would otherwise need to install grassroots units to not invest and
purchase credits instead. This aspect of the ABT program is very important because our analysis
estimates that only one refinery would need to install a grassroots hydrotreater whereas without
the ABT provisions, there could be as many as 20 grassroots units. This one aspect has
important implications for leadtime because as discussed in the previous subsection, revamps
require two years or less whereas grassroots FCC postreater units require approximately three
years to install. We are convinced that this aspect of the ABT program will be utilized to the
maximum extent possible because refineries revamping their postreaters in lieu of installing
grassroots postreaters results in the most cost-effective mechanism for meeting the 10 ppm
annual average standard.
An important question is whether refiners will not invest in a grassroots unit trusting that
the credits will be freely available. For the NPRM, we conservatively assumed that refiners
would only rely on credits if they could generate them internal to the company. As discussed in
Section 4.4, we assessed how the sulfur credits were being traded under Tier 2 and we found that
over half the sulfur credits were freely traded between companies (as opposed to only being used
within companies), and many single-refinery companies had sulfur levels above 30 ppm (single-
refinery companies must purchase credits from other companies). Because we set up the Tier 3
credit trading program to work just like the Tier 2 credit trading program, we are confident that
there will be widespread trading within and between refining companies which means that few
grassroots units will be need to be built for Tier 3.
A second aspect of the ABT program that helps with leadtime is the provision for
generating early sulfur credits and banking them for later use. This provision allows refineries to
reduce their gasoline sulfur to less than 30 ppm prior to January 1, 2017 and bank the credits for
later use. Based on comments that we received on the proposed rule, we are allowing Tier 2
credits which are generated during the years 2012 and 2013 to also be used to show compliance
for Tier 3. This effectively extends the early credit generation period for Tier 3 to encompass the
years 2012 to 2016, which is 5 years. Analyzing the 2012 gasoline quality data that refiners
reported to EPA, we found that gasoline sulfur levels in the U.S. averaged 26.7 ppm. Thus,
refiners have already begun overcomplying with Tier 2 by 3.3 ppm, and are therefore already
generating early credits for Tier 3. If refiners do nothing more but continue to overcomply with
4-38
-------
Tier 2 by 3.3 ppm over the 5 years of early credit generation, refiners will have generated enough
credits to delay the completion of their capital projects by more than one year. Furthermore
since those credits generated in 2012 and 2013 will expire in 2017 and 2018 respectively,
refiners will have an incentive to either use them themselves or trade them in 2017 and 2018.
Thus refiners that may need to count on them to delay their capital investment are likely to be
able to have access to them.
We believe that refiners will generate a lot more early credits with their existing gasoline
sulfur control units than the 3.3 ppm we observed in 2012. As we discussed in our cost analysis,
to produce more diesel fuel in response to a greater demand for diesel fuel relative to gasoline,
refiners are undercutting the swingcut portion of FCC naphtha at their refineries.E This action to
shift what historically was blended into the gasoline pool to the diesel fuel pool, also
dramatically reduces the sulfur content of the gasoline pool. If the entire swingcut portion of
FCC naphtha is undercut to the diesel fuel pool, the amount of sulfur in the gasoline pool is
reduced by about 50 percent. Our cost analysis estimates that at almost one quarter of U.S.
refineries, refiners are fully undercutting the FCC naphtha to diesel fuel today. At many other
refineries, our cost analysis estimates that refiners are partially undercutting their FCC naphtha.
These refineries will be able to reduce the sulfur of their gasoline well below their current levels
and generate a large number of early credits for Tier 3. Even for the subset of refineries where
FCC naphtha is not being undercut, refiners can assess how much activity or catalyst life is left
in its FCC postreater catalyst and compare this time with the time to the next turnaround when
the FCC postreater catalyst is scheduled to be replaced. If there is spare catalyst life, the refiner
could elect to increase the severity of their postreaters to reduce their gasoline sulfur levels to
under 30 ppm. With this strategy, the refiner would generate early sulfur credits. Also, when the
refiner replaces the catalyst in its Tier 2 postreater, it can elect to do so with a more active
catalyst which would allow the refinery to produce gasoline at sulfur levels below 30 ppm and
generate more early credits for Tier 3.
Based on the early actions refiners are either already taking, or could take, to reduce their
gasoline sulfur levels, we believe that refiners would be able to reduce their gasoline sulfur to as
low as 20 ppm, on average, without making any capital investments. By averaging 20 ppm for
2.5 years prior to 2017, refiners would be able to delay completion of all capital investments for
Tier 3 until mid 2019. If we add the 3.3 ppm of credits during 2012, 2013 and first part of 2014,
refiners would be able to delay completion of all capital investments in Tier 3 until 2020. Thus,
the early credit provisions in-effect can provide nearly 6 years of leadtime for full compliance
with the fuels program. This will allow ample time for refiners to complete their investment and
schedule their tie-ins during normal shutdown activities. It effectively provides even more lead
time than 5 years that the refining industry requested in their comments. The delay in the
program implementation will also help to distribute the demand on the E & C industry over more
years ensuring that the E & C industry would not be overwhelmed. Thus, the Tier 3 program
with a very flexible ABT program provides ample leadtime.
E The term swingcut means that this portion of the FCC product pool can be blended into gasoline or diesel fuel
while still meeting the fuel quality specifications for either fuel regardless of where this swingcut is blended.
4-39
-------
A third aspect of the Tier 3 ABT provisions which helps with leadtime is that small
refiners and small volume refineries (i.e., refineries processing less than or equal to 75,000 net
barrels per day of crude oil) are exempted from complying with the 10 ppm average sulfur
standard until 2020. This provides an estimated 36 refineries, of the total 108 refineries, nearly
6 years of lead time; again more than the 5 years that the refining industry requested in their
comments. As a group, we believe that these refiners and refineries are disproportionally
impacted when it comes to their cost of compliance and ability to rationalize investment costs in
today's gasoline market. Giving these refiners and refineries additional lead time provides more
time to invest in desulfurization technology, take advantage of advancements in technology,
develop confidence in a Tier 3 credit market as a means of compliance, and avoid competition
for capital, engineering, and construction resources with the larger refineries. The small refiner
and small volume refinery exemption until 2020 reduces the number of refineries which will
need to make a significant capital investment to comply with Tier 3 prior to 2020 to a total of 49
non-small refineries (48 revamps and 1 grassroots unit), thus 15 refineries could wait to take
action until 2020 (see Table 4-8 below). The provisions for small refiners and small volume
refineries are discussed in more detail in Section V.E.I of the preamble. Although the small
refiners and small volume refineries are not required to comply with Tier 3 until 2020, they can
still generate early credits (from January 1, 2017 through December 31, 2019) relative to 30 ppm
for sale to other small refiners/small volume refineries, and relative to 10 ppm for sale to non-
small refiners. Such credits generated relative to 10 ppm could provide another pool of early
credits for Tier 3.
In summary, the ABT program provides ample flexibility for complying with Tier 3. The
averaging provisions will allow refiners that only need to revamp their Tier 2 postreaters to
overcomply and generate credits which will allow refineries that otherwise need to install
grassroots units to comply solely through the purchasing of credits. The banking provisions,
which allow refiners to generate early credits, effectively delays investments for compliance to
potentially as late as the year 2020. Finally, the small refiner and small refinery provisions delay
compliance for approximately 30 refiners until 2020. The provisions also allow them to generate
and sell credits during this period if they so choose. All these ABT provisions effectively
address the leadtime concerns. Furthermore, were we to shift the start date back another 2 years
as the refinery industry suggests in their comments, it would provide nearly 8 years of leadtime
for refinery changes that require just 2 or 3 years to complete. Refiners would not have to even
begin taking action for Tier 3 for a couple of years. Given that the lead time and associated
programmatic flexibility we are finalizing is sufficient to allow industry to readily comply; we do
not expect that a delay in the start date of the fuel standards would change the cost of compliance
discussed in Chapter 5. Any further delay in the program start date would simply delay the
actions to comply. Furthermore, delaying the start of the program would forego significant
emissions, air quality, and health benefits.
4.5.2.3 Impact of Turnaround Timing
In their comments to the proposed Tier 3 rulemaking, the oil industry stated that the time
it takes to comply with the proposed 10 ppm gasoline sulfur standard must include the time it
takes to tie-in the revamps and grassroots units. The oil industry suggested that the leadtime for
Tier 3 be increased to 5 years to allow for refiners to make their investments needed to comply
with Tier 3 and tie-in those new investments. We agree that the need to make tie-ins must be
4-40
-------
considered when assessing the feasibility of leadtime, and even when we factor the time needed
to do so, our analysis shows that refiners can comply with Tier 3 with the leadtime provided.
This is true because the final rule effectively provides nearly 6 years of leadtime to complete
capital projects, as described above, and also because the capital projects do not have to be
completed prior to installing the necessary tie-ins for new Tier 3 units.
When a refiner builds a grassroots unit or some sort of revamp that involves a new
reactor or perhaps an added distillation column, the new vessels and associated equipment must
be "tied-in" to the rest of the refinery. The tie-in usually involves connecting a pipe from the
existing unit to the new unit installed. However, a pipe cannot simply be added while the
refinery is operating. Instead, the refiner will add the necessary pipe for making the tie-in when
the refinery is shutdown for regular maintenance. The revamp or grassroots unit does not have
to be started up at that time. Instead, the connection pipe just needs to be added and blocked off
with a sealing-type valve and a blind flange (essentially a flat piece of steel) is bolted on as a
precaution against a leaky valve. This is a very simple process that would take several pipefitters
a half a day of work to complete including completing all the necessary safety protocols.F Once
this piping has been added, the refiner can restart its refinery. Then when the refiner is ready to
complete the tie-in to the completed revamp or grassroots unit, the refiner would remove the
blind flange and connect a pipe that connects the existing part of the refinery to the newly
installed grassroots postreater unit or revamp postreater subunit. This last step can either occur
when the refinery is shutdown or still operating. At that point the refiner would only need to
open the block valve to complete the tie-in of the grassroots unit or revamp to the existing
refinery. One refiner who owns a number of refineries informed us that it installed the tie-ins for
a possible Tier 3 rule when it installed its Tier 2 units.
On its webpage, the American Petroleum Institute (API) reports that the average time
between major turnarounds is 4 years when the U.S. refineries perform maintenance on the FCC
unit.41 An Energy Information Administration (EIA) study makes a similar finding, which is that
refiners target 3-5 years, or 4 years on average, between refinery turnarounds.42 This means
that on average, 25% of U.S. refineries shutdown to perform maintenance on its FCC units each
year. Most often, refiners conduct maintenance turnarounds on their refineries during the spring
when the demand for gasoline and diesel fuel is at their lowest. The EIA study also found that
over 25% of the time refiners need to conduct turnarounds earlier than targeted because a
maintenance issue forces the earlier turnaround. However, the EIA study did not estimate how
much earlier the turnaround occurred so it was not possible to estimate an actual average
turnaround schedule, which is likely less than 4 years.
F Since most refiners have already completed their scoping studies with the vendor companies which license the
desulfurization technologies, they likely already understand what steps would need to be taken to tie-in their
revamps and grassroots units with their existing refinery. Installation of tie-ins is relatively simple refinery change
that can be engineered and installed in a short period of time. For this reason, we believe that refiners can begin
making their tie-ins as soon as the spring of 2014,
G This analysis is focused on the projected emissions increase (step 1) and the net emissions increase calculation
(step 2). Even if a project results in a significant emissions increase, the source may be able to demonstrate that the
net emissions increase is insignificant and thus the project does not constitute a major modification.
4-41
-------
If refiners reduce their gasoline sulfur to 20 ppm immediately after the Tier 3 rule is
finalized, as we anticipate they are capable of, then the substantial number of early credits
generated would allow refiners to delay their unit start-ups sufficiently for refiners to not only
make the tie-ins, but even complete their capital projects as necessary. However, we also
conducted another analysis which assesses the ability of refiners to comply with Tier 3 if they
generate no or few early credits. Using the estimate that turnarounds are scheduled at refineries
every 4 years and the output from our refinery-by-refinery model, we analyzed how the need for
refiners to tie in the new revamps and grassroots units affects the ability for refiners to comply
with the Tier 3 program start date. Our refinery-by-refinery cost model projects how refiners
will comply with the 10 ppm average gasoline sulfur standard at their refineries while
minimizing the costs for doing so under the ABT program. Our modeling projects that the ABT
program will result in a preference of revamps over grassroots units, and refiners desulfurizing
their gasoline to 5 ppm at some refineries and generating credits that would allow refiners to not
reduce the sulfur at other refineries requiring grassroots units. Also, small refiner and refinery
provisions delay the Tier 3 start date for the refineries covered by these provisions until the year
2020. Table 4-6 summarizes our estimate for the numbers of refineries which will need to
comply in 2017 and also in 2020, when the Tier 3 gasoline sulfur program is fully phased-in.
Table 4-6 Summary of Investment Types in 2017 and 2020 Assuming no Early Credit Use
Refineries
Complying in
2017
Small
Refiners/Refineries
Complying in
2020
Total
10 ppm
5 ppm
10 ppm
5 ppm
10 ppm
5 ppm
Grassroots
Units
0
1
0
0
0
1
Revamps
31
21
8
7
39
28
No
Investments
19
21
40
Table 4-6 shows that to comply in 2017 without the use of early credits, 1 grassroots unit
will be installed and 54 existing postreaters will be revamped to comply with the 10 ppm
gasoline sulfur standard. Table 4-6 shows that some refineries would overcomply with the 10
ppm gasoline sulfur standard by producing gasoline which contains 5 ppm sulfur and generate
credits which would be purchased by refiners to show compliance because it would be more
expensive at their refineries to reduce their gasoline sulfur to 10 ppm. Thus, 19 refineries will
4-42
-------
not need to reduce their gasoline sulfur to comply in 2017 because they purchase credits instead.
Thirty six small refineries and refineries owned by small refiners will not need to invest for 2017
because they don't need to comply until 2020.
With an estimate of the number of revamps and grassroots units needed for 2017, we can
superimpose the tie-in timing that we believe would occur as refiners comply. Since the Tier
rule was finalized by early 2014, we believe that refiners will begin to make their Tier 3 tie-ins
starting in the spring of 2014. Thus, refiners will have the years of 2014, 2015 and 2016 to make
their tie-ins during a regularly scheduled refinery turnaround, thus roughly three-quarters of the
tie-ins could occur before the January 1, 2017 compliance date. The other quarter of revamped
refineries would need to wait until the spring of 2017 to make their tie-ins. For the one
grassroots unit, the tie in could be made anytime during the years from 2014 to 2017, but
because we project that a grassroots unit requires three years until unit start-up, we project that
the grassroots unit would start-up in the spring of 2017. Table 4-7 summarizes the unit startup
timeline that accounts for the need for refiners to tie-in their revamped or grassroots units. Since
we do not have data for when individual refineries are shutting down their FCC units and
existing postreaters, we conducted a simple averaged analysis which assumes that each of the
refineries is producing the same volume of gasoline and that each refinery earns credits for a year
of compliance. To avoid confusion with how credits are generated under Tier 2, we call these
compliance units (one compliance unit is equal to one refinery earning 20 ppm (30 minus 10
ppm) credits or purchasing credits for one year).
Table 4-7 Example Case of Compliance Unit Balance for Tier 3
# Revamps
# Grassroots
Compliance
Unit
Balance51
Revamps
installed 2
years after
T3
Published
(April 2016)
39
26
Program
Start Jan.
1,2017
Tie-ins made
so that
Revamps can
start-up - 4
Months after
Program Start
(April 2017)
13
-4.3
Grassroots
units start up
16 Months
after
Program
Start
(April 2017)
1
-0.33
Total
Compliance
Units'
21.4
a Compliance units are a simplified representation for credits, whereas one refinery's compliance with Tier
3 (producing 5 or 10 ppm gasoline instead of 30 ppm gasoline) represents one compliance unit for each year of
compliance.
Table 4-7 shows that 39 of the total 52 revamps will have completed their tie-ins during
the years 2014, 2015 and 2016 and will have the ability to start up immediately upon completion
of construction of their revamped Tier 3 units. The startup of these revamps is estimated to be at
the end of April 2016, about two years after Tier 3 is published. We believe that this is a
4-43
-------
conservative assumption because many revamps will not require the full two years to design,
construct and start up. For example, we don't discount the time required to complete these
projects due to the fact that most refiners already completed their scoping studies before Tier 3
was finalized. Even with this conservative set of assumptions, these refineries are estimated to
generate 26 compliance units - 26 compliance units are generated because those 39 refineries are
operating for two thirds of a year (May through December) before the compliance date of
January 1, 2017. For the other 13 revamps and the one grassroots unit, they are assumed to start
up at the end of April of 2017, which is one third of a year after the official program start date
which would require the purchase of about 5 compliance units. The 26 compliance units
generated greatly exceed the 5 compliance units needed to be purchased by the revamped and
grassroots refineries estimated to start-up after January 1, 2017. It is important to point out that
we did not assume any generation of early compliance units (early credits) in this analysis, thus
no compliance units were created from 2012 through 2016, except for the revamps which started
up in early 2016.
While many refiners likely could begin to make their tie-ins in 2014, which would likely
be before they have completed their finished design, many other refiners could also wait until
2015 before they start to do so. We also assessed the leadtime feasibility if we assume that tie-
ins don't begin until the spring of 2015, thus tie-ins are made in the spring of 2015, 2016, 2017
and 2018. Table 4-10 shows the number of refineries which would start up in the spring of each
year based both on the time to construct the revamps and the grassroots unit and taking into
account when the tie-ins are estimated to occur. Case 1 of Table 4-8 assumes that no early
compliance units (early credits) are generated by refiners. Case 2 shows the compliance unit
balance for Tier 3 if US refineries (including all the refineries not investing for 2017) average
just 1 ppm lower sulfur starting in 2012, generating early compliance units over those 5 years.
Case 3 shows the compliance unit balance for Tier 3 assuming that US refiners will average 3
ppm lower sulfur starting in 2012 (data provided by refiners to EPA shows that the US refining
industry averaged 26.7 ppm sulfur in 2012, which is greater than 3 ppm lower than what the 30
ppm Tier 2 sulfur program requires). This analysis in Table 4-8 is really a business as usual
case. Case 4 that we conducted assumed that all refineries average 27 ppm from 2012 until mid-
2014, and 20 ppm from mid-2014 until they start up, therefore generating 3 and 10 ppm worth of
compliance units, respectively.
4-44
-------
Table 4-8 Example Cases of Compliance Unit Balances for Tier 3
# Revamps
# Grassroots
Case 1
Compliance
Unita Balance
(no early
credits)
Case 2
Compliance
Unita Balance
(1 ppm early
credit)
Case 3
Compliance
Unit3 Balance
(3 ppm early
credits
Case 4
Compliance
Unita Balance
(3/10 ppm
early credits
Early
Compliance
Unitsa
2012-2016
27
81
175.5
Revamps
installed 2
years after
T3 Published
(April 2016)
26
17.3
16.5
14.7
8.7
Tie-ins made
so that
Revamps can
start-up
(April 2017)b
13
-4.3
-4.1
-3.7
-2.2
Grassroots
and Revamp
units start up
(April 20 18)b
13
1
-20
-17.7
-15.9
-9.3
Total
Compliance
Unitsa
-7.0
21.7
76.1
172.7
a Compliance units are a simplified representation for credits, whereas one refinery's compliance with Tier
3 (producing 5 or 10 ppm gasoline instead of 30 ppm gasoline) represents one compliance unit for each year of
compliance.
b The reason why the values in these columns are decreasing as early credit increase is due to the fact that
the sulfur level declines below 30 ppm reducing the number of compliance units needing to be purchased
As shown in the upper row in Table 4-8, we conservatively assume no early compliance
units are generated in Case 1 other than the start-up of the 26 revamps in early 2016. This case
shows that the 17.3 early compliance units generated by the revamps which start up in April of
2016 are insufficient to supply the 24.3 needed compliance units. To show how easy it would be
to satisfy the demand for purchasing compliance units, Case 2 shows the compliance unit
balance if we assume that all US refineries generate early compliance units by reducing their
4-45
-------
gasoline sulfur by only 1 ppm starting in 2012 (the gasoline they are producing is 29 ppm instead
of 30 ppm). For this Case, the very modest additional compliance unit generation that occurs
from 1 ppm of early credit generation (added to the revamps which start up in the spring of 2016)
easily shifts the compliance units balance to a positive balance because the 43.5 early
compliance units generated exceeds the 21.8 compliance units demanded. Case 3 assumes that
refiners produce gasoline which averages 27 ppm sulfur, slightly higher than the level already
being achieved in 2012. For this partial compliance case the 95.6 compliance units generated
exceeds the 19.6 compliance units demanded by a large margin. Case 4, which assumes that
refineries average 27 ppm until mid-2014 and 20 ppm afterwards, shows that the 184 compliance
units generated exceed the 11.5 compliance units demanded by about a factor of seventeen.
These analyses demonstrate that if the US refining industry generates a very modest number of
early credits, that there is ample leadtime for refiners to comply with Tier 3.
4.5.2.4 Hardship
While we project that there will be plenty of credits available to allow refiners to
complete their capital investments and tie in the new or revamped unit to the rest of their
refinery, we provide another option to refiners as a safety valve in case there is a shortfall of
credits. In the case where a refiner cannot complete its project by January 1, 2017 and credits are
not available or are prohibitively expensive, the refiner may file a hardship waiver. Details about
our hardship provisions can be found in Section 5.E.2 of the preamble.
4.5.3 Permitting Analysis
To meet the Tier 3 standards, it will be necessary for some U.S. refineries to install new
equipment and/or modify existing equipment and processes resulting in increased emissions of
some regulated air pollutants. Refinery projects designed to meet the new fuel standards could
trigger preconstruction air permitting requirements under the Clean Air Act and the EPA's New
Source Review (NSR) regulations.
EPA has updated our refinery-by-refinery assessment of the physical and operational
changes that are likely to be needed to allow each active refinery in the U.S. to produce gasoline
that complies with the final Tier 3 fuel specifications. We have also assessed the likely effects of
those changes on refinery emissions. This updated assessment is described in more detail below.
Using this updated assessment, we were able to update our understanding of the potential scope
of the major NSR permitting requirements refiners might face under the final Tier 3 program. In
general, our assessment indicates that only a small number of refineries will likely need to make
modifications of a type and size that would trigger the need for a PSD or nonattainment NSR
permit.
In our updated analysis, we adjusted several inputs to reflect the existence of a
nationwide average, banking, and trading (ABT) program and refined our estimates regarding the
physical and operational changes that will be required at each refinery. The modifications at a
given refinery could include revamps to existing FCC pre- or post-treatment unit(s) or the
installation of a new grassroots post-treatment unit for sulfur reduction. Based on the updated
projections of refinery-specific changes, we re-estimated the increased demand for energy (i.e.,
fuel to generate process heat, steam, and electricity), hydrogen, and sulfur recovery associated
4-46
-------
with meeting the final Tier 3 standards. Having received no comments suggesting that they
should be changed, we re-applied the representative industry emission factors for NAAQS
pollutants, their precursors, and GHGs for each emitting process and combined them with
estimates of incremental activity to estimate the emissions changes at each equipment unit (or
group of similar units) at each refinery.
Resulting emissions increases were compared to pollutant-specific "major modification"
permitting thresholds under the NSR regulations, including Prevention of Significant
Deterioration (PSD) and Nonattainment New Source Review (NA NSR), as applicable.
Refineries projected to trigger PSD and/or NA NSR for specific regulated NSR pollutants were
identified.
In summary, the permitting analysis results support the conclusion that NSR permitting
impacts associated with the final Tier 3 fuel standards are limited, with only 3 refineries
projected to trigger PSD and/or NA NSR for criteria pollutants and an additional 6 refineries
projected to trigger PSD for GHGs only under the worst case impact scenario.
In comparison, for the Tier 2 program, EPA expected the need for NAAQS-related NSR
permits might be widespread among refineries. For the final Tier 3 gasoline sulfur standard,
however, only about 3 refineries would need air permits that address NAAQS pollutants.
This number could be lower if these refineries apply emission controls, such as selective
catalytic reduction (SCR) for NOx, to reduce the emission increases below the significance level.
For refineries that do need a major source NSR permit for NAAQS pollutants, the permitting
process is expected to take 9-12 months. For an in depth assessment of stationary source
implications, refer to Section V.K of the preamble.
4.5.3.1 Calculation Approach
A "major modification" is a physical change or change in the method of operation that
results in a significant emissions increase and a significant net emissions increase.*3 In
accordance with the major NSR applicability procedures contained in the federal regulations, the
calculation approach for determining whether a significant emissions increase will occur depends
upon the type of emissions units involved in the project. Three different tests can potentially
apply: (1) the actual-to-projected-actual test for projects that involve only existing emissions
units, (2) the actual-to-potential test for projects that only involve construction of new emissions
unit(s) and (3) the hybrid test for projects that involve multiple types of emissions units. Under
the hybrid test, the appropriate calculation is performed by emissions unit depending on new vs.
existing unit status. The terms "existing emissions unit" and "new emissions unit" are defined in
G This analysis is focused on the projected emissions increase (step 1) and the net emissions increase calculation
(step 2). Even if a project results in a significant emissions increase, the source may be able to demonstrate that the
net emissions increase is insignificant and thus the project does not constitute a major modification.
4-47
-------
the regulations. "Replacement units" meeting specified criteria are treated as existing emissions
units for the purpose of calculating emissions increases.H
The actual-to-projected-actual test involves calculating baseline actual emissions and
projected actual emissions for each emissions unit affected by a project and calculating the
project emissions increase as the sum of the differences between projected actual emissions and
baseline actual emissions for each unit. The definition of projected actual emissions provides that
the owner or operator shall exclude, in calculating the increase from a particular emissions unit,
that portion of the unit's emissions following the project that the unit could have accommodated
in the selected baseline actual emissions period and that are also unrelated to the project,
including any increased utilization due to demand growth. The actual-to-potential test compares
the potential to emit of a unit with the baseline actual emissions and does not provide for the
exclusion of any emissions in calculating the increase. For new emissions units other than those
qualifying as replacement units, baseline actual emissions are zero.
Because individual refinery project details were not available for this analysis, EPA used
a simplified approach to approximate the results from the application of the actual-to-projected-
actual and/or or actual-to-potential tests for affected units at each source. The approach relied
upon estimates of incremental energy, hydrogen and sulfur recovery demands associated with
meeting the final Tier 3 fuel standards and the application of scaling factors to adjust from base
year to projected maximum annual rates and design rates for existing and new units, respectively.
Representative industry emission factors were identified for each affected emissions unit
category (i.e., process heaters/boilers, hydrogen plants and sulfur recovery units) and emissions
increases were calculated as the product of the production or activity data (e.g., MMBtu energy
demand) and corresponding emission factors (e.g., Ib/MMBtu). The activity data and emission
factors used in the analysis are discussed in further detail below.
Activity Data
EPA performed a refinery-by-refinery analysis of process and equipment changes likely
to be implemented to meet the final Tier 3 fuel standards and from this analysis generated
refinery-specific estimates of energy demand, hydrogen demand and increased sulfur recovery
unit (SRU) loading for 108 U.S. refineries. Energy demand includes fuel needed to generate
process heat, steam and electricity. Hydrogen demand is associated with increased hydrotreating
of post-fluid catalytic cracking (FCC) naphtha and light straight run (LSR) streams, which may
involve new or revamped hydrotreater units. Increased SRU loading results from the increased
fuel desulfurization and associated H^S generation.
H The actual-to-projected actual test was adopted in the federal regulations as part of the 2002 NSR Reform
rulemaking. At this time, the Reform Rule provisions are available to sources in a large majority of state/local
jurisdictions either by delegation or adoption into state/local rules (most of which are SIP approved). There may be
isolated cases where, at the time that Tier 3 projects are being evaluated for NSR applicability, Reform Rule
applicability procedures are not available. For the purpose of this analysis, it was assumed that the federal NSR
applicability procedures were available for all affected sources because: 1) this is the case for most refineries
presently, and 2) the analysis is sufficiently conservative such that applying alternative applicability calculations is
not expected to significantly affect the overall conclusions.
4-48
-------
In our updated analysis, we adjusted the analysis performed prior to the NPRM to reflect
a nationwide average, banking, and trading (ABT) program which is being finalized and refined
our estimates regarding the physical and operational changes that will be required at each
refinery. To estimate the Tier 3 emissions impacts, it was first necessary to establish the
production volumes of the units that would be affected by the standards. A "normal operating"
case reflecting calendar year volumes was developed using refinery-specific actual base-year
(2011) production data and production-normalized impact data (e.g., MMBtu per 1000 bbl
gasoline production and MMscf hydrogen per 1000 bbl gasoline production). Normal operating
production rates were then scaled up to reflect maximum annual capacity utilization rates and
further adjusted using "overdesign" factors in certain cases as summarized below. This scaling
was done under the assumption that, for the purpose of determining NSR applicability, each
refinery would project emissions based on the maximum achievable production rate for existing
units and would use design capacity to calculate potential to emit for new units. It was not tied to
any particular assumption about overall gasoline demand.
• For refineries projected to meet Tier 3 sulfur specifications by revamping existing
hydrotreater units, demands were scaled up to 92 percent of the refinery's maximum
design production. If a particular refinery ran its FCC at a utilization level above 92
percent in the base year, it was assumed that that refinery was already operating at its
annual capacity and no adjustment was made.
• For refineries projected to meet Tier 3 standards by installing new FCC post treatment
units (post-treaters), the same scale-up approach documented above for revamps was
applied, and an additional overdesign factor of 15 percent was applied to represent a
maximum stream day capacity.
• For refineries projected to install new LSR treating units, the same approach was applied
as documented above for new FCC post-treaters.
• For SRUs, the additional sulfur loading was calculated based on the incremental sulfur
reduction required to meet the proposed standards and the scaled-up production estimates
documented above for FCC revamps.
From the scaled utility demand data, low and high impact cases were designed for
estimating emissions increases under each scenario as follows.
• Low case: Each refiner buys its additional hydrogen, high octane blendstocks and
electricity externally.
• High case: Each refiner generates the required additional hydrogen, high octane
blendstocks and electricity internally.
The low case represents external sourcing of some of the energy and chemical demand
associated with meeting the proposed standards, meaning that emissions associated with these
activities do not occur from the refinery itself and are not accounted for in determining NSR
applicability. The high case represents internal production of all required energy and chemical
demand, meaning that these impacts do occur at the refinery itself and are accounted for in
determining NSR applicability.
4-49
-------
Emission Factors
The emission factors used to estimate Tier 3 projected emissions increases are
documented in Table 4-9.
Table 4-9 Emission Factors
Unit
Heaters &
Boilers
H2 Plant
SRU
Pollutant
NOX
SO2
CO
voc
PM10
PM25
H2SO4
CO2e
C02
NOX
S02
CO
VOC
PM10
PM25
H2SO4
C02
EF
0.040
0.012
0.040
0.0055
0.0075
0.0075
0.0004
130.6
25.0
0.271
3.792
0.882
0.039
0.020
0.017
0.290
806.9
Units
Ib/MMBtu
Ib/MMBtu
Ib/MMBtu
Ib/MMBtu
Ib/MMBtu
Ib/MMBtu
Ib/MMBtu
Ib/MMBtu
tons/MMscfH2
Ib/ton S
Ib/ton S
Ib/ton S
Ib/ton S
Ib/ton S
Ib/ton S
Ib/ton S
Ib/ton S
Reference
Assumed average for new/modified units equipped with ultra-
low NOX burners; consistent with NSPS subpart Ja for natural
draft heaters (40 CFR 60.102a).
Calculated based on 8 ppmv annual average SO2 limit in NSPS
subpart Ja (40 CFR 60.102a).
Recent B ACT determinations - burner design and good
combustion practices.
AP-42 Chapter 1.4.
AP-42 Chapter 1.4.
AP-42 Chapter 1.4.
Assumes 2% oxidation of SO2 to SO3 (non-SCR).
40 CFR 98 Tables C-l, C-2 (CO2 EF for fuel gas; CH4 & N2O
EFs for petroleum).
Bonaquist, Dante. Analysis of CO2 Emissions, Reductions, and
Capture for Large -Scale Hydrogen Production Plants, White
Paper. Praxair, Oct. 2010. (Available in docket).
Calculated industry average EF based on 2005 NEI.
Calculated industry average EF based on 2005 NEI.
Calculated industry average EF based on 2005 NEI.
Calculated industry average EF based on 2005 NEI.
Calculated industry average EF based on 2005 NEI.
Calculated industry average EF based on 2005 NEI.
Calculated as 5% of SO2 EF corrected for mol. wt.
Technical Support Document for the Petroleum Refining
Sector: Proposed Rule for Mandatory Reporting of Greenhouse
Gases. EPA OAR, Sept. 8, 2008, p. 11. (Available in docket).
4-50
-------
Pollutants Evaluated
The regulated NSR pollutants identified in the table above were determined to have some
probability of triggering major NSR as a result of projects designed to meet the final Tier 3 fuel
standards. Certain other regulated NSR pollutants, including total reduced sulfur (TRS), F^S and
reduced sulfur compounds (RSC) may be emitted at very low rates from potentially affected
refinery units, but were determined to have little or no probability of exceeding significant
emission rate thresholds. Therefore, these pollutants were not included in the quantitative
analysis.
Heaters and Boilers
The emission factors selected for refinery process heaters and boilers represent
combustion emissions from refinery fuel gas (RFG) and/or natural gas. The worst case (higher
emitting fuel) was assumed where fuel-specific factors were available. For NOX, an emission
factor of 0.040 Ib/MMBtu was selected to represent average emissions from new and modified
refinery process heaters and boilers. In most cases, emissions at this level are achievable using
ultra-low NOX burners and without the addition of add-on controls such as selective catalytic
reduction (SCR). With the use of SCR, emissions substantially below this level, on the order of
0.01 Ib/MMBtu or less are achievable. An 862 emission factor of 0.012 Ib/MMBtu was
calculated based on the NSPS subpart Ja annual average SO2 limit of 8 ppmv for fuel gas
combustion units. A CO emission factor of 0.040 Ib/MMBtu was selected based on recent BACT
determinations that represent burner design and good combustion practices.1 VOC and PM
emission factors are based on EPA AP-42 Chapter 1.4: Natural Gas Combustion and are
consistent with permit precedent for refinery heaters/ For sulfuric acid mist, it was
conservatively assumed that 2 percent of 862 emissions oxidize to 863 and then condense to
form H2SO4. A GHG emission factor of 130.6 Ib/MMBtu CO2 equivalent (CO2e) was calculated
as the sum of CC>2, CFLi and N2O emission factors reported in 40 CFR 98 (EPA GHG Reporting
Rule, Subpart C) for fuel gas multiplied by their respective global warming potentials.
Hydrogen Plants
In addition to emissions associated with process heat, hydrogen production by steam
methane reforming (SMR) generates CC>2 emissions as a byproduct. The selected CC>2 emission
factor of 25 Ib/MMscf Fb is based on data from Praxair, a major hydrogen plant engineering and
design firm and hydrogen supplier to the refining industry in the U.S.
Sulfur Recovery Units
Emissions from SRUs can vary significantly depending on design and upstream
variables, making published emission factors inappropriate in general. To account for this
variability, EPA derived industry average emission factors for most pollutants based on total
1 Based on a review of recent BACT determinations documented in the EPA RACT/ BACT/ LAER Clearinghouse
available at: http://cfpub.epa.gov/RBLC/index.cfm7actioFHome.Home
1 Available at: http://www.epa.gov/ttn/chief/ap42/ch01/final/c01s04.pdf
4-51
-------
refinery SRU emissions reported in the 2005 National Emissions Inventory (NEI) database
combined with sulfur plant production data from the U.S. Energy Information Administration
(EIA). For sulfuric acid mist, it was conservatively assumed that 5 percent of 862 emissions
oxidize to 863 and then condense to form IrbSO/t. For GHGs, a CO2 emission factor from the
EPA Technical Support Document for the proposed GHG Reporting Rule for the petroleum
refining sector was used.
Attainment Status
The attainment status of each refinery location for ozone and PM (PM2.5 and PMio) was
determined using nonattainment area designation information from the EPA Green Book website
as of November 1, 2013 supplemented by information in relevant notices for final redesignation
actions that were not yet reflected on the website.K Based on the attainment status and
classification (for ozone nonattainment areas), NA NSR and PSD triggers were resolved for each
refinery. For serious and severe ozone nonattainment areas, or areas required to maintain NSR
program requirements consistent with those designations under the prior 1-hour ozone standard
as a result of the "anti-backsliding" policy, major modification thresholds of 25 tons per were
assumed for the precursors NOX and VOC, versus 40 tons per year in attainment areas and areas
classified as marginal or moderate nonattainment.L
4.5.3.2 NSR Applicability Results
The results of the Tier 3 projected NSR applicability analysis are summarized below for
the lowest and highest NSR impact scenario/case combinations. In total, 108 refineries
determined to be potentially affected by the final standards were included in the analysis.
No ABT Scenario; Low Case
• 4 refineries trigger PSD and/or NA NSR
• 2 of the 4 refineries trigger PSD and/or NA NSR for criteria pollutant(s)
• 2 of the 4 refineries only trigger PSD for GHGs
Primary ABT Scenario; High Case
• 9 refineries trigger PSD and/or NA NSR
• 3 of the 9 refineries trigger PSD and/or NA NSR for criteria pollutant(s)
• 6 of the 9 refineries only trigger PSD for GHGs
The pollutants driving major NSR applicability are GHGs and NOX. None of the other
pollutants evaluated, including 862, CO, VOC, PM2.5 and PMio and H2SO4, exceeded major
K Available at: http://www.epa.gov/oaqps001/greenbk/
L The significant emission rate (SER) threshold in severe ozone nonattainment areas is 25 tpy of NOX and VOC.
While the SER in serious nonattainment areas is 40 tpy of NOX and VOC, increases in both serious and severe
nonattainment areas are also affected by CAA §182(c)(6), which limits aggregated NOX and VOC emissions over a
five-year period to less than 25 tpy to avoid major NSR. For the purpose of this analysis, a SER of 25 tpy for NOX
and VOC was conservatively assumed for both serious and severe ozone nonattainment areas.
4-52
-------
modification thresholds at any refinery. It should also be noted that in several cases, emissions
increases shown to trigger NSR were very close to the respective thresholds, indicating that there
would likely be opportunities to design these projects in such a way to avoid NSR applicability,
at least for some pollutants.
Example Calculations
The following example calculations illustrate the approach used to determine NSR
applicability on a refinery-by-refinery basis. The calculations are based on hypothetical Tier 3
project impacts.
Assumptions
• Refinery located in an area designated as attainment/unclassifiable for all criteria
pollutants (PSD area)
• Incremental demand estimates reflect "high case" assumptions, i.e., hydrogen and
electrical power demand generated internally
Tier 3 Incremental Refinery Demand Estimates:
Total energy demand: 1,500,000 MMBtu/yr
Hydrogen demand: 400 MMscf/yr
SRU production: 150 tons S/yr
Emissions Increase - NOS:
[(1,500,000 MMBtu/yr) x (0.040 Ib/MMBtu) + (150 tons S/yr) x (0.271 lb/ton)] x (1
ton/2,000 Ib) = 30.0 tons NOx/yr
The applicable significant emission rate threshold for NOX is 40 tpy; therefore, the
project does not trigger PSD for NOX.
Emissions Increase -
[(1,500,000 MMBtu/yr) x (0.0075 Ib/MMBtu) + (150 tons S/yr) x (0.017 lb/ton)] x (1
ton/2,000 Ib) = 5.63 tons PM2.5/yr_
The applicable significant emission rate threshold for PM2.5 is 10 tpy; therefore, the
project does not trigger PSD for PM2.5.
Emissions Increase - GHGs (CO2e):
[(1,500,000 MMBtu/yr) x (130.6 Ib/MMBtu) + (150 tons S/yr) x (806.9 lb/ton)] x (1
ton/2,000 Ib) + (400 MMscf H2/yr) x (25 tons/MMscf) = 108,01 1 tons CO2e/yr
4-53
-------
The applicable "subject to regulation " threshold for GHGs is 75,000 tpy CO 26 and the
applicable significant emission rate threshold is any increase on a mass basis; therefore, the
project triggers PSD for GHGs.
4-54
-------
References
1 Meyers, Robert A., Handbook of Petroleum Refining Processes, McGraw Hill, 1997.
2 40 CFR 80 Subpart H.
3 Shorey, Scott W., AM - 99-55, Exploiting the Synergy Between FCC and Feed Pretreating Units to Improve
Refinery Margins and Produce Low-Sulfur Fuels, National Petroleum and Refiners Association's 1999 Annual
Meeting.
4 Conversation with Woody Shiflett, Advanced Refining Technologies, October 2011.
5 Barletta, Tony, Refiners must optimize FCC feed hydrotreating when producing low-sulfur gasoline, Oil and Gas
Journal, October 14, 2002.
6 Conversation with Woody Shiflett, Advanced Refining Technologies, October 2011.
7 Shorey, Scott W., AM - 99-55, Exploiting the Synergy Between FCC and Feed Pretreating Units to Improve
Refinery Margins and Produce Low-Sulfur Fuels, National Petroleum and Refiners Association's 1999 Annual
Meeting.
8 Brunei, Sylvette, On the hydrodesulfurization of FCC gasoline: a review, Applied Catalysis A: General 278
(2005) 143 - 172.
9 Leonard, Laura E., Recombination: A Complicating Issue in FCC Naphtha Desulfurization, Prepared for the
AIChE 2006 Spring National Meeting, April 26, 2006.
Petroleum Refinery Process Economics, Maples, Robert E., PennWell Books, Tulsa, Oklahoma, 1993.
11 Nocca, J.L., et al, Cost-Effective Attainment of New European Gasoline Sulfur Specifications within Existing
Refineries, November 1998.
12 Prime G, A Sweet Little Process for Ultra-Low Sulfur FCC Gasoline without Heavy Octane Penalty, IFF
Industrial Division.
13 Debuisschert, Quentin, Prime G+ Update, 12th European FCC Conference - Grace Davidson Seminar, Seville
Spain, May 2004.
14 Beck, J.S., Advanced Catalyst Technology and Applications for Higher Quality Fuels and Fuels, Prepr. Pap. Am
Chem Soc., Div. Fuel Chem, 2004 49(2), 507.
15 McGihon, Ron, Exxon Mobil, FCC Naphtha Desulfurization - New Developments, Presentation at the 2009
Technology Conference, October 5 &6, Dubai, United Arab Emirates.
16 Ellis, E.S., Meeting the demands of low sulfur gasoline, Petroleum Technology Quarterly Spring 2002.
17 Successful Start-Up of New Scanfining Unit at Statoil's Mongstad Refinery, November 19, 2003.
4-55
-------
1 &
Greeley, J.P., Zaczepinski, S., Selective Cat Naphtha Hydrofining with Minimal Octane Loss, NPRA 1999
Annual Meeting (this document available from docket A-97-10).
19 Halbert, Thomas R., Technology Options for Meeting Low Sulfur Mogas Targets AM-00-11, Presented at the
2000 Annual Meeting of the National Petrochemical and Refiners Association, March 2000.
20 McGihon, Ron, Exxon Mobil, FCC Naphtha Desulfurization - New Developments, Presentation at the 2009
Technology Conference, October 5 &6, Dubai, United Arab Emirates.
21 Refining Processes 2004, Hydrocarbon Processing.
22 Upson, Lawrence L., Low-sulfur specifications cause refiners to look at hydrotreating options, Oil and Gas
Journal, December 8, 1997.
23 Krenzke, David L., Hydrotreating Technology Improvements for Low-Emissions Fuels AM-96-67, Presented at
the 1996 Annual Meeting of the National Petrochemical and Refiners Association, March 1996.
24 UOP SelectFining™ Process - New Technology for FCC Naphtha HDS, 2009.
25 CDTECH, FCC Gasoline Sulfur Reduction, CDTECH, Sulfur 2000, Hart's Fuel and Technology Management,
Summer 1998.
26 Rock, Kerry J., Putman, Hugh, Global Gasoline Reformulation Requires New Technologies, Presented at Hart's
World Fuels Conference, San Francisco, March 1998.
Rock, Kerry L., et al, Improvements in FCC Gasoline Desulfurization via Catalytic Distillation, Presented at the
1998 NPRA Annual Meeting, March 1998.
28 Greenwood, Gil J., Next Generation Sulfur Removal Technology AM-00-12, Presented at the 2000 NPRA Annual
Meeting, March 2000.
29 Meier, Paul F., S Zorb Gasoline Sulfur Removal Technology - Optimized Design AM-04-14, Presented at the
2004 NPRA Annual Meeting, March 2004.
30 Printed Literature by Phillips Petroleum Shared with EPA September 1999.
31 Fatal, Raj, Advanced FCC Feed Pretreatment Technology and Catalysts Improves FCC Profitability AM-02-58,
Presented at the 2002 NPRA Annual Meeting, March 2002.
32 Conversation with Woody Shiflett of Advanced Refining Technologies October 2011.
33 Letter to Margo Oge, EPA, from Mike Ricca, Baker Hughes, July 25, 2011.
34 Letter to Caryn Muellerleine, EPA, from Richard Kelly, Marvel Oil Company, July 13, 2011.
35 The requirements for transmix blenders are contained in 40 CFR 80.84(d).
36 Graphs of transmix gasoline product sulfur levels at Kinder Morgan transmix processing facilities e-mail from
James Holland, Kinder Morgan, August 24, 2011.
4-56
-------
37 Moncrief, Philip and Ralph Ragsdale, "Can the U.S. E&C Industry Meet the EPA's Low Sulfur Timetable,"
NPRA 2000 Annual Meeting, March 26-28. 2000, Paper No. AM-00-57.
38 Judzis Jr., Arvids, Start-up of First CDHDS Unit at Motiva's Port Arthur, Texas Refinery AM-01-11, technical
paper presented at the 2001 National Petrochemical and Refiners Association Annual Meeting, March 2001.
39 Nocca, Jean-Luc, Prime G+: From Pilot to Start-up of Worlds First Commercial 10 ppm FCC Gasoline AM-02-
12, Technical Paper presented at the 2002 National Petrochemical and Refiners Association Annual Meeting, March
2002.
40 Bedell, Richard, Power point presentation to the Energy Summit 2007, LSU Center for Energy Studies, October
24, 2007.
41 Fact Sheet entitled "Refinery Turnarounds," American Petroleum Institute (API) webpage:
http://www.api.org/oil-and-natural-gas-overview/fuels-and-refining/refineries/refmerv-turnarounds: downloaded
June 21, 2013.
42 Energy Information Administration, Refinery Outages: Description and Potential Impact on Petroleum Product
Prices, March 2007.
4-57
-------
Chapter 5 Fuel Program Costs
5.1 Methodology
This chapter provides a summary of the methodology used and the results obtained from
our cost analyses of the final Tier 3 gasoline sulfur control program. We start by summarizing
the refinery models used for our analysis. We then describe our detailed methodology for
estimating the sulfur control costs for our proposed sulfur program followed by the results.
Finally, we discuss and compare the results of several other cost analyses.
5.1.1 Overview
When we began our planning for estimating the cost of additional reductions in gasoline
sulfur, we considered two different options. One option for estimating the costs would be to
utilize a linear programming (LP) model, while the second option would be to develop a
refinery-by-refinery cost model. While the LP refinery models are necessary and appropriate for
many analyses, they also have several important limitations of relevance here. When used to
model the cost of nationwide fuel control programs on the entire refining industry, LP models are
usually used to model groups of refineries in geographic regions called Petroleum
Administration for Defense Districts (PADDs). The LP refinery model averages the costs over
the refineries represented in the PADDs; however, the technology chosen by the refinery model
would normally be the lowest cost technology found by the refinery model. This may represent
an unreasonable choice of technologies for individual refineries because of how refineries are
configured and based on the sulfur control technologies installed for compliance with the Tier 2
gasoline sulfur program. While the choice of technologies can be limited based on an
approximate analysis of what mix of technologies would best suit the group of refineries
modeled in each PADD, this would only provide an approximate estimate of the cost incurred.
Based on the quality of input data to these LP models and the assumptions made for complying
with a regulatory requirement, LP refinery models may overestimate or underestimate the
program costs. For example, an LP refinery model would not be a sensible tool for estimating
the credit averaging and trading between refineries. This could be partially overcome by
iterating between PADD refinery model runs, thus estimating the number of credits traded
between PADDs and estimating the level of sulfur control in each PADD. However, the need to
make multiple runs per PADD for each case, coupled with the need to run multiple control cases
for different sulfur standards, would be very time consuming, costly and still would only result in
approximate estimates of the sulfur levels achieved and the cost incurred.
For this reason, EPA developed a refinery-by-refinery cost model which models the
capability for each refinery to revamp existing or install new sulfur control technologies
available to them to reduce their gasoline sulfur levels. Rather than start from scratch, we started
from a refinery-by-refinery cost model developed by APT (Mathpro) for EPA to estimate the
cost of benzene control under MSAT2. However, instead of using the representations of benzene
control technology contained in the model, we obtained information about gasoline
desulfurization and represented the cost of this desulfurization in the refinery-by-refinery cost
model.
5-1
-------
We believe the refinery-by-refinery cost model best estimates the cost of individual
refineries, especially when considering an averaging, banking and trading (ABT) program and
therefore is the best analysis tool for estimating nationwide costs. However, certain inputs
necessary for estimating costs. Because some refinery-specific information is not publicly
available, it was necessary to find another way to estimate certain inputs necessary for estimating
costs with the refinery-by-refinery model. The inputs and outputs from LP refinery cost
modeling provide this needed information and it was utilized in the refinery-by-refinery cost
model. The information from LP refinery modeling used in the refinery-by-refinery cost model
is described in Section 5.1.2, and in more detail in other Sections below.
Since the refinery-by-refinery cost model contains confidential business information for
each refinery, we could not publish the model or present some of the details of the model here.
Therefore, to ensure its viability, the refinery-by-refinery cost model that we developed for the
proposed rule was subjected to peer reviews by two refinery industry consultants (involving a
total of three peer reviewers). Our review of most of the suggested changes recommended by the
initial set of peer reviewers suggested that there would be little to no change in our
desulfurization cost estimate (some of the changes would increase the estimated costs, while
others would reduce the estimated costs). Also, after modifying the modeling to respond to this
first round of peer review comments, we then went on to make a number of additional model
improvements to support the final rule. This additional work was then put through a second
round of peer review that was used to modify the modeling for this FRM. The peer review
comments and our assessment of the peer review comments are contained in reports submitted to
the docket. l 2 3 4
The refinery-by-refinery cost model focuses on reducing sulfur from the FCC naphtha
because of its high sulfur content. To comply with the 30-ppm Tier 2 sulfur control program,
most refiners installed FCC naphtha hydrotreaters (referred to as FCC postreaters) or FCC feed
hydrotreaters (referred to as FCC pretreaters) to reduce that unit's sulfur contribution to their
gasoline pool. If refiners installed an FCC postreater under Tier 2, we modeled refiners
revamping those units. However, if refiners relied on FCC pretreaters to comply with Tier 2, we
assumed that grassroots FCC postreaters would have to be installed at those refineries to reduce
its gasoline pool down to 10 ppm. However, since adding grassroots FCC postreaters is
relatively expensive for the amount of sulfur reduction obtained, the ABT analysis we conducted
avoided almost all of these types of investments. Refineries with both pre and postreaters today
could achieve further gasoline sulfur reductions less than 10 ppm (we used 5 ppm) at a relatively
low incremental cost and sell the credits to those refiners who are operating refineries which
would otherwise be faced with grassroots postreater investments. In addition to addressing the
sulfur in the FCC naphtha, we believe that at some refineries, refiners may need to reduce the
sulfur in light straight run (LSR) naphtha.
To better understand the desulfurization costs, we evaluated several alternative gasoline
sulfur control scenarios or cases besides the primary final rule cost analysis. For a 10-ppm
average sulfur standard, also we assessed the costs based on each refinery achieving the 10-ppm
standard with no averaging among refineries and another with only intra-company transfers of
sulfur credits. To provide credits for averaging and trading under the 10-ppm average standard,
we evaluated refiners reducing their gasoline sulfur down to 5 ppm. Since we had estimated
costs for each refinery to get to 5 ppm sulfur, we also report out the cost for a 5 ppm average
5-2
-------
gasoline sulfur standard assuming no averaging between refineries (at an average of 5 ppm, the
opportunity to comply through credit transfers would be limited). Although we considered
reducing the 80 ppm cap standard, we do not provide any cost assessment for lowering the 80
ppm cap standard, nor do we estimate sulfur control costs which include a lower cap standard.
These different scenarios are summarized in Table 5-1.
Table 5-1 Sulfur Control Cases Evaluated for the Final Rule
Description
Final Rule Fuels
Program
Limited Credit
Trading (Proposed
rule)
No Credit Trading
Stringent Sulfur
Standard
Sulfur Standard
lOppm
10 ppm
10 ppm
5 ppm
ABT Credit Trading
Nationwide
Intra-company
None
None
Cap Standard
80 ppm
80 ppm
80 ppm
80 ppm
We made a series of improvements to the cost analysis since we completed the proposed
rule cost analysis. Some of the improvements were made to address the comments of the peer
reviewers. Others were conceived and incorporated to further improve the robustness of the
refinery modeling work by incorporating more data which we believe allowed us to better model
each individual refinery. The following describes the most important improvements that we
made to the refinery-by-refinery cost model:
We updated the EIA data for individual refineries in the refinery-by-refinery cost model
from 2009 to 2011. The data that we updated included crude oil sulfur content and API gravity,
refinery-specific throughput volumes for the atmospheric crude tower, the FCC unit, cokers and
hydrocrackers We also incorporated new refinery-specific data into the refinery cost model
which includes purchases and sales of pentanes plus (natural gas liquids orNGLs), and sales into
the petrochemical market.
Refinery blendstock volumes for the reformer, alkylation unit, isomerization unit, and the
naphtha hydrotreater are now based on actual throughput volume data from the Office of Air
Quality Planning and Standards (OAQPS). OAQPS requested, and the Office of Management
and Budget (OMB) approved, the collection of refinery operations data by OAQPS from refiners
which included throughput data for many refinery units. The data collected by OAQPS was for
the year 2010. We obtained that data and entered it into the refinery-by-refinery model and this
new data provides us with a much clearer picture of how these units are being utilized in
individual refineries and provides us with a much more robust modeling tool. Previously we
were using projected PADD-average use estimates by an LP refinery modeling run made by
Mathpro for the MSAT2 cost analysis. A very important outcome of using this actual unit
throughput data is that we removed the uncertainty associated with the volumes of gasoline
blendstocks which make up each refinery's gasoline pool. Since we better understand the
volumes of these gasoline blendstocks, we can much better deduce certain other practaces that
refiners may be using, such as undercutting their heavy naphtha streams into the jet and diesel
pools which will likely have implications for complying with Tier 3.
5-3
-------
We attempted to better match each refinery's gasoline production in the refmery-by-
refinery cost model to actual refinery gasoline production volume as reported to EPA. In trying
to match individual refinery gasoline volumes, we estimated if the practice of undercutting FCC
naphtha and heavy naphtha is being practiced. Since we often had excess gasoline blendstock
material, we also estimated hydrocracker operation (naphtha, intermediate, or diesel modes) as a
means to match gasoline volumes. More often than not, heavy naphtha volumes tend to exceed
reformer throughput volumes, so for those refineries that have excessive gasoline volumes we
assume that the excessive heavy naphtha volume is sold. For refineries with insufficient gasoline
volume, this excess volume is assumed to be blended into gasoline (but not reformed).
We updated the refinery-by-refinery model with 2012 refinery capacity data, and
included some 2013 and later data for announced expansions. Based on the comments of one
peer reviewer, we focused on including the expansions of the refineries recently modified to
process more tar sands from Canada. Because we incorporated the refinery expansions that
involved processing more tar sands, we needed to not only process more volume, but we needed
to adjust the quality of crude oil processed by the refinery to represent the heavier tar sands.
The volume of light and heavy straight run naphtha are based on the API gravity of the
crude oil slate being refined by each refinery. This replaces the previous method of relying on
similar correlation for the average quality of crude oil refined in each PADD.
We developed a means to adjust desulfurization costs to account for the cases when a
refinery's modeled desulfurization situation differed from the typical case for which the vendors
provided us information. For example, for reducing a refinery's gasoline sulfur from 30 ppm to
10 ppm, the refiner would typically need to reduce its FCC naphtha from 75 to 25 ppm.
Depending on the amount of FCC naphtha blended into its gasoline, the amount of sulfur control
that the refiner would need to achieve in its FCC naphtha could be larger or smaller than this.
We linearly adjusted the desulfurization cost to account for the variances from the typical case.
However, where the level of desulfurization required exceeded 96% and 99% for single stage
and double stage units, respectively, we assumed an exponential increase in hydrogen demand
and octane loss. If we did not make this adjustment, we would be underestimating the cost for
those refineries which must achieve a very steep rate of desulfurization.
We incorporated in our refinery-by-refinery cost model refiner actions that they took for
complying with the Mobile Source Air Toxics rulemaking to reduce the content of benzene in
their gasoline. This affected the volume of benzene precursors sent to the reformer or the volume
of benzene extracted from the gasoline pool.
We obtained more information from vendors of gasoline desulfurization equipment and
included this information in the final rule cost analysis. We increased the off site factor for the
vendor costs to 0.35 based on discussions with engineering companies.5
In the refinery-by-refinery cost model, we updated utility (natural gas and electricity) cost
projections to be based on AEO 2013, the most recent projections available at the time that we
were conducting the final rule cost analysis.
5-4
-------
We updated the octane costs used in the refinery-by-refinery model. To do so we
improved, updated and reran the LP refinery model for octane costs, As part of the LP refinery
analysis, we updated our analysis of the changes in gasoline qualities associated with gasoline
sulfur control. The LP refinery model incorporated improvements in how the model blends up
reformate.A
The several errors that the peer reviewers found in the version of the refinery-by-refinery
model that was used for the proposed rule have been corrected. These errors had an insignificant
impact on the costs of the Tier 3 program.
Since we made a series of new modifications to the refinery-by-refinery cost model, we
had the refinery-by-refinery cost model peer reviewed again by three peer reviewers. The
following describes the most important improvements that we made to the refinery-by-refinery
cost model based on the second round of peer reviews:
Two peer reviewers commented that inside battery and outside battery limit costs that we
estimated do not reflect extra costs which could be incurred. Such excess costs could include
adding pilings for foundations at construction sites with poor soil, adding a control room for
operations control, adding a storage tank, and long pipeline runs in the case that the Tier 3 unit
needs to be placed an abnormally long distance from associated refinery units. To account for
such cost add-ons, we applied a 20% contingency factor.
One peer reviewer commented that some FCC feed pretreaters are too small to hydrotreat
the entire FCC feed, yet our cost analysis assumed that that the entire FCC feed was being
hydrotreated. We adjusted the sulfur level of the feed and the output of FCC units in the cases
where the FCC pretreaters were relative small in throughput capacity and could only treat a part
of the FCC feed.
One peer reviewer commented that we only accounted for the sulfur content of vacuum
gas oil and not very high sulfur light coker gas oil which also is fed to FCC units. For those
refineries with cokers, we increased the sulfur content of the feed to the FCC pretreater or FCC
unit to account for refining light coker gas oil in addition to vacuum gas oil.
One peer reviewer commented that the volumetric yield loss from the reformer was based
on too low of severity and that many reformers are operating at a higher severity than that which
would cause a higher yield loss. We developed a methodology for estimating reformer severity
which resulted in higher average reformer severity.
One peer reviewer commented that the regression that we developed for estimating light
straight run volume from crude oil had a poor fit. The peer reviewer provided a regression based
on much more crude assays with a much better fit of the data and we used that regression to
estimate light straight run volume for each refinery.
A The refinery model now estimates reformate distillation properties, including RVP, based on actual feed qualities.
Before the improvements, the refinery model estimated a fixed set of reformate distillation properties based on a
typical set of feed qualities.
5-5
-------
We discuss how we applied these changes to our refinery-by-refinery cost model in the
following sections and a more complete discussion of the peer review comments and our
response to them are contained in our response document to the peer review comments.
5.1.2 LP Refinery Modeling Methodology and Results
Although we used the refinery-by-refinery cost model to estimate gasoline
desulfurization costs, certain input information was needed. Without access to detailed refinery-
specific information, we relied on outputs from our LP refinery modeling for some important
inputs. Perhaps the most important input is the cost for making up the octane loss that occurs
with desulfurization. In addition, certain refinery operations information from the LP refinery
model was used for estimating the volume of gasoline produced in the refinery-by-refinery
model, including the utilization factors of individual refinery units, and the percentage that
straight run naphtha, FCC naphtha and hydrocrackate comprises of the feed volume of their
respective units.
LP refinery models are detailed mathematical representations of refineries. They are
used by individual refining companies to project how best to operate their refineries. They are
also used by government agencies, such as EPA and DOE, as well as by refining industry
associations and individual companies, to estimate the cost and supply impacts of fuel quality
changes. LP refinery models have been used for these purposes for decades and a certain
protocol has been established to conduct these studies.
Refinery modeling output from the Haverly GRTMPS refinery model was used in the
refinery-by-refinery cost analysis. The primary reason for using recent LP refinery modeling
analysis was to estimate the cost of making up the octane loss associated with desulfurization as
well as estimate how gasoline qualities would be affected by the octane recovery to feed into the
emissions inventory impact analysis discussed in Chapter 7. The cost of octane has decreased
due to expected increased use of ethanol under the RFS2 rulemaking, making historical octane
cost data of limited usefulness.
The first step in conducting an LP refinery modeling analysis was the development of a
base case. The base case is a refinery modeling case that calibrates the refinery model based on
actual refinery unit capacity and input and output data. The base year for this study was the year
2004 for the Haverly model. Because much of the information available for establishing the base
case is only available for PADDs of refineries, the LP refinery modeling was conducted on a
PADD-wide basis. Refinery capacity information for 2012 from the Energy Information
Administration (EIA) was aggregated by PADD and entered into the LP refinery model. The
feedstock volumes, including crude oil and gasoline blendstocks, were obtained from the EIA
and entered into each PADD's model. Similarly, product volumes such as gasoline, jet fuel, and
diesel fuel were obtained from EIA and entered into the cost model. The environmental and
ASTM fuel quality constraints in effect in the base year were imposed on the products. This
includes the Reformulated Gasoline program and the 500-ppm highway diesel fuel sulfur
standard, and the first year of the Tier 2 gasoline sulfur standard. This information was input
into the LP refinery cost model for each PADD and each PADD model was run to model the
U.S. refinery industry for the base year. The gasoline quality for each PADD refinery model was
then compared to the actual gasoline quality for conventional and reformulated gasoline which is
5-6
-------
available from the RFG database. Each model was calibrated to closely approximate the
gasoline quality of each PADD.
The second step in modeling is the development of a reference case. The purpose of the
reference case is to model the refining industry operations and cost in a future year, which is the
year that the control program is modeled to be in effect (serving as a point of reference to the
control cases for estimating costs and other impacts). The reference years for the Haverly
refinery modeling were 2018 and 2030 for the control cases that we ran. The reference cases
were created by starting with the base case for each PADD and adjusting each base case to
model the future year, accounting for the changes between the two years.
Two different types of adjustments were made to the base case refinery models to enable
modeling the refining industry for the reference case. First, the refinery model needed to reflect
future product volumes and energy prices which are projected year-by-year by EIA in its Annual
Energy Outlook (AEO). For the 2018 and 2030 Haverly LP refinery modeling cases, we relied
on the early release of the AEO 2013 (the later actual AEO 2013 was not yet available). The
projected U.S. refinery production was entered into the reference case version of the LP refinery
model. The utility and crude oil and other feedstock prices which are projected by EIA for 2018
and 2030 were also entered into the refinery model, as well as the estimated product prices which
were based on the forecasted product prices.
The second adjustment made to model the reference cases was the application of fuel
quality changes. Environmental programs which have been implemented or which will largely
be implemented by the time that the prospective fuels control program would take effect were
modeled in the reference case. These fuel quality changes include limits such as the 30-ppm
average gasoline sulfur standard, 15-ppm caps on highway and nonroad diesel fuel and the
MSAT2 benzene control program, in addition to the environmental programs which were already
being modeled in the base case. This also included the fact that California gasoline was already
averaging 10 ppm sulfur or less as a result of prior changes to their predictive model used for
gasoline certification, well in advance of their 20 ppm cap on gasoline sulfur taking effect. As a
result, our Tier 3 gasoline standards are not proposed to apply in California. Thus, for this
analysis we only assumed further sulfur control on gasoline volumes produced by California
refineries for distribution outside of California. Also, the implementation of EPAct included a de
facto ban on MTBE by rescinding the RFG oxygenate requirement. The RFS2 renewable fuels
volumes were modeled for 2030 based on the projections made by the Energy Information
Administration (EIA) in its 2013 early release of the Annual Energy Outlook (AEO). For the
2030 reference case, we modeled 14.4billion gallons of corn and cellulosic ethanol, and 8.3
billion gallons of renewable diesel and biodiesel. As discussed below, for the 2018 case, which
was used for modeling octane costs, we modeled all gasoline containing 10 volume percent
ethanol and a small volume of E85.
The third step in conducting the LP refinery modeling was to run the control cases. The
control cases are created by applying a specific fuel control standard to each PADD reference
case. To single out a specific cost or other impact, the sole difference between the control case
and the reference case is the parameter change being studied.
5-7
-------
One control case was run to model the octane loss associated with desulfurizing gasoline
using 2018 as the year of analysis. Since we solely wanted to identify the cost of recovering lost
octane for the refinery-by-refinery modeling, this case was run by reducing the octane value of
the FCC naphtha by one octane number, and this was the sole change relative to the reference
case. The control case was run with capital costs evaluated at a 15 percent rate of return on
investment (ROI) after taxes.6 This case was run with E10 and a small volume of E85, and we
substituted 2013 natural gas liquid prices (ethane, propane and butane) which are much lower
compared to the historical price relationship from previous years. The lower natural gas prices
increases the octane cost provided by the reformer since the reformer produces natural gas
liquids as byproducts. The octane cost estimated by the LP cost model is $0.3 I/octane number-
barrel (0.74 c/octane number-gallon) as summarized in Table 5-39 in the appendix and the
related tables numbered 5-41 through 5-49. Because the octane loss associated with a specific
sulfur control technology may be lower or higher than 1 octane number, we scaled the octane
cost based on the relative estimated octane loss on the FCC naphtha (i.e., a 1A octane loss of the
FCC naphtha was estimated to cost $0.155/octane number-barrel.
The oil industry and one of the peer reviewers commented that we should use the
wholesale price difference between premium and regular gasoline grades to develop an octane
cost. Using this method, they estimated that the cost of octane would be $l/octane number-
barrel. We believe that the premium minus regular grade pricing method for estimating octane
cost overstates the cost of making up lost octane in the FCC naphtha pool. Premium grade
gasoline is produced in smaller batches than regular grade and must be handled specially and
separately to avoid compromising its octane content. For example, when shipping premium
gasoline in pipelines, the interface between premium and regular grades of gasoline must be
downgraded into the regular grade gasoline to ensure that premium gasoline's high octane
content is not compromised - this downgrading increases the production cost of premium
gasoline. Furthermore, the pricing between regular and premium grades of gasoline tends to
reflect higher profit margins on premium fuel.
However, it was further analysis of our LP refinery modeling work which provided the
most significant reason why the price differentials between premium and regular grade gasoline
overstates the cost of octane when the octane loss is in fact occurring in the FCC naphtha. While
the octane cost for making up octane loss in the FCC naphtha is $3 I/octane number-barrel from
the LP model, the octane cost determined by the premium-regular grade differential using the LP
model is $0.50/octane number-barrel, which is 60 percent higher (see Table 5-40). We believe
that the LP refinery model is estimating a higher octane cost for the premium-regular grade
differential because of the cost of producing premium gasoline, which is 6 octane numbers
higher than the regular grade. And yet, the LP model is not capturing the additional cost
inflating factors mentioned above such as smaller tankage, special handling and distribution and
profit. Our conclusion from this analysis is that the premium-regular grade price differential is a
poor indicator of the cost of making up the small amount of lost octane in FCC naphtha for
B Normally we conduct the refinery modeling assuming an after-tax 15% ROI and adjust the costs to reflect a
before-tax 7% ROI to report the costs. However, in this case because the new capital investments were so minimal,
we omitted the capital cost amortization adjustment because its effect on costs was judged to be negligible.
5-8
-------
desulfurization and by using it would overstate the cost. Despite our confidence in the octane
cost that we generated, we recognize the need to quantify the impacts on our costs if octane costs
are indeed higher than what we estimated, so in Section 5.2 where we present the Tier 3 fuel
cost, we provide a sensitivity cost estimate assuming that octane costs are $0.50/octane number-
barrel. A table in an appendix at the end of this chapter summarizes the data output from the
refinery modeling from which we calculated the octane cost for use in the refmery-by refinery
cost model. Other output from the LP refinery modeling are also summarized in tables contained
in the appendix.
Another peer reviewer also evaluated octane cost and concluded that while recent octane
costs using premium minus regular grade price differences were in the $ I/octane number barrel
range, modeled future octane costs for 2015 and 2020 were appreciably lower. This peer
reviewer's estimate for future octane costs could be much lower and provided a range of $0.25 -
$0.5/octane number-barrel. The octane cost that we are using is right in the middle of the
projected octane cost projected by this peer reviewer and our octane cost sensitivity is at the high
end of this range.
We also conducted LP refinery modeling to estimate how the addition of ethanol affects
gasoline quality which allowed us to establish a reference case for our air quality analysis. This
refinery modeling is described, and the resultant projected changes in gasoline quality are
summarized, in Chapter 7 in subsection 7.1.3.
5.1.3 Summary of Refmery-by-Refmery Model Methodology
The purpose of the refinery-by-refinery cost model is to project how each refinery would
reduce the sulfur in its gasoline pool to 10 ppm or lower and to estimate the cost for doing so.
To do this we created a U.S. refining industry refinery-by-refinery spreadsheet cost model using
refinery-specific unit operations information. This spreadsheet cost model also allowed us to
model how costs would be affected by the ABT provisions.
The building of the refinery-by-refinery model consisted of two major steps. The first
step was to estimate baseline operating conditions for each refinery. This involves estimating the
volumes and sulfur levels of the gasoline blendstocks that comprise each refinery's gasoline. We
chose to use information from 2010 and 2011 for modeling the baseline operating conditions for
the refineries as it was the latest years for which we had data for refiner operations and yields.
Additionally, EIA projections indicated that gasoline demand is expected to be essentially flat
between 2011 and when the Tier 3 sulfur control program takes effect, alleviating the need to
adjust refinery operating throughputs and yields for future changes in gasoline demand.6 As a
final adjustment to our estimated gasoline volumes and sulfur levels, we calibrated the model to
actual refinery gasoline volume and sulfur levels to ensure our model's accuracy.
A critical step in characterizing refinery operations was to estimate the FCC naphtha
volume, the sulfur level of the FCC naphtha, and the amount of sulfur reduction needed in FCC
naphtha to meet a 10-ppm sulfur standard at each refinery. We also incorporated in our refmery-
by-refmery model the impacts that FCC pretreaters have on FCC naphtha yields and sulfur
levels, as well as the impact of refinery-specific crude oil sulfur levels on FCC naphtha yields.
Similarly, we also used the refinery-by-refinery cost model to estimate the volume levels of light
5-9
-------
straight run naphtha (LSR) and natural gas liquids (NGL) that may require additional
hydrotreating.
The second step involved applying the various sulfur control technologies to each
refinery as necessary to meet the 10-ppm sulfur standard. We expect that the majority of the
sulfur reductions necessary to comply with a 10-ppm gasoline sulfur standard will come from
reducing the sulfur level in their FCC naphtha. We first identified the sulfur control technology
used by refiners at each refinery for complying with Tier 2. We then obtained information from
vendors of FCC naphtha desulfurizing equipment for revamping or adding new capital as the
basis for complying with additional gasoline sulfur control. Furthermore, we used our refmery-
by-refmery model to estimate the extent that refiners will need to add additional LSR/NGL
hydrotreating capacity at their refineries.
This allowed us to generate a cost estimate for the sulfur control technology in each
refinery. The capital costs for installing the sulfur control technologies in each refinery were
evaluated based on a before-tax, 7 percent return on investment (ROI). In the following sections,
we present the various steps that were used in this refinery-by-refinery modeling analysis.
5.1.3.1 Estimating Individual Refinery Gasoline Blendstock Volumes
To effectively model sulfur control costs for each refinery in our refinery-by-refinery cost
model, it was necessary to understand the sulfur levels and volumes of the various blendstocks
which make up each refinery's gasoline. Each refinery blends up its gasoline pool from the
various gasoline blendstocks that are produced from the refinery units installed at each refinery.
However, information on the volumes and sulfur levels of each gasoline blendstock produced by
each refinery is not publicly available, so it was necessary to seek other sources of information.
Fortunately, instead of solely relying on refinery unit capacity data to estimate unit operations,
we were able to obtain individual refinery unit throughput volumes for most of the refinery units.
From the Energy Information Administration (EIA), we obtained actual year 2011 refinery
specific throughput volumes for the crude, FCC, cokers and hydrocracking units. For other
refinery units, including the alkylation unit, reformer isomerization unit and naphtha
hydrotreater, we obtained throughput data from the Office of Air Quality and Standards
(OAQPS). We used this information to estimate the extent that each refinery process unit is
utilized, followed by a unit-specific analysis for estimating how the respective refinery unit
produces material for blending into gasoline. After the unit-by-unit estimates were completed,
we performed a check for each refinery by comparing our estimated gasoline volumes with
reported gasoline volumes for each refinery, using refinery gasoline production volumes that
refiners report to EPA and made further appropriate adjustments to our refinery-by-refinery
model.
5.1.3.1.1 Principal Refinery Unit Throughput Volumes
To estimate the production volumes for each of the refinery's gasoline blendstocks, we
obtained average unit throughput volumes from EIA and from the Office of Air Quality Planning
and Standards (OAQPS), a sister office at the EPA. With this information, the refmery-by-
refmery model effectively models each refinery's gasoline blendstock yields. Our use of this
information significantly improved our model's ability to estimate FCC naphtha, as well as other
5-10
-------
gasoline blendstocks that each refinery makes. The FCC, coker and hydrocracker unit
throughputs versus actual capacity (capacity utilization) that we obtained from EIA for each
domestic refinery are aggregated and presented on a PADD-average basis in Table 5-2. We also
present the actual throughput volume aggregated by PADD in Table 5-3. This data is presented
on a PADD-average basis to protect refinery confidential business information (CBI).
Table 5-2 Process Capacity Utilization"
Total U.S.
PADD lb
PADD 2
PADD 3
PADDs 4/5
excluding California
Crude
Throughput
0.827
0.722
0.840
0.847
0.762
FCC
Throughput
0.797
0.750
0.763
0.821
0.777
Coker
Throughput
0.803
0.474
0.838
0.813
0.739
Hydrocracker
Throughput
0.827
0.434
0.782
0.860
0.774
aActual unit throughput rates as a fraction of maximum unit capacity on a PADD basis
Table 5-3 2011 EIA Refinery Unit Throughputs (1,000 BPSD)
PADD 1
PADD 2
PADD 3
PADDs 4/5 excluding
California
Crude
906
3,427
7,893
1,362
FCC Units
419
998
2,572
282
Coker
<3a
512
1,326
124
Hydrocracker
<3a
239
985
99
" Since there are less than three refiners in this PADD with these units, the data was not reported to protect
CBI information.
In the model, we also adjusted the refinery capacity information to account for refinery
expansions or refinery shutdowns that we were aware of and are scheduled to occur over the next
several years. Refinery expansions include those announced for WRB Refinery in Wood River
Illinois Motiva in Port Arthur, TX, Valero in Port Arthur, TX, Tesoro in Mandan, ND, Holly
Frontier in Woods Cross, UT and BP in Whiting, IN. For these expansions, there is limited
public data on which of the specific process unit capacities would be increased, though each
expansion project has information on the crude unit capacity increase. Since the data was
limited, we increased all of the existing individual process unit capacities by the fractional
increase in crude oil unit capacity at each of the expanding refineries. Refineries that we believe
are permanently shutdown in PADD 1 were removed from our analysis but, consistent with
recent import/export trends, PADD 3 refineries or European gasoline likely would supply any
lost capacity to PADD 1 as a result of this lost production. PADD 1 refineries that were
presumed to be permanently shutdown are; Giant refinery located in Yorktown, Virginia, the
Sunoco refineries in Westville, New Jersey, the Sunoco refinery in Markus Hook (was part of the
Philadelphia complex), and the Hovensa refinery on St. Croix of the Virgin Islands
The next step was to calculate actual unit throughput rates for the other refinery processes
that produce gasoline blendstocks. These units include alkylation, dimerization, polymerization,
isomerization, naphtha reforming. All of the feedstocks for these processes are primarily
5-11
-------
supplied by the crude and FCC Units. We used OAQPS throughput data for Alkylation and
Isomerization units, while Poly/Dimersol units were assumed to run at full capacity.
Additionally, there were seven refiners that did not have alkylation throughput charge rates in the
OAPQS data. Since it is likely that refiners are using the alkylation units at these refineries, in
response to peer reviewer comments, we assumed that the alkylation units were being used at the
same utilization rate as the FCC unit at those refineries. Since the throughput data for alkylation
units is for the production volume, alkylate volume was the same as the throughput volume. The
results of the capacity utilizations of these downstream units are aggregated summarized on a
PADD-average basis in Table 5-4.
Table 5-4 Other Unit Process Capacity Utilization"
PADD 1
PADD2
PADD 3
PADDs 4/5
excluding
California
Reformer
Throughput
0.499
0.580
0.599
0.609
Alkylation
Throughput
0.597
0.727
0.800
0.691
Isomerization
Throughput
0.960
0.651
0.443
0.730
Poly/Dimersol
Throughput
1.000
1.000
1.000
1.000
a Actual unit throughput rates as a fraction of maximum unit capacity on a PADD basis
With these inputs the refinery-by-refinery model now contained estimates of the
feedstock charge rates for all of the gasoline blendstock producing units. However, for some
units, estimating refinery unit capacity and capacity utilization may does not translate directly
into the gasoline blendstock volume produced by a specific refinery unit. This is because some
refinery units may also produce products other than gasoline blendstock. Additionally, some
processes have volume loss of feedstock due to process reactions and conversions that take place
that increase or decrease the density and therefore the volume of products. To take this into
account, a gasoline fraction yield factor has to be applied for some of the processes to convert the
process charge rate into the yield of gasoline blendstocks. The crude, the FCC, the coker and
hydrocracker units all produce some naphtha from its feed, and we needed to estimate the
naphtha volume. We will discuss how we estimated the FCC naphtha here separate from the
other units because the FCC naphtha is usually kept separate from the other naphtha streams
until it is blended into the gasoline pool. We will discuss the contribution of the crude, coker and
hydrocracker units to naphtha when we present how we estimate light naphtha and heavy
naphtha volume below.
5.1.3.1.2
FCC Naphtha Volume
The FCC unit produces significant volumes of naphtha, a gasoline blendstock. The
conversion percentage to naphtha is affected by the severity of the operation of the FCC unit. As
shown in Table 5-5, an initial estimate was made for the portion of FCC feedstock converted to
naphtha. This initial estimate ranged from 53 to 57 percent depending on the PADD and was
estimated based on a year 2018 refinery modeling run output using the Haverly LP refinery
model.
5-12
-------
Table 5-5 FCC Naphtha Fraction of Feed
PADD 1
PADD2
PADD 3
PADDs 4/5 excluding California
FCC Units Average
0.543
0.574
0.572
0.535
While having an initial estimate is useful, the range among individual refineries can be
quite large, but we didn't have access to refinery-specific data for this. However, we were able
to take steps to use the data that we did have access to better estimate each refinery's FCC
naphtha production. There is expected to be differences in FCC naphtha production between
refineries with and without FCC pretreaters. Thus, as the first step we took to better estimate
FCC naphtha production from each refinery's FCC unit, we 1 differentiated between refineries
that have an FCC feedstock pretreater and those that do not. We also quantified the gasoline
blendstock fraction yield for FCC units that have both feed pretreater and postreater units. Per
comments from the peer review of our model, we limited the amount of FCC feedstock that
could be pretreated at a few refineries that did not have enough capacity to pretreat all of their
FCC feedstock. Historically, refiners have installed FCC feed pretreaters for economic reasons,
as pretreaters increase FCC unit conversion to a high value gasoline blendstock while decreasing
the production of low value light cycle oil and residual material. FCC feed pretreaters also have
the benefit of reducing sulfur from the FCC feedstocks, resulting in the production of lower
sulfur FCC naphtha and ultimately lower sulfur gasoline. In developing our refinery-by-refinery
model, we quantified the impact that FCC feed pretreating and postreating has on FCC naphtha
yields and sulfur levels based on our evaluation of information we received from technology
vendors. The results of this analysis are shown in Table 5-6.
Table 5-6 FCC Unit Gasoline Blendstock Fraction Yields
PADD 1
PADD 2
PADD 3
PADD 4/5 excluding
California
Average of All
FCC Units
0.543
0.574
0.572
0.535
FCC Units with
No Pretreater
0.532
0.547
0.551
0.523
FCC Units with a
Pretreater Only
NA
0.650
0.650
0.608
FCC Units with a
Pretreater and
Postreater
0.588
0.605
0.608
0.574
The volume of FCC naphtha is affected in other ways by how refiners operate their
refineries. One way that FCC naphtha volume is affected is if they choose to produce increased
volumes of propylene.7 8 9 10 Propylene is a valuable feedstock chemical used for the production
of polypropylene. All FCC units produce some propylene, however, if refiners opt to use a
particular catalyst additive called ZSM-5, the FCC unit cracks even more heavier hydrocarbons
(primarily C6 and C7 hydrocarbons) to light olefms and more propylene is produced. The extent
that more olefms are produced depends on the amount of ZSM-5 which is added to the FCC unit.
We obtained refinery-specific data from EIA for the volume of propylene sales by each refinery.
We compared the propylene sales data to the volume of propylene expected to be produced by
5-13
-------
FCC units which do not use the ZSM-5 catalyst, and if the propylene sales volume exceeded the
expected propylene production volume, we assumed that ZSM-5 catalyst is being used. The
reason why this is important is that as the propylene production increases, the volume of FCC
naphtha decreases, which lowers the fuel volume needing to be desulfurized. It also reduces the
severity of desulfurization because FCC naphtha comprises a smaller percentage of the gasoline
pool and the FCC naphtha would not need to be desulfurized to as low a level to meet the same
gasoline sulfur target.
The two critical criteria for implementing this propylene-related adjustment to FCC
naphtha volume are the point at which higher propylene production fractions of the FCC naphtha
indicates the use of ZSM-5, and the rate at which FCC naphtha production volume decreases
relative to the rate that propylene production volume increases. For the former criterion, we used
the recommendation made by a peer reviewer, who consulted refiners on the use of ZSM-5. The
peer reviewer recommended that we use 8.5 percent as the point at which the propylene
production indicated the use of ZSM-5 catalyst.11 Thus, when propylene sales by a refinery
exceed 8.5 percent of the FCC naphtha volume, we apply an adjustment to lower the FCC
naphtha volume. For the second criterion which estimates how much to adjust the FCC naphtha
volume as propylene production exceeds 8.5 percent, we used information from the Jacobs data
base which models the use of a ZSM-5 catalyst. The Jacobs database estimates that for every 1
percent increase in propylene volume above 8.5 percent, there is a reduction of 1.4 percent in
FCC naphtha. After implementing this adjustment, the production volume of FCC naphtha was
reduced in 22 refineries from a range of 0 to 10 percent, and the volume weighted average was
2.9 percent.
Another way that refiner decisions impact FCC naphtha volume is if the refiner elects to
undercut its FCC naphtha to the light cycle oil (diesel fuel) pool in response to relatively higher
demand for diesel fuel. Diesel fuel demand is increasing in the US and US refiners are also
exporting more diesel fuel. At the same time gasoline demand is decreasing in the US and US
refiners tend not to export as much gasoline. This trend is expected to continue in the future
based on Energy Information Administration's (EIA's) projections in its Annual Energy
Outlook. Table 5-7 summarizes both the historical and projected demand in petroleum and
renewable fuels.
5-14
-------
Table 5-7 Historical and Projected US Petroleum and Renewable Fuels Demand
(billion gallons)
Total Gasoline
Total Diesel
Total Jet
Ethanol
Biodiesel/Renewable
Diesel
Refinery Gasoline
Refinery Distillate
(Jet + Diesel)
Distillate/Gasoline
Ratio
2009
132.8
53.9
20.2
11
0.3
121.8
73.8
0.606
2012
134.6
57.8
21.1
13
1
121.6
77.9
0.641
2018
130.9
66.6
22.9
16
3.84
114.9
85.7
0.746
2030
112.5
68.7
24.5
22
8.34
90.5
84.9
0.938
Table 5-7 summarizes historical and future total demand for gasoline, diesel and jet fuel
including renewable fuels and that which refiners produce after renewable fuels are subtracted
out (refineries don't produce these fuels). Refinery gasoline production was essentially flat from
2009 to 2012, but after 2012, it is expected to decline. Conversely, refinery distillate production
has been increasing and it is expected to increase further in the future. Thus, the changing
refinery production profile, when expressed as a ratio of distillate production (jet and diesel fuel)
to gasoline production, is expected to increase from 0.61 to 0.94 during the period from 2009 to
2030. As a point of comparison, Europe demands much more diesel fuel than gasoline, and its
ratio of distillate to gasoline demand is currently over 2.0.
If this trend towards lower relative gasoline demand continues, refiners will continue to
change how they refine crude oil into gasoline and diesel fuel. One way that refiners are
expected to respond to this trend is by undercutting the FCC naphtha to the diesel fuel pool. The
refinery unit information that we obtained from OAQPS has allowed us to better calibrate our
model to how refiners are operating their refineries. Despite the improved modeling of US
refineries, we frequently found that refinery gasoline production exceeded their actual gasoline
production volume as reported to EPA. One refiner which operates multiple refineries shared
with us that they use FCC naphtha undercutting as a means to rebalance its gasoline and diesel
fuel production volumes. One vendor of FCC naphtha desulfurization shared with us that
refiners are indeed undercutting FCC naphtha into diesel fuel at many refineries, particularly in
the wintertime. Finally, the peer reviewers shared that undercutting FCC naphtha to diesel fuel
is a commonly used practice by refiners. As a verification that undercutting of FCC naphtha is
occurring, we compared the T90 value of US gasoline produced in 2011 versus 2004. We found
that the volume-weighted average T90 of US gasoline was 323F in 2011 compared to 333F in
2004. Thus, the gasoline pool is getting lighter which supports our premise that undercutting of
FCC naphtha is occurring. Therefore, we used undercutting of FCC naphtha in our refmery-by-
refmery model as a means to help balance the modeled gasoline volume with actual gasoline
volume. The practice of undercutting is important because if the FCC naphtha swingcut is fully
cut into the diesel fuel pool, an estimated 16% of FCC naphtha volume, and more importantly,
half of the FCC naphtha sulfur would be shifted to the diesel fuel pool.12 13 It is important to
note that we do not project that this is a strategy that refiners will adopt for complying with Tier
5-15
-------
3, but instead refiners will pursue undercutting FCC naphtha as a means to improve its day-to-
day refinery margins independent of Tier 3.
We used output from the LP refinery model to guide us in assumptions for undercutting
the heavy gasoline streams in the refinery-by-refinery model. The LP refinery model estimates
that US refineries undercut the heavy naphtha (heavy naphtha is the feedstock for the reformer)
swingcut by 84% in 2009 and projects that 100% will be undercut in 2018. Conversely, the LP
model estimates that refineries undercut 22% of the FCC naphtha swingcut in 2009 and will
undercut 68% of the FCC naphtha swingcut in 2018. This information from our LP model
suggests that the economics favor undercutting the heavy naphtha to jet fuel over undercutting
FCC naphtha to diesel fuel. For this reason, we chose to undercut the heavy naphtha to jet as a
first step to match the model's gasoline production volumes with actual gasoline volumes (to the
extent that reformer throughput volume allows it) before we resorted to undercutting FCC
naphtha. After applying FCC naphtha undercutting to further balance the gasoline volume in the
refinery-by-refinery model, we ended up undercutting 34% of the FCC naphtha swingcut to
diesel fuel. There is general agreement between the LP refinery model and our refmery-by-
refmery model since the undercutting in the refinery-by-refinery model is in the range of 22% to
68% undercutting which represents the undercutting between 2009 and 2018. While the LP
model projects that refiners will be undercutting a larger portion of the FCC naphtha to diesel
fuel in the future, we did not attempt to account for any cost reduction associated with further
undercutting of FCC naphtha to diesel fuel, so we are likely being conservative with this aspect
of our cost analysis. We present a sensitivity case for the potential savings due to full
undercutting of the FCC naphtha in subsection 5.2.
5.1.3.1.3 Poly Gas and Alky late
For the polymerization and alkylation units the throughput volume from the OAQPS data
base was used for these units for each refinery. Since this volume represents the output from
these units, no adjustments to the volumes were necessary
5.1.3.1.4 Heavy Naphtha and Reformate
A series of steps were taken in the refinery-by-refinery model to effectively estimate the
volume of heavy naphtha which, for the most part, serves as the feed to the reformer. The heavy
naphtha volume was estimated from the volume of three components of heavy naphtha,
including heavy straight run naphtha, heavy coker naphtha, and heavy hydrocrackate. The feed
to the reformer was based on OAQPS throughput data, which often did not match the heavy
naphtha volume. With the use of actual unit throughput volume, it removed a lot of uncertainty
about how refiners are actually operating their reformers. Having good estimates of both the
volume of heavy naphtha and feed to the reformer allowed us to better estimate the total volume
of heavy naphtha sent to the reformer and volume of heavy naphtha which is likely blended
straight to gasoline, instead of being sent to the reformer, or undercut to the jet and kerosene
pool. An additional step was added to estimate the fate of benzene precursors based on refiners'
actions they took to comply with the Mobile Source Air Toxics (MSAT2) 0.62 volume percent
benzene standard. The aggregated summary of refinery action on MSAT2 (we used refinery
specific information which is confidential business information) can be found in a report entitled
5-16
-------
Summary and Analysis of the 2011 Gasoline Benzene Pre-Compliance Reports on EPA's
webpage (www.epa.gov/otaq/regs/toxics).
Heavy straight run naphtha comes from the distillation of crude oil in the atmospheric
crude oil tower. The heavy straight run naphtha volume was estimated for each refinery by
applying the results of a regression analysis of API gravity versus heavy straight run naphtha
volume based on the assays of 13 crude oils. Table 5-8 summarizes the regression equations used
for each cut of heavy straight run naphtha, and also provides a regression based on the total
volume of HSR naphtha.
Table 5-8 Values from Regression Analysis Used to Estimate
Heavy Straight Run Naphtha Volume
All HSR
160/220
220/285
285/350
350/400
x- variable
0.0090
0.0031
0.0035
0.0023
0.0014
Intercept
-0.106
-0.0448
-0.0484
-0.0131
0.0027
r2
0.88
0.81
0.83
0.85
0.75
Based on a regression of crude quality (API gravity) versus crude assays, we were able to
estimate the fraction of crude oil that is heavy straight run. The volume of heavy straight run
naphtha ranges from about 7 to 36 percent of the crude oil input depending on crude oil quality
processed by the refinery, and whether or not the refinery may be undercutting the heaviest part
of this stream into the jet or diesel fuel pool. To demonstrate the results of the regression
analysis, the volume of the various portions of heavy straight run naphtha are estimated for two
different API gravities representing two hypothetical crude oils in Table 5-9. The two API
gravities chosen to represent light and heavy crude oils were 20 and 35, respectively.
5-17
-------
Table 5-9 Estimated heavy straight run naphtha volume cuts at two API Gravities
Crude Oil fraction for
specific distillation
cuts (Temperature F)
160/220
220/285
285/350
350/400
API Gravity
20
0.017
0.023
0.033
0.031
35
0.064
0.076
0.068
0.052
The information in Table 5-9 shows that lighter crude oils (API gravity 35) contain a
larger amount of heavy naphtha than heavier crude oils. An important reason why we chose to
conduct our regression analysis for separate heavy naphtha cutpoints was to capture how, as the
API gravity changes, the fraction of each heavy naphtha cutpoint varies. For example, the
350/400 swingcut cutpoint comprises about 30% of the heavy naphtha volume for the 20 API
crude oil, but only 20% of the of the heavy naphtha volume for the 35 API crude oil.
The volume of heavy coker naphtha was estimated based on fractions of coker unit feed
from information in the Jacobs data base that we use with the Haverly refinery cost model. For
the 160/220, 220/350 and 350/400 distillation cuts, the fractions of coker feed are 0.022, 0.097
and 0.033, respectively. While cokers can produce somewhat lower and higher amounts of
naphtha depending on how the coker units are being operated, we believe that these fractions are
about average. Because the feed volumes to the cokers are actual throughput volumes from EIA
data, the volume of coker naphtha that we estimate at each refinery is likely to be reasonably
accurate.
The volume of heavy hydrocrackate was also estimated based on information in the
Jacobs data base as well as from a peer reviewer, depending on feedstock type. We estimated
different production profiles depending on the feed type. Standalone hydrocrackers which
process gas oil are very flexible units capable of producing a range of gasoline and diesel fuel
production volumes. We incorporated the range of gasoline production volumes in our refinery -
by-refinery model to represent this flexibility. One important reason why we included this range
of gasoline production volume is to help balance the gasoline volume produced at each refinery.
Our refinery-by-refinery model estimates lower and higher gasoline volumes than actual gasoline
5-18
-------
production, and differences in hydrocracker naphtha production volumes, based on how
hydrocrackers are operated, is likely one reason why our model is not matching actual refinery
gasoline production volume. We also model the heavy naphtha production volume when
hydrocrackers are used to process residual fuel. When processing residual fuel, the naphtha
production profile is more predictable for residual fuel hydrocrackers and the conversion to
naphtha tends to be much lower. The last type of hydrocracker that we modeled is an FCC feed
hydrotreater which is converted over to a mild hydrocracker. This may increasingly be an
approach used by refiners to increase their distillate to gasoline ratio. In doing so, it would also
aid Tier 3 sulfur control. However, since we had no data with which to estimate which refiners
might convert their FCC feed hydrotreaters to mild hydrocrackers, we did not assume this in our
cost analysis, but only show the potential impact in a sensitivity analysis discussed in section 5.2.
The fraction of feed to the hydrocracker which is produced as heavy naphtha is summarized in
the Table 5-10 for each type of hydrocracker that we modeled.
Table 5-10 Summary of Heavy Naphtha Production by Hydrocracker Type and Operating
Mode
Naphtha cut
(F)
49/200*
200/285
285/400
Hydrocracking of Gas Oil - Operating Modes
Naphtha
0.42
0.21
0.16
Middle
0.23
0.15
0.23
Diesel
0.06
0.08
0.17
Resid
0.03
0.04
0.09
Gas Oil
FCC Feed HT
to Mild
Hydrocracker
0.02
0.03
0.07
* The naphtha outpoint data for hydrocrackers included the light hydrocrackate cut (49/160) along with the light
portion of the heavy naphtha cut.
The heavy naphtha distillation cutpoints were sometimes different depending on the
source of the heavy naphtha. For example, the heaviest naphtha cut for the hydrocracker is
285/400, while the heaviest naphtha cut for the heavy straight run naphtha is 350/400. It was
necessary to create a consistent set of cutpoints before we could combine the various streams
from different units, so we normalized the cutpoints to create four distinct heavy naphtha
volumes. The cutpoints for the heavy naphtha volumes are 160/200, 200/285, 285/340 and
340/400. The cutpoints were either adopted from the Jacobs data base, or specified to satisfy a
certain need. The benzene precursor fraction was established as 160/200, (changed from
160/180 used in the NPRM) based on benzene precursor boiling point information provided by a
peer reviewer. The swingcut was established as 340/400 (changed from 350/400 used in the
NPRM) to capture the ability for jet fuel or kerosene to accommodate the lighter blendstocks and
still meet the flashpoint specifications. In cases when a heavy naphtha cutpoint provided does
5-19
-------
not match the outpoints we are using, we adjusted the volume using a simple reapportionment
based on the temperature range that the outpoint represents. Using the heaviest naphtha cut
produced by the hydrocracker as an example, to convert 285/400 to 285/340 and 340/400, the
volume was apportioned (340-285)/(400-285), or 48% for the lighter cut, and 52% for the
heavier portion.
Refiners must comply with MSAT2 benzene standard which requires refiners to reduce
their gasoline to an average of 0.62 volume percent benzene. How refiners comply with MSAT2
determines the fate of the 160/200 heavy naphtha volume. Depending on what happens with the
volume, it can affect our estimated refinery gasoline volume. In their compliance report to EPA,
refiners indicated how that they would comply with MSAT2 benzene standard listing one of
several possible actions. These actions can be divided into two categories; those who would
route the benzene precursors to the reformer and those who would not. A second point of
interest is whether the benzene precursor stream or the benzene-rich reformate stream is blended
into the refinery's gasoline pool or not, which makes a difference in terms of matching a
refinery's gasoline production volume. Table 5-11 summarizes the refinery options for the
benzene precursors or the benzene-rich reformate stream.
Table 5-11 Summary of Refiner Actions for Complying with MSAT2
Fate of Benzene Precursors
Do not Route to Reformer
Route to Reformer
Keep in Gasoline Pool
1) Send to Isom unit
2) Route around
Reformer
3) Send to Benzene
Saturation unit
1) Reform and Extract
2) Benzene Alkylation
Do not Keep in Gasoline Pool
Send Benzene Precursors to
another Refinery
Send Benzene-Rich Reformate
to another Refinery or
Petrochemical Plant
The options for compliance with MSAT2 are: continue to reform the benzene precursor
cut and extract the benzene, routing the benzene precursors around the reformer to a benzene
saturation unit, route the benzene precursors around the reformer to an isomerization unit, and
finally, exporting the 160/200 substream. In the case of exporting the 160/200 substream, it is
not clear whether the stream would be exported before or after being reformed. So we used a
reformer balance, as described below, to estimate whether this stream is reformed or not before
being exported. Table 5-12 summarizes refiner plans for complying with MSAT2.
5-20
-------
Table 5-12 Refinery MSAT2 Compliance Plans
Number of
Refineries
MSAT2 Compliance Plans
Extraction
15
Route
Around
Reformer to
Gasoline
39
Isomerization
16
Export
16
Benzene
Saturation
10
The reformate volume is based on throughput volume from data we obtained from
OAQPS. The reformer throughput volume is the volume of heavy naphtha which is fed to the
reformer. The OAQPS reformer throughput data shows that on average, US refiners are
operating their reformers at 69% of capacity, with some reformers operating at full capacity
while others are shutdown (this usually is one train of multiple reformer trains at larger
refineries). Consistent with our LP refinery modeling results, refiners are operating their
reformers at a much lower throughput volume than capacity due to the much increased blending
of ethanol which is very high in octane. Furthermore, demand for premium gasoline has
diminished in recent years, which also has reduced the demand from octane-producing units such
as the reformer.
The reformer throughput volume usually did not match the heavy naphtha volume that
we estimated for each refinery, which we expected because throughput volume was so low
relative to reformer capacity. While there was a very good match between the heavy naphtha
volume and the reformer throughput for 22 refineries, for the rest of the refineries there was a
discrepancy. While 13 refineries did not have sufficient heavy naphtha volume to match the
reformer throughput volume, it was far more common for the heavy naphtha volume to exceed
reformer throughput volume, even after accounting for refiners' plans for complying with
MSAT2. The most likely reason why refiners were not feeding all the heavy naphtha to the
reformer is because refiners are undercutting the 340/400 heavy naphtha swingcut to the jet pool.
When the heavy naphtha volume exceeded the reformer throughput volume and our model
estimated a larger gasoline volume than actual, we assumed that the refinery is undercutting the
heavy naphtha swingcut to the jet pool. Based on our criteria for undercutting the heavy
naphtha, our model estimates that about one third of the heavy naphtha swingcut is being
undercut to the jet pool. If after assuming that the refinery is undercutting the heavy naphtha
swingcut and the refinery's heavy naphtha volume still exceeds the reformer throughput volume,
we assumed that the refinery is routing a portion of the heavy naphtha around the reformer and
blending it straight to gasoline. Alternatively, we assume that the heavy naphtha stream which is
not being reformed is being exported from the refinery if the model's gasoline production is
exceeding actual gasoline production.
5-21
-------
The volume of reformate produced by the reformer is not the same as the feed volume to
the reformer because of the cracking and volume shrinkage which occurs due to the reforming
reactions. Furthermore, the change in volume that occurs in the reformer is dependent on the
severity of the reformer. Reformer severity is measured by the research octane number (RON)
of the reformate, which can vary between 90 to 105. As the severity of the reformer increases,
the volume of the product reformate decreases as a fraction of the feed. This relationship also
varies depending on the type of reformer. Table 5-13 summarizes the estimated volume
shrinkage for different reformers operating at different operating severities.
Table 5-13 Reformate Production Fraction of Feed for Different Reformer Severities
Severity
90
95
100
Reformer Type and Fraction of Feed Volume that Reformate Volume
Represents
Semi Regen
0.87
0.845
0.78
Cyclic
0.87
0.845
0.78
Continuous
0.895
0.87
0.805
The volume of reformate is clearly dependent on the severity that a refiner is operating its
reformer. Unfortunately, there is no available data on the severity at which refinery reformers
are operating, so we developed a means to estimate it. Our assumption is that reformer severity
is dependent on the number of octane increasing units that a refinery has and whether the
refinery has an aromatics plant. If a refinery has an aromatics plant (refiners would likely want
to maximize aromatics production if they have an aromatics plant), or if the refinery does not
have an isomerization unit nor an alkylation plant (if the sole octane producing unit in a refinery
is a reformer, then the refiner would likely need to operate it at high severity to provide sufficient
octane), then we assume that a refinery's reformer is operating at 100 severity. If a refinery has
either an alkylation plant or an isomerization unit and no aromatics plant, then we assume the
refinery's reformer is operating at 95 severity. However, if a refinery has both an alkylation and
isomerization unit and no aromatics plant, then we assume that the refinery's reformer is
operating at 90 severity. Based on our assumptions for estimating reformer severity, the
reformers at 14 refineries are operating at 90 severity, the reformers at 66 refineries are operating
at 95 severity, and the reformers at 27 refineries are operating at 100 severity. Once we assigned
a severity to each refinery's reformer and knowing the type of reformer each refinery has (based
on information from the Oil and Gas Journal), we determined the fraction that the product
reformate is of the throughput volume as summarized in Table 5-13.
5-22
-------
5.1.3.1.5 Light Naphtha
The light straight run naphtha (LSR), light coker naphtha, light hydrocrackate and
pentanes plus (natural gas liquids) all are light naphtha streams which are principally comprised
of five- and six-carbon hydrocarbons which boil in the range from 49 F to 160 F. The LSR
comes directly from crude oil and the pentanes plus (natural gas liquids or NGL) come from
natural gas wells, while light coker naphtha and light hydrocrackate come from the coker and
hydrocracker units, respectively.
The volume of LSR for each refinery was based on a regression analysis of volume of
LSR versus API gravity for the crude oil that it processes. Crude oil assays contain crude API
gravity numbers along with the corresponding LSR yields fraction for the crude oils, which
enables someone to develop a correlation between crude API and LSR yield. During the peer
review of our model, one of the peer reviewers provided a regression analysis of the LSR versus
crude oil API gravity using a large number of crude oil assays. For the final rule we adopted the
equation recommended by the peer reviewer. The resulting equation is: LSR Yield = 2E-06 X3
- 8E-5X2 + 0.0022X + 0.0001, with X being crude API gravity number. This equation has a
predictive R value of 0.8197 for estimating the percentage of LSR from Crude API gravity. By
utilizing data obtained from EIA for each refinery's average annual crude oil API number, we
were then able to estimate each refiner's LSR yield fraction. Our regression analysis estimates
that the percentage of LSR in crude oil varies from 2 and one half percent, for heavy crude oils,
to over 8 percent for light crude oils.
The NGL that refiners purchase comes from natural gas wells after it has been separated
and cleaned up at natural gas processing plants. Refiners report to EIA this volume of NGL that
they purchase at each refinery and we obtained and entered this data into our refinery-by-refinery
model and added this volume to the rest of the light naphtha assuming that it all boils in the
temperature range of 49 F to 160 F.
The volume of light naphtha produced by the coker unit is estimated based on data from
the Jacobs data base. That data base estimates that light coker naphtha comprises 2.1 percent of
the feed volume to the coker unit. While the coker units may produce a somewhat higher or
lower percentage of light naphtha than what we estimated, we believe that this percentage is
average for coker units in general.
The volume of light naphtha from hydrocrackers is estimated based on naphtha
production estimates for hydrocrackers operating in different modes as described above in
Section 5.1.3.1.4. In that Section, we present the volume of lightest cut of the heavy naphtha
produced by hydrocrackers, and that cut also contains the volume of light naphtha. To calculate
the volume of light naphtha from that cut, we used a simple ratio of the distillation cut points.
The light naphtha portion of the 49/200 cut is calculated to be (160-49)7(200-49), or 74 percent
of that volume fraction provided for that cut.
For modeling purposes, the light naphtha (LSR, light coker naphtha and light
hydrocrackate) were split into a five carbon stream and a six carbon stream. This split was
necessary to facilitate matching the RVP of the gasoline produced by the refinery, as described in
subsection X that discusses matching refinery gasoline volume. These five and six carbon
5-23
-------
streams have several possible fates in each refinery. If a refinery has an isomerization unit,
which converts straight chain hydrocarbons to branched chain hydrocarbons for its octane
benefit, these streams could be fed to that unit. If a refinery does not have an isomerization unit,
or that unit is capacity limited, the five and six carbon hydrocarbons are blended directly into
gasoline, unless removing five carbon hydrocarbons is necessary to control RVP.
The volume of C5 and C6 hydrocarbons fed to a refinery's isomerization unit was based
on its throughput volume in 2010 from OAQPS data. To ease the task of balancing RVP, the six
carbon stream was sent to the isomerization unit first and then the five carbon stream sent to the
isomerization unit to satisfy the throughput volume. Any leftover volume was blended straight
to gasoline, unless some 5 carbon hydrocarbons needed to be removed to balance RVP. For the
isomerization units in 6 refineries, the Oil and Gas Journal specifies that these are only C5
isomerization units. For these isomerization units, we modeled that these isomerization units
were solely being fed C5s. The volume of isomerate is estimated to be 98.5 percent of the feed
to the isomerization unit (there is 1.5 percent volume loss in the isomerization unit).
5.1.3.1.6 Other Purchased Blendstocks
Some additional gasoline blendstocks are purchased and blended into the gasoline pool
and we attempt to account for this. The gasoline blendstocks typically purchased include ethanol
and butane. Gasoline quality data reported by refiners to us contains the percentage of ethanol
blended into each refinery's RFG, but does not contain the amount of ethanol that is blended into
the conventional gasoline pool at the refinery, and it certainly does not contain the ethanol
blended into gasoline at terminals. In the 2011 gasoline quality database, the ethanol volumes
only averaged 3.3 percent of US refinery gasoline production, which would result in an over
estimation of our refinery and program costs if we did not account for the rest of the ethanol. To
capture the impact of the blended ethanol, we added ethanol into each refinery's gasoline pool
until the total ethanol blended into each refinery's gasoline reached 10 percent. .
To estimate the butane volumes in our refinery-by-refinery model we used an RVP
balance equation. This equation states that the product of the overall RVP and volume of a
refinery's annual gasoline pool is equal to the sum of the product of the RVP and volumes of the
non-butane components plus the product of the RVP and the volume of the butane blendstocks.
We raised each of the RVP values to an exponent to capture the true impact of hydrocarbon's
volatility. The value of the exponent that we used is 1.15 which is the value recommended by
vendors and consultants. This equation can be rearranged to solve for the volume of butane
blendstocks as shown in Equation 5-1 below.
5-24
-------
Equation 5-1 RVP Butane Balance Equation
Butane =D* (AL15-BL15)/(CL15-AL15)
Where:
Butane = Volume of Butane added in each refinery in BPSD
A = Blended gasoline RVP average
B = Non-butane blendstock RVP average
C = Butane RVP
D = Volume of gasoline produced
The gasoline production volumes and RVP of the blended gasoline are reported to EPA
by refiners for each refinery and were used for the A and D terms in Equation 5-1. For the
butane stream, our LP refinery modeling shows that this stream is comprised predominantly of
normal butane. We assumed that only normal butane would be blended into gasoline in our
refinery modeling analysis. The non-butane blendstock RVP was estimated by multiplying each
individual gasoline blendstock RVP times the gasoline blendstocks volume fraction of each
refineries gasoline pool (CG and RFG) using 2011 ethanol volumes and taking the sum of all of
these values. The RVP value for each of these streams is shown in Table 5-14 below. With this
information we were then able to estimate the volume of butane added to the gasoline blendstock
at each refinery. The annual volumes of butane added by refineries on a PADD level are listed
in Table 5-15. The volume of butane blended into gasoline at each individual refinery varies
based on the annual average gasoline RVP that the refinery produces (the RVP of CG and RFG
gasoline are volume weighted together), as well the variance in gasoline blendstock streams that
a particular refinery uses to produce CG and RFG gasoline.
Table 5-14 PADD Average RVP's of Gasoline Blendstocks
LSR
Heavy Naphtha
Reformate[update]
FCC Naphtha
Isomerate C5
Isomerate C6
NGL
Polymerization Gasoline
Alkylate, C4
Dimersol
Ethanof
PADD 1
12.0
3.0
4.5
4.6
13.0
7.2
12.6
2.8
3.2
5.8
13.5
PADD 2
12.0
3.0
6.6
4.6
13.0
7.2
12.6
2.8
3.2
5.8
10.713.5
PADD 3
12.0
3.0
5.0
4.6
13.0
7.2
12.6
2.8
3.2
5.8
10.713.5
PADD's4/5a
12.0
3.0
6.2
4.6
13.0
7.2
12.6
2.8
3.2
5.8
10.713.5
"Excluding California data
3Ethanol's RVP was established by taking ethanol's blending RVP of 20 to the root of 1.15
5-25
-------
Table 5-15 PADD Average Gasoline Data
Non-butane blendstock RVP (B), psi
Gasoline Pool Volume (D), BPSD
Volume Butane Added, BPSD
Blended Gasoline RVP average (A), psi
PADD 1
6.4
738.2
21.4
9.4
PADD 2
6.3
1744
95.1
10.38
PADD 3
5.7
3487.8
143.1
9.31
PADD's4/5a
6.3
500.3
25.7
10.2
"Excluding California data
5.1.3.1.7 Calibrating the Blendstock Volumes in the Refinery-By-
Refinery Model
As we discussed earlier, we took steps to balance each refinery's gasoline production
volume in the refinery-by-refinery model against actual gasoline volume production volume
reported by refiners to EPA. We estimated each refinery's gasoline volume by estimating the
volume of each gasoline blendstock which comprises gasoline. Because we used individual
refinery unit throughput volumes to estimate the gasoline blendstock volume for each refinery
unit, we believe that we employed a robust methodology for estimating gasoline volume.
However, the modeled gasoline volume for most refineries was higher than their actual gasoline
volume. Thus, it was necessary to identify and make additional adjustments that would better
match refiners' gasoline production.
We adjusted gasoline volume based on certain inputs and outputs that we identified. EIA
collects data on natural gas liquid (termed pentanes plus) purchases for each refinery and we
added the 2011 volume to each refinery. We entered the information of refinery actions for
complying with the MSAT2 gasoline benzene standard in the refinery model. The MSAT2
actions which involved exports from the refinery, which included extraction of benzene, or
exports of benzene precursors or the benzene-rich reformate, and these impact gasoline volume.
We also adjusted for gasoline exports from Gulf Coast refineries. We did not have refinery-
specific data on Gulf Coast exports, so we apportioned the 428 thousand barrel per day exports
(EIA data) from Gulf Coast refineries by each Gulf Coast refinery's overproduction of gasoline.
While the throughput volume fairly well defines the gasoline blendstock volumes for
most refinery units, we needed to estimate the gasoline blendstock volume for others.
Hydrocrackers can be operated in different modes that preferentially produce naphtha or
distillate fuel. While we did not know which mode each refinery operates its hydrocrackers,
when the modeled gasoline volume for a refinery with a hydrocracker was too high relative to its
actual gasoline production, we assumed that the refinery was operating its hydrocracker in
distillate maximization mode. Or, when the modeled gasoline volume for a refinery with a
hydrocracker was too low, we assumed that the hydrocracker was being operated in naphtha
mode. We also assumed a middle mode for hydrocrackers when the modeled gasoline volume
for refinery with a hydrocracker was close to matching its actual gasoline volume.
Another gasoline blendstock volume which we adjusted was heavy naphtha from the
atmospheric crude distillation unit, the hydrocracker and the coker that would otherwise be feed
to the reformer. Refiners have the option of feeding the heaviest portion of this heavy naphtha
5-26
-------
stream to the reformer or cutting it into the distillate pool. For any refinery, we only cut this
heaviest portion of the heavy naphtha pool into distillate fuel if the volume of heavy naphtha
exceeded the reformer throughput volume and the modeled gasoline volume exceeded the actual
gasoline volume for that refinery. There were times when the heavy naphtha volume was
insufficient compared to the reformer throughput volume, so we assumed that unfinished oils
which are purchases that are made by refineries are
The last refinery's unit production volume that we adjusted is the FCC unit and we did so
in several different ways. The production of FCC naphtha from the FCC unit is affected by
whether the feed to the FCC unit is hydrotreated with an FCC pretreater - FCC units produce
greater naphtha when the feed is hydrotreated. As a percentage of feed, the FCC naphtha volume
varied from 53.5 to 65%. Another adjustment we made was to adjust for refinery use of a
propylene increasing catalyst named ZSM-5. When maximizing propylene production, the
ZSM-5 reduces FCC naphtha volume. We identified 21 refineries which we estimate are using
the ZSM-5 catalyst. The last volume adjustment we made to FCC naphtha was to undercut the
heaviest portion of the FCC naphtha, called the swingcut. This volume is called the swingcut
because it can either be blended into the gasoline pool or the distillate pool. We only undercut
the FCC naphtha to distillate fuel if the modeled refinery volume exceeded the actual refinery
volume. In the refinery-by-refinery model, some or all of the FCC naphtha swingcut was being
undercut to distillate fuel at 45 out of the total 108 refineries that we modeled.
After having completed these steps that we took to balance the modeled gasoline volume
against actual gasoline production volume, the modeled gasoline volume for many refineries did
match actual gasoline production volume. However, for many refineries the modeled refinery
gasoline volume did not match the actual gasoline production volume. We reviewed the
modeling for the refineries whose modeled gasoline volume did not match actual gasoline
volume and we were able to identify several reasons why there were discrepancies. First, the
refineries which we modeled as tar sands refineries were short of gasoline relative to actual
gasoline volume. This is because tar sands are so heavy that they contain very little naphtha.
Another group of refineries with a volume discrepancy have aromatics extraction units and they
tend to have insufficient heavy naphtha feed for the reformer. We believe that the volume
discrepancy exists because these refineries purchase heavy naphtha for feeding to their reformers
for making additional aromatics for their aromatics plants, but we do not have data for heavy
naphtha purchase (the only category that may reflect this is unfinished oils, but the volumes did
not match the heavy naphtha shortfall in many of these cases. To avoid improperly modeling
refinery costs because of our inability to accurately model refinery gasoline volume, we used
actual refinery gasoline production volume in the year 2011 for refinery gasoline volume instead
of the modeled gasoline volume estimated by the refinery-by-refinery model.
5.1.3.2 Refinery Blendstock Sulfur Levels
After determining the volume of each gasoline blendstock stream, we next estimated the
sulfur level of each of the gasoline blendstocks for our modeling analysis using information we
collected from literature reviews and discussions with refinery consultants and technology
providers. We also considered the blendstock sulfur levels estimated for the MSAT2 rule and
the estimates derived from our refinery-by-refinery model to estimate the sulfur levels of the
blendstock streams. Establishing these sulfur levels is important as this sets a baseline for the
5-27
-------
refmery-by-refmery model that represents our estimate for the current operations of each
refinery. This allowed us to project what changes refiners would have to make in their refineries
to comply with the Tier 3 standards, and project the new investments and operating costs
associated with these changes. The following section contains further details on how the sulfur
content of each of the blendstocks was estimated. The results of this analysis can be found in
Table 5-9 at the end of this section.
The first stream we considered was the butanes that are used as a gasoline blendstock.
The butanes used as gasoline blendstock within a refinery come from a variety of sources. Much
of the butane used as a gasoline blendstock is distilled from the crude oil or other blendstock
streams within the refinery. Refiners remove the butanes from crude oil and sometimes gasoline
blendstocks which contain some butane (i.e., FCC naphtha, hydrocrackate) and then blend them
back into the gasoline up to the RVP or vapor/liquid limit applicable to the gasoline market that
the gasoline is being sold into. During the summer months refiners usually have excess butane
which cannot be blended into the gasoline pool because of the tighter RVP standards. Many
refiners store the excess butanes and then blend them back into gasoline in the winter months
when the volatility limits for gasoline are less stringent. Other sources of butanes used as
gasoline blendstocks are natural gas processers and crude oil drilling operations. The butanes
from these sources are produced in downstream units which separate the various hydrocarbon
components. Some of these downstream units "sweeten" or desulfurize the butanes using an
extractive desulfurization unit prior to shipping them in pipelines or selling them directly to
refiners. The sweetening process reacts the hydrocarbon mercaptan compounds to disulfide
compounds reducing their odor and corrosivity. The sweetening process, however, does not
lower the sulfur level. If the source natural gas well is very high in sulfur, the operator may need
to use an extractive treatment technology which actually removes the sulfur from the butane
stream. If the purchased butanes are not treated, then refiners treat the butanes. This treatment
generally lowers the sulfur level of the butanes to under 5ppm. Purchased butanes that are
blended into gasoline have a sulfur limit of 30 ppm and those that are shipped through pipelines,
regardless of their end use, have a limit of 140 ppm. Furthermore, many refiners have extractive
desMerox units on site that are capable of removing sulfur from butanes that are either purchased
or generated internally from refinery units. For dimersol, and poly gas blendstock streams, we
used the same sulfur levels that we estimated for our MSAT2 rulemaking. The sulfur levels for
these streams are inherently low due to the dynamics of process reactions in the dimersol and
polymerization units. Furthermore, it is unlikely that refiners have altered these processes in
their refineries since our analysis for the MSAT2 rule was completed.
Alkylate blendstocks usually have a small amount of sulfur, usually less than 15 ppm.
The primary source of sulfur in alkylate is the sulfuric acid that is used as a catalyst in the
alkylation process. Finished product coalescers and knockout drums are used by refiners to
remove impurities, including sulfuric compounds, from the alkylate product as it leaves the
alkylation unit. This separation is imperfect, and a small quantity of the sulfuric compounds
which remain in the alkylate account for the majority of its sulfur content. Prior to the enactment
of the Tier 2 standards, the alkylate produced by most refineries contained 10 to 25 ppm sulfur
which assumes that there was some carryover of sulfuric compounds into the alkylate. Based on
our discussions with gasoline desulfurization technology vendors, however, refiners have
installed new acid coalescers and knock out drums in recent years. These new units improved
the removal of residual sulfuric compounds and can produce an alkylate blendstock with a 5-ppm
5-28
-------
sulfur level. This adjustment by refiners seems to be a low cost method for reducing the sulfur
content of alkylate. For our refinery-by-refinery baseline analysis, we assumed that refiners have
already installed improved acid knockout drums and are currently producing alkylate which
averages 10 ppm sulfur. We also assumed that Hydrofluoric Acid (HF) alkylation processes had
the same alkylate yield per feedstock throughput as a sulfuric acid alkylation unit in our refinery
by refinery model. We assumed that the sulfur level of alkylate from an HF units also averages
10 ppm sulfur, even though HF processing units use hydrofluoric acid as the processing catalyst,
instead of using sulfuric acid.
The coker unit produces a gasoline blendstock with a significant amount of sulfur. These
units convert the heavy portion of crude oil, called residuals, into gasoline and diesel blendstocks
through the use of heat and pressure. The gasoline blendstock produced by the coker can contain
more than 3,000 ppm sulfur. This stream is normally split into two different streams. The
stream which contains the six to nine carbon hydrocarbons is processed in the naphtha
hydrotreater, which reduces the sulfur level of this blendstock to below 1 ppm. This stream is
then routed to the reformer for octane improvement. The five and six carbon hydrocarbon
portion of coker naphtha is called light coker naphtha and usually contains on the order of
several hundred ppm sulfur. Because of the instability of this stream due to its high olefm
content, it is generally processed by the naphtha hydrotreater and sent to the isomerization unit if
the refinery has one. After being processed in the hydrotreater, the sulfur content of this stream
is reduced to approximately 1 ppm. If a refinery does not have an isomerization unit, we assume
that light coker naphtha is already being desulfurized due to the unstable and very high sulfur
levels of this naphtha stream. These treating pathways were assumed for each refinery in the
refinery-by-refinery baseline analysis.
Straight run naphtha is a gasoline blendstock which contains a moderate amount of
sulfur. Straight run naphtha is the product stream from the atmospheric crude oil tower with a
boiling point that falls within the boiling range of gasoline. The heaviest portion of straight run
naphtha is higher in sulfur relative to the lighter portion of the straight run naphtha. The heavy
portion of straight run naphtha is normally hydrotreated in the naphtha hydrotreater before being
reformed by the reformer in order to improve its octane before being blended into gasoline.
After this processing, the reformate has a sulfur level of less than 1 ppm. The light straight run
naphtha (LSR) contains the five and part of the six carbon hydrocarbons and can range from zero
to hundreds of parts per million sulfur before any extractive desulfurization or hydrotreating.
LSR that is routed as feedstock to isomerization units has its sulfur lowered to 1 ppm by
hydrotreating in the naphtha hydrotreater. This hydrotreating is necessary to allow this material
to be processed in the isomerization unit, as the catalysts in these units require low sulfur
feedstocks to function properly. Some refiners, however, do not have isomerization units or they
produce LSR volumes that are greater than the capacity of their isomerization units. For many of
these cases, refiners can treat LSR using sulfur extraction which, in most cases, reduces the
sulfur level of LSR to well under 10 ppm sulfur. Even cases where there is insufficient capacity
in the isomerization units it may be desirable for refiners to hydrotreat as much of the LSR as
possible since it is more cost-effective to reduce the sulfur content of the LSR than the FCC
naphtha. Refiners can either hydrotreat this volume of LSR in the naphtha hydrotreaters or in
FCC naphtha postreaters. After considering the volume of LSR which is already being
hydrotreated or treated with extractive desulfurization, we assume that LSR contains 5 ppm
sulfur.
5-29
-------
Natural Gas Liquids (NGL) have a composition that is similar to LSR, as it is comprised
primarily of pentanes and hexanes. NGLs are produced from natural gas processers and crude
oil drilling operations and the sulfur content of the NGLs can vary depending on its source,
although we estimate that this stream ranges from zero to about lOOppm sulfur. While some of
the NGLs are treated to remove sulfur by the NGL producers before being purchased by the
refineries we did not have sufficient information to be able to determine the extent to which this
treatment is occurring. At most, if not all refineries, refiners are extracting the sulfur from NGLs
to prevent their gasoline from having odor issues due to the mercaptans which may be present in
NGLs. Extraction is likely to reduce the total sulfur content of NGLs to low levels. For the
control case in our refinery-by-refinery model we assumed that NGL liquids are being treated in
a similar manner as LSR.
We also assumed that all ethanol blended into gasoline has a sulfur content of 5 ppm.
Ethanol produced at ethanol plants should naturally have a negligible amount of sulfur. Before
being sold, however, a denaturant is added to the ethanol. This denaturant most commonly used
is natural gasoline, a C5 to C7 naphtha produced during natural gas processing. Natural gasoline
has a sulfur content that ranges anywhere from a few parts per million to a couple hundred parts
per million sulfur. We conservatively assumed that the natural gasoline used as an ethanol
denaturant is not hydrotreated and has an average sulfur level of 250 ppm. Ethanol contains 2
percent denaturant, which results in denatured ethanol having a sulfur level of 5 ppm.
After determining the sulfur level for each of the gasoline blendstock streams as
discussed above we can use this information, along with the gasoline production volumes and
sulfur levels for the United States in 2011, to determine the sulfur level of the FCC naphtha
stream on a national average basis. To do this we used the following equation, referred to as
Equation 5-2 hereafter:
5-30
-------
FCC Naphtha Sulfur ppm = [(A*B) - (C*D+E*F+G*H+I*J+K*L+M*N+O*P+Q*R+S*T)] / Z
Where:
A = Refinery Total Gasoline Yield, BPSD
B = Refinery Total Gasoline Sulfur level, ppm
C = Butane to Gasoline, BPSD
D = Butane Sulfur, ppm
E = Alkylate BPSD
F = Alkylate Sulfur, ppm
G= Reformate BPSD
H= Reformate Sulfur, ppm
I = Coker Naphtha, BPSD
J = Coker Naphtha Sulfur, ppm
K= Hydro-crackate BPSD
L= Hydro-crackate Sulfur, ppm
M= Light Straight Run (LSR) and Natural Gas Liquids (NGL), BPSD
N =LSR and NGL Sulfur, ppm
O= Dimersol, BPSD
P= Dimersol Sulfur, ppm
Q= Polymerization BPSD
R= Polymerization Sulfur, ppm
S= Ethanol, BPSD
T = Ethanol Sulfur, ppm
Z= FCC Gasoline Yield, BPSD
Equation 5-2 Calculating FCC Naphtha Sulfur Content for Refinery-By-Refinery Model
We used this equation to assess two cases; a baseline case where the 30 ppm Tier 2 sulfur
standards were fully implemented and a control case that reflects the 10 ppm Tier 3 sulfur
standards. The only terms in Equation 5-2 that change between the two cases are the national
average sulfur level and the sulfur levels of the LSR, NGL, and FCC naphtha streams. The
national average sulfur levels for the two cases were set at the sulfur limits under the Tier 2 and
Tier 3 programs — 30 ppm and 10 ppm, respectively. For the baseline case we assumed that the
sulfur level of the LSR stream is 15 ppm. This reflects our assessment of how these streams are
currently being handled as discussed earlier in this section. We estimate that 66 percent of the
volume of NGL and LSR are hydrotreated before being blended into gasoline and have a very
low sulfur content of approximately 1 ppm. Another 25 percent of LSR is being treated by
extractive desulfurization and is low in sulfur, likely in the 0 to 10 ppm range. The remaining 10
percent is untreated and has a sulfur content of approximately 100 ppm.
For the Tier 3 control case we assumed that all of the NGLs and LSR were either A)
hydrotreated or treated with extractive desulfurization and therefore had an average sulfur
content of 5 ppm, or B) the refinery would comply by purchasing credits and their LSR would
remain at its Tier 2 levels. This information allowed us to solve Equation 5-2 for the FCC
naphtha content. The resulting FCC naphtha sulfur numbers, along with our estimation of the
gasoline blendstock sulfur levels and percent of total gasoline volume made up by each
blendstock are shown in Table 5-16.
5-31
-------
Table 5-16 Sulfur Levels for Gasoline Blendstocks in the Refinery-By-Refinery Model
Gasoline Blendstocks
FCC Naphtha
Reformate
Alkylate
Isomerate
Butane
Light Straight Run Naphtha (LSR) and
Natural gas Liquids (NGL)
Hydrocrackate
Ethanol
Coker Naphtha
Other Gasoline Blendstocks
Total/Sulfur Average
Baseline Tier 2
Case
Percent
of Total
Volume
37.2
22.5
12.7
3.2
4.0
5.2
3.0
9.9
2.2
0.2
100
Sulfur
Levels
30
ppm
80a
0.5
10
0.5
5
15
8
5
0.5
10
30
Proposed Tier 3
Case Year 20 17
Percent
of Total
Volume
36.0
21.8
12.5
3.1
4.0
4.9
2.9
12.5
2.1
0.2
100
Sulfur
Levels
10
ppm
21a
0.5
10
0.5
5
5
8
5
0.5
10
10
Proposed Tier 3
Case Year 2030
Percent
of Total
Volume
35.0
21.2
12.1
3.1
3.8
4.8
2.8
15
2.0
0.2
100
Sulfur
Levels
10
ppm
21a
0.5
10
0.5
5
5
8
5
0.5
10
10
"These values are calculated using Equation 5-2; all other sulfur levels are assumed.
The numbers in the table above represent national averages. While this is useful
information, it is insufficient for us to be able to model the implications of the proposed Tier 3
standards for an individual refinery. Each refinery has a unique combination of processing units
that will determine the cost and operational changes necessary for that refiner to comply with our
proposed sulfur limit. While each of these processing units may impact the cost for refiners to
lower the sulfur content of the gasoline they produce we believe these costs will be dominated by
the units responsible for the desulfurization of the FCC naphtha, and to a lesser extent the NGLs
and LSR. This is because these are the only streams we anticipate would see significant sulfur
reduction under the proposed Tier 3 sulfur standards. The units that are used to desulfurize these
streams include the FCC unit pre- and postreaters and the naphtha hydrotreaters. It is important,
therefore, to have a good understanding of which of these units are in place in each refinery, as
well as the type and capacity of these units, in order to allow us to most accurately estimate the
cost of the Tier 3 sulfur standards to the refining industry. We used the above FCC naphtha
sulfur balance information as the basis of our vendor request for refiner modifications to FCC
postreaters under Tier 3. However, for the vendor requests, we used a preliminary model, where
the FCC naphtha levels under Tier 2 averaged 75 ppm, while FCC naphtha levels under Tier 3
averaged 25 ppm for 10-ppm sulfur gasoline, representing a 50 ppm sulfur reduction, close to the
5-32
-------
same delta presented in the table above.0 The following section discusses our assessment of the
desulfurization equipment currently being used in refineries.
5.1.3.3 Assessment Existing Desulfurization Equipment at Refineries
Since the desulfurization cost of the Tier 3 program is largely impacted by the cost of
lowering sulfur in FCC gasoline, it is important to understand what refiners are already doing to
lower the sulfur content of the FCC gasoline blendstock to meet the Tier 2 sulfur standards. This
was important to our analysis of the cost for each individual refiner to reduce the sulfur content
of their gasoline to meet the proposed Tier 3 sulfur standard. Refiners that already have an FCC
pretreater or postreater can revamp these units for a lower cost than installing grass roots units.
It was also important to determine which refineries have an FCC feed pretreater, since these units
increase the refineries FCC conversion and production of FCC naphtha and also lower the sulfur
level of the FCC naphtha. To compile this information we analyzed capacity information for
FCC naphtha pretreaters and postreaters for each refinery listed in the OGJ, EIA and OAQPS
database sets. If one of the databases showed that a refinery had FCC pretreating and/or post-
treating capacity, while the other did not, we assumed that the refinery did have the units listed
with a capacity as reported. Our OAQPS database had approximately six refiners that have FCC
post treaters, and that were not listed in either the EIA or OGJ data sets. We assumed that these
refiners had an existing postreater. For refineries that have FCC naphtha postreaters we next
determined which vendor's FCC naphtha desulfurization technology is installed in each refinery.
To do this we conducted a public database search using OGJ, company web postings and, other
refinery publications. To supplement this data we also had extensive discussions with many
refiners who provided us with information on the type and capacity of the desulfurization
technology currently installed in their refineries, as well as how their operations might be
adjusted to meet the new Tier 3 sulfur standards. The various FCC naphtha desulfurization
technologies that we identified as currently being used by refiners are CD Tech's Cd Hydro and
CDHDS, Axens Prime G and Prime G+, UOP's ISAL and Selectfming, Exxon's Scanfining I or
II and Sinopec's S-Zorb. For refineries for which we could not find or obtain information on the
type of desulfurization they were using, Axens was chosen as the default as they have the largest
market share of desulfurization units in the U. S. To confirm the accuracy of our work we
reviewed our assessments with one of the main technology vendors. Our desulfurization
technology selection assumptions were adjusted based on feedback from the vendor. The
aggregated results of this assessment are summarized in Table 5-17.
c Because the technology vendors provided us with cost data only for the increment of reducing FCC naphtha sulfur
content from 75 ppm to 25 ppm and in some cases from 75 ppm to 10 ppm, we modeled all refineries, regardless of
their current sulfur level, using the same technology costs. In reality, those with finished gasoline sulfur levels
higher than 30 ppm would have slightly higher costs and those with finished gasoline sulfur levels lower than 30
ppm would have slightly lower costs. We are trying to obtain additional information that would enable us to adjust
our cost analysis to reflect actual refinery starting sulfur levels.
5-33
-------
Table 5-17 Postreater Technologies Used By Refineries
Refiners with Existing
Postreater
CD Tech
15
IFP/Axens
38
Scanfming
15
UOP ISAL
2
S-Zorb
5
The next step of our analysis was to determine which refineries use FCC feed
hydrotreating technology (pretreaters) in addition to post-treating units. FCC feed hydrotreating
was primarily installed at refineries not as a sulfur control technology but because of the
economic benefits that can be obtained from hydro-treating FCC feed. Hydrotreating the FCC
feed increases the crackability of this stream by saturating the components with hydrogen
resulting in a higher paraffin content in the feed stream. Hydrotreating also removes FCC feed
impurities such as nitrogen, metals, con-carbon and sulfur, which improve FCC unit catalyst
effects. An additional benefit of FCC feed pretreating is that it reduces the sulfur content of the
FCC feedstock by 70 to 90 percent, resulting in the production of FCC naphtha with lower sulfur
levels than what would be produced using FCC feed that is not hydrotreated.
Our analysis indicates that approximately 54 refiners are currently using FCC feed
pretreaters. Of the 54 refineries with pretreaters, 35 also have FCC postreaters installed to
comply with the Tier 2 gasoline sulfur standard. The technologies used by these 35 refineries are
shown in Table 5-18. FCC naphtha produced from an FCC pretreater operating at standard
severity generally produces a gasoline with a sulfur content that exceeds the Tier 2 standards.
According to information from vendors, the average FCC naphtha sulfur level of refineries with
an FCC feed pretreater operating at standard conditions without a postreater ranges from 200 to
500 ppm. Further reductions in the sulfur level of the FCC naphtha are possible using only an
FCC pretreater by operating the pretreater at a higher severity or higher pressure (if the unit is
designed to do so). This appears to be the case for the 19 refineries using only FCC pretreaters
to comply with the current Tier2 sulfur standard. These high pressure FCC pretreating units
were designed to be able to run at a high severity to further increase the crackability of the FCC
feed and therefore increase the conversion rate of the FCC unit. These more severe conditions
also further reduce the sulfur level of the FCC naphtha. The naphtha produced from these units
operating with high severity or high pressure has an average sulfur content ranging from 75 to
100 ppm, allowing these refineries to produce gasoline that meets the Tier 2 sulfur standards.
Operating FCC pretreaters at the high severities necessary to meet the Tier 2 standards, however,
also results in increased operating cost, as the pretreater requires more frequent catalyst
changeouts, consumes more hydrogen, and operates higher temperatures than pretreaters
operating under standard conditions.
Table 5-18 Technologies Used By Refiners with FCC Pre and Postreaters
Refiners with FCC Pretreater
and Naphtha Postreater
CD Tech
9
IFP/Axens
17
Scanfming
6
UOP ISAL
1
S-Zorb
2
Our analysis also showed that there are several refineries that have an FCC unit but have
installed neither an FCC naphtha postreater nor an FCC feed pretreater. These are small
refineries, or refineries that produce a refinery gate gasoline with a sulfur level below the Tier 2
cap of 80 ppm sulfur, but above the 30-ppm average. These refiners are relying on buying or
5-34
-------
sharing sulfur credits from other refineries that are over-complying with Tier 2 and make
gasoline with a sulfur level less than 30 ppm.
Finally, some refineries do not have an FCC unit and therefore have not installed FCC
postreaters to comply with the Tier 2 sulfur standards. These refiners primarily use reformate,
alkylate, LSR, butanes, and pentanes to make gasoline. Since these blendstocks all have low
sulfur content this allows refiners to produce gasoline with a low enough sulfur content to meet
the Tier 2 sulfur standards.
A summary of the number refineries which fall into differing categories of how they are
complying with Tier 2 is shown in Table 5-19.
Table 5-19 Refinery FCC Naphtha Desulfurization Unit Characterization
FCC Treatment Units Installed
No FCC Unit
FCC Unit, No Pretreater or Postreater
FCC Unit With Postreater Only
FCC Unit With Pretreater Only
FCC Unit With Pretreater and Postreater
Number of Refineries
14
4
42
16
35
5.1.3.4 Estimate of FCC Naphtha Sulfur Levels
After we had determined the desulfurization technology in place at each refinery, we
sought to calculate the sulfur content of the FCC naphtha, which is the feedstock for FCC
postreaters. It is important to understand the sulfur level of FCC naphtha because it helps to
determine the extent that existing FCC postreaters are operating at their maximum hydrotreating
capacity and how refiners might invest to comply with Tier 3. Some refineries may have excess
capacity in their FCC naphtha pretreater or postreaters that would allow them to produce
gasoline that would meet the proposed Tier 3 standards without having to revamp existing units
or add grass roots units. These refineries will have much lower cost impacts than refineries that
have to make more significant capital investments. To estimate the sulfur of the FCC naphtha we
must start upstream of the FCC unit and first understand the sulfur level of the crude oil which is
the source for the vacuum gas oil that serves as the feedstock for the FCC units.
The sulfur level of the FCC feedstock is dependent on several items; 1) the sulfur level of
the crude oil being processed by the refinery, 2) whether or not a refinery has a coker unit and 3)
whether or not the refinery has an FCC feed pretreater. In refineries, the crude unit directly
supplies heavy atmospheric gas oils (HGO) and vacuum gas oils (VGO) streams, which are the
bulk of material used as FCC feedstock in most refineries. The sulfur level of these two streams
can be estimated using a regression correlation that is based on the actual crude unit sulfur level
and actual VGO sulfur level from crude assays. The peer reviewers agreed that our calculations
were adequate for estimating the FCC feed sulfur levels of HGO/VGO feedstocks based on crude
sulfur levels. However, if a refinery has a coker unit, the FCC unit will also likely use heavy
coker gas oil (HCGO) as a feedstock. A peer reviewer pointed out to us that the sulfur level of
HCGO is usually much higher than HGO and VGO, since coker units use heavy residual stocks
5-35
-------
as feedstock. As such, the sulfur level of HCGO cannot be estimated from the same correlation
used for HGO and VGO. As a result, for our final rule analysis we adjusted the FCC feed sulfur
levels higher to account for the refining of higher sulfur HCGO in FCC units.
To arrive at each refiner's FCC feed sulfur level, we determined the blended FCC
feedstock sulfur value based on the volumes and sulfur levels we estimate for HGO/VGO and
HCGO, as describe below.
The first step in determining the sulfur level of the FCC feedstock was to input the crude
sulfur level for each refinery into our refinery-by-refinery model. For this, we obtained
confidential business information (CBI) from EIA on the annual average crude sulfur levels that
each refinery processed in 2011. This data, which is reported to EIA for each refinery, was used
as the baseline crude sulfur level in our refinery-by-refinery analysis. Using this data, we then
determined what each refiner's HGO/VGO feed sulfur level would be, using a regression co-
relation we built from data on crude sulfur levels and FCC feedstock material, as discussed
below.
The boiling point range that we assumed for VGO also contained some residual material,
representing FCC feed with residual content. This was done to reflect the imperfect distillation
cuts in crude towers and that some refiners use small amounts of residual material as FCC
feedstock. The balance of the residual material, however, was excluded from the feed to FCCs
since this material makes a poor feedstock due to its high aromatics, metals and concarbon
content. Each of these materials negatively affects the FCC gasoline conversion yields. Most
refiners today do not directly use residuals as feedstock to their FCC units, but instead send them
to be processed in coker units or use the residual material for fuel oil and asphalt production.
The boiling point ranges that we used for HGO and VGO are listed Table 5-20.
Table 5-20 Boiling Ranges of FCC Feedstocks from the Crude Unit
Heavy Atmospheric Gas Oil
(HGO)
Vacuum Gas Oil (VGO)a
TBP Initial
600°F
800°F
TBP Final
800°F
1,000°F
a Contains some residual material
For our FCC HGO/VGO feed sulfur regression, we used various crude oil assays that we
obtained from Jacobs Engineering. We used data from five specific crude types, including West
Texas intermediate, Bonny Light, Saudi Heavy, Alaskan North Slope, and Mayan, and three
blended crude assays. The equation for this regression, along with the estimated FCC feed sulfur
contents for various crude oils are shown in Equation 5-3 and Table 5-21.
Equation 5-3 FCC HGO/VGO Feed Sulfur Content Based on Crude Sulfur Content
FCC Feed Sulfur Weight Percent = (Crude Sulfur Weight Percent)0'8 * 1.1858 + 0.0409
5-36
-------
Table 5-21 FCC HGO/VGO Feedstock Sulfur For Various Crude Sulfur Levels
Crude Sulfur Level (Weight %)
0.11%
0.28%
0.85%
1.3%
1.4%
1.6%
2.8%
3.04%
FCC HGO/VGO Feed Sulfur Level
(Weight %)
0.24%
0.47%
1.08%
1.50%
1.59%
1.77%
2.74%
2.93%
If a refiner did not have a coker unit, we used the above method to estimate the FCC
feedstock sulfur level prior to any FCC feed pretreating. However, for refiners with a coker unit,
we had to adjust the above numbers higher to account for the levels of sulfur from HCGO.
For refiners with a coker unit, the HCGO yield from the coker unit was set at 25% of the
feedstock, based on yield estimated from one of the peer reviewers. The sulfur level of HCGO
was set to be equal to the feedstock sulfur level of vacuum residual bottoms (VTB), which is
used as feedstock to the coker unit. The VTB sulfur level was assumed to be 2.2 times the
refiners EIA crude sulfur level, based on correlations recommended by the peer reviewer. Using
this approach, we estimated the HCGO volume and sulfur level that is used as FCC feedstock in
refineries with coker units. Then, the HGO/VGO volume of feedstock was determined by
subtracting any HCGO from the total EIA FCC feedstock volume that the refiner used as charge
in 2011. The resulting HCGO and HGO/VGO feedstocks were then volume averaged with the
corresponding sulfur levels to determine the final FCC feedstock sulfur level for the refiner.
This resulted in a modest rise in FCC unit feedstock sulfur levels from our NPRM analysis
approach. Table 5-22 shows the FCC feedstock sulfur results, using this methodology.
5-37
-------
Table 5-22 PADD Average Effects of HCGO on FCC feed sulfur levels
PADD
1
2
O
4&5,
excluding California
California
National Average
FCC Feed not adjusted for
HCGO, wt. % sulfur
1.15
1.46
1.52
1.30
1.54
1.44
FCC Feed adjusted for
HCGO, wt. % sulfur
1.21
1.59
1.70
1.35
1.54
1.54
With FCC feed sulfur estimated, the next step in our analysis was to consider the impact
of pretreaters on the FCC feed sulfur levels for refineries that have these units. There are several
factors that must be considered to determine the impact of the pretreaters on the FCC sulfur
level, including the pressure at which the unit operates, the severity at which it is run, and
whether or not the FCC naphtha will be postreated.
To inform our understanding of how FCC pretreaters operate, we obtained guidance from
technology and catalyst providers. From these discussions we learned that the capability for
FCC pretreaters to remove sulfur from the gas oil feed varies significantly depending on the
pressure at which the unit operates. FCC pretreaters can generally be subdivided into high
pressure units (1400 psi and above), medium pressure units (900 to 1,400 psi), and low pressure
units (below 900 psi). High pressure FCC pretreaters are capable of removing about 90 percent
of the sulfur contained in the gas oil feedstock to the FCC unit, while low and medium pressure
units are capable of removing 65 to 80 percent of the feed sulfur. Information we received from
the vendors also indicated that refiners with both a pretreater and a postreater are typically
producing FCC naphtha that ranges from 200 to 450 ppm before being processed by the
postreater. Having a postreater allows these refineries to not have to operate their pretreaters at a
high severity for compliance with the 30 ppm Tier 2 sulfur standard since the sulfur will further
be reduced to levels necessary to meet the applicable standards in the postreater. For a subset of
these refineries, the pretreater is undersized and not able to pretreat the entire FCC feed charge
volume. In response to peer reviewer comments, we therefore, limited the amount of FCC feed
that was processed for several refineries based on their unit capacity. Refineries with only a
pretreater are making lower sulfur FCC naphtha in the 75 to 100 ppm range, according to vendor
estimates. With this information we used our refinery-by-refinery model to estimate the
pretreater desulfurization rates required to get FCC naphtha sulfur levels within the ranges
specified. We estimated that FCC units with a pretreater and a naphtha postreater are operating
their pretreaters at a severity which results in a 76 percent desulfurization of the FCC feed
stream. This number represents the national average. While the actual severity at which the
5-38
-------
pretreating units are run varies on a refinery-by-refinery basis this average was used in our
modeling for all refineries with both pretreating and postreating units due to a lack of refinery-
specific information. For FCC units with a feed pretreater but no postreater we calculated the
FCC naphtha sulfur level required to make a refinery gate gasoline that meets the Tier 2
standard. To do this calculation we used the gasoline yields from our refinery-by-refinery model
along with the gasoline blendstock sulfur levels discussed in subsection 5.1.3.2 and shown in
Table 5-16. These calculations showed that refiners with FCC feed pretreating units, but no
postreaters, need to produce FCC naphtha that averages about 85 ppm on a national level. This
sulfur level corresponded to these refiners operating their pretreaters at a severity that results in a
reduction of sulfur in the FCC feed stream of approximately 91-92 percent. This number is close
to the estimate we received from the vendors for this category of refineries and therefore was
used in our refinery-by-refinery model to determine the FCC feed sulfur level for refiners with
pretreaters.
After we have calculated the sulfur level of the FCC feed we had to then take into
consideration the impact the FCC unit itself has on the sulfur level of the FCC naphtha. We
reviewed several literature sources14'15 and found that the FCC naphtha sulfur level can be
estimated by dividing the FCC feed sulfur level by 20 for refineries with an FCC feed pretreating
unit. For refineries without an FCC feed pretreater, the FCC naphtha sulfur levels can be
calculated by dividing the desulfurized FCC feed sulfur level by 10. In these cases the effect of
the FCC unit itself on the sulfur level of the FCC naphtha is lower, as the FCC feed has already
been through a desulfurization process. These factors, when combined with the sulfur levels of
the FCC streams as discussed above, allow us to calculate on a refinery-by-refinery basis the
sulfur level of the FCC naphtha before any post-treating operations. The results of this analysis
are summarized in Table 5-23.
Table 5-23 FCC Naphtha Sulfur Levels for Various Refinery Configurations
No Pretreater or Postreater
Pretreater Only
Postreater Only
Pretreater and Postreater
PADD 1
N.A.a
N.A.a
1119
45
PADD 2
-<3b
79
1233
<3b
PADD 3
-<3b
52
556
601
PADD 4/5
-<3b
61
1039
293
" N.A. - not applicable, no units of this type in the PADD
* Since there are less than three refiners in this PADD with the described configuration, the data was
removed to protect potential CBI concerns.
An additional adjustment to the FCC naphtha sulfur levels was made because of
undercutting of the FCC naphtha into the diesel fuel pool. If in the refinery-by-refinery model
we estimated that a refinery is fully undercutting its FCC naphtha to the diesel fuel pool, the
sulfur level of the FCC naphtha was reduced to half of what it would be without undercutting. If
the amount of FCC undercutting was somewhere between none and all of the swingcut being
undercut to diesel fuel, we proportionally adjusted the FCC naphtha sulfur level. For example, if
in the refinery-by-refinery model we estimated that half of the FCC naphtha swingcut was being
undercut, we estimated that the FCC naphtha sulfur level is 75 percent of what the FCC naphtha
sulfur level would be without undercutting.
5-39
-------
This information, along with the information described in previous sections (e.g.,
gasoline blendstock volumes and sulfur levels, desulfurization equipment currently in place at
refineries, and crude oil sulfur levels) allowed us to conduct the best analysis for the baseline
case for our refinery-by-refinery model. This baseline case reflects what we believe the current
operating conditions are at each refinery, including any modifications they have made to meet
the Tier 2 sulfur standards. The next step of our analysis was to project what further changes,
either to operations, adding new equipment or revamping existing units, each refiner would have
to make to meet the proposed Tier 3 standards. After these changes are estimated, we can then
estimate the cost associated with each of these changes, and ultimately the cost of the program.
5.1.3.5 Cost Inputs for the Sulfur Control Technologies
After we determined the sulfur levels of the gasoline blendstocks for each refinery and
the sulfur levels that these blendstocks would have to achieve to meet the Tier 3 sulfur standards,
the next step in our refinery-by-refinery analysis was to project the changes to refinery units and
unit operations each refinery would have to make to comply with the Tier 3 sulfur standard. One
step refiners would take to further reduce gasoline sulfur levels would be to desulfurize the light
straight run and natural gas liquids. The costs to reduce the sulfur content of these streams is
relatively low and would therefore be a cost-effective way to further reduce the sulfur content of
the finished gasoline. In addition, because ethanol tends to be a relatively low sulfur blendstock
(assumed to be 5 ppm in our refinery-by-refinery model), increasing the amount of ethanol in the
gasoline pool lowers the overall sulfur content of the gasoline. For the NPRM analysis, we
projected that 50 percent of all gasoline produced by refiners in 2017 would contain 15 percent
ethanol and that it would be almost entirely El5 by 2030. However, for our final rule analysis
we are basing our ethanol use on AEO2013, which assumes widespread E10 use with very little
E15 use and a small amount of E85 use. Reducing the sulfur content of the LSR and NGL
streams and taking advantage of ethanol blending, however, would fall far short in enabling
refiners to comply with the Tier 3 gasoline sulfur standards. Refineries with an FCC unit would
still have to reduce the sulfur content of their FCC naphtha blendstock to meet the Tier 3
standards.
For each refinery we considered two cases. In the first case each refinery had to meet the
Tier 3 gasoline sulfur standard of 10 ppm. To meet this standard, as discussed in Section 5.1.3.2,
we estimated that they would have to reduce the sulfur level of their FCC naphtha stream from
75 to 25 ppm. We also considered a case where each refinery would reduce the sulfur level of
their gasoline to 5 ppm. This information was used to help us determine which refineries might
reduce the sulfur level of their gasoline below our proposed 10-ppm standard to generate credits
for our ABT scenarios summarized in Table 5-1 and discussed in Section 5.2.1.
Our refinery-by-refinery model assumed that reducing the sulfur content of the FCC
naphtha to 25 ppm and 10 ppm for the two cases discussed above would require that each
refinery that produces FCC naphtha have an FCC naphtha postreater. For companies that
already have an FCC naphtha postreater we assumed that all that would be necessary to meet the
Tier 3 sulfur standards was to revamp their existing FCC postreating units. We received cost
information from several vendors for revamping FCC postreating units and assumed a revamp
cost for each refinery in line with the cost projections quoted by the vendor of the technology
already in place in their refinery. We assumed that refineries with FCC units that currently do
5-40
-------
not have an FCC postreater would have no choice other than to add a new grass roots FCC
postreating unit. We ultimately only received cost information from one vendor for the cost of
adding a new grass roots FCC postreating unit that fit the sulfur reduction requirements of the
proposed Tier 3 program. We therefore assumed the cost for each refinery that would need to
add a new FCC post-treating unit would be in line with this estimate. More details on the costs
used in our refinery-by-refinery model for the desulfurization of LSR and NGL, ,as well as the
new FCC post-treating units and revamps, can be found in the following sections.
5.1.3.5.1 Cost to Revamp Existing FCC Naphtha Postreaters
To estimate the cost for revamping existing FCC postreating units or for adding new
postreating capacity, we contacted several technology vendors for cost estimates and reviewed
literature, including cost information provided for the Tier 2 rulemaking. Because no two
refineries are exactly the same, the cost for new FCC postreater units or revamps to existing units
will vary significantly from refinery to refinery. Some of the factors that have the most
significant impact on the cost of FCC postreaters are the technology that the refiner used to
comply with Tier 2, the volume of FCC naphtha, the sulfur content of the FCC unit feed and the
level of desulfurization in the existing postreater, and the location of the refinery. Based on
feedback from vendors we considered three categories of FCC postreaters based on whether the
FCC naphtha (the feed for the existing Tier 2 postreater) contained low (0 - 400 ppm) medium
(400 - 1,200 ppm) or high (>1,200 ppm) levels of sulfur.
For revamp postreaters, we asked the vendors to evaluate desulfurizing FCC naphtha
from a Tier 2 sulfur level of 75 ppm, down to 25 and 10 ppm, for Tier 3 finished gasoline sulfur
levels of 10 and 5 ppm, respectively. This corresponds to a delta FCC naphtha sulfur reduction
of 50 ppm (75 ppm minus 25 ppm) for the 10 ppm standard, and 65 ppm (75-10 ppm) for the 5
ppm level. For new standalone Tier 3 post treaters, we asked the vendors to evaluate a FCC
naphtha sulfur reduction from a Tier 2 sulfur level of 100 pm, down to levels of 25 and 10 ppm,
(reductions of 75 and 90 ppm), for the 10 and 5 ppm finished gasoline levels, respectively. For
our FRM analysis, we applied the vendor cost estimates based on these sulfur ranges, to the
sulfur reductions needed for every refinery. Since each refinery has a specific FCC naphtha
sulfur reduction level needed for the Tier 3 standards, we adjusted the vendor cost estimates, so
as to apply them to each refinery's particular FCC naphtha sulfur level reduction amount that
was needed for Tier 3. We applied the vendor cost estimates to each refiners FCC naphtha sulfur
reduction level needed for Tier3, by assuming that the vendors estimate scale linearly for octane
and hydrogen for sulfur reduction levels that are less than 96 percent of the FCC naphtha Tier 2
sulfur levels. As such, the vendor cost estimates for octane and hydrogen were scaled based on
the relative increase or decrease in a refiners FCC naphtha sulfur level, versus the 50 and 65 ppm
sulfur delta used in the vendor estimates for a 10 and 5 ppm standard. All of the other vendor
cost estimates, such as capital cost, steam, electricity usage etc, were assumed to not vary from
what was supplied by the vendors. This cost adjustment was applied to all of the vendor
estimates for new and revamp FCC Naphtha post treaters. The starting (Tier 2 ) FCC naphtha
sulfur level for each refiner was determined based on our modeling of the volumes and sulfur
levels of each of the gasoline blendstocks that the refinery uses to produce finished gasoline.
We obtained information from several technology providers for the revamp costs of
existing FCC postreaters. One of the technology providers, however, declined to provide us with
5-41
-------
information applicable for Tier 3 sulfur control; they merely provided us with historical
information for controlling sulfur from pre-Tier 2 uncontrolled sulfur levels. As a result, their
capital and operating costs were extraordinarily high relative to the rest of the cost information
we received from the other technology providers that was specific to Tier 3. In addition, the
information they provided did not contain sufficient detail to enable us to adjust it to be of
relevance for Tier 3. Given the magnitude of their cost estimate, we also believe it is likely that
their cost estimate represented not only a grass roots FCC postreater, but also significant refinery
investment in other refinery processes such as FCC feed pretreating, coker unit expansion, etc.
Thus, we were unable to use this information to estimate the revamp costs for FCC post treating
for the refineries that employ that company's technology, and our peer reviewers concurred. The
other technology providers submitted information applicable to Tier 3, but not necessarily
covering all of the scenarios refineries would experience. Therefore, as discussed below, we
used and adjusted the information we were provided as necessary to apply it across all the
refineries. Some of the submitted information only had cost information for our medium sulfur
(400 - 1,200 ppm) FCC feed case. Another technology provider did not provide cost estimates
for producing FCC naphtha with a sulfur level of 25 ppm, corresponding to a finished gasoline
with 10 ppm sulfur, therefore, we needed to interpolate their cost information. Because the cost
information provided by the technology providers was labeled CBI, this cost information cannot
be listed individually, however we aggregated the cost information we received for FCC
postreater revamps to meet 10 ppm and 5 ppm sulfur levels in gasoline. The aggregated
information is summarized in 5-25 and 5-26.
One of the vendors we contacted for a cost estimate for FCC naphtha desulfurization
technology provided information for several potential FCC postreater revamp cases. The first
case was a no capital costs case where refiners made no equipment modifications, but rather
solely made operational changes using their existing equipment installed for Tier 2. The second
case we requested was one where refiners would incur only minor capital costs and was intended
to be used for analyzing program options with moderate octane costs. The third case we
requested was one where refiners were willing to incur greater capital costs in order to minimize
operating costs and octane loss. The majority of the vendors only supplied cost estimates for the
third case, which included adding an additional catalyst reactor bed to the existing FCC
postreater unit (i.e. revamping their existing FCC postreater). This addition of catalyst reactor
bed ensures that refiners will be able to run their existing FCC postreater at 4 to 5 year catalyst
cycle lengths, which is a critical feature for FCC unit operations.
The costs for the FCC postreater revamps submitted by one of the vendors, however
showed that for low (0 - 400 ppm) and medium (400 - 1,200 ppm) sulfur FCC naphtha sulfur
levels, the second case, with low capital costs, resulted in the lowest cents per gallon costs for
meeting the proposed 10-ppm Tier 3 standards. According to this vendor, these cases also had a
4 to 5 year catalyst cycle length, equivalent to the higher capital cost cases even though a second
stage reactor was not required. We therefore assumed that refineries using this vendor's
technology would choose the minor capital cost pathway for meeting the 10-ppm Tier 3 standard
when they had low or medium sulfur levels in their FCC feed. The high capital cost cases for
producing gasoline to meet a 5 ppm sulfur standard from low and medium sulfur FCC feeds were
found to have the lowest cost on a cents per gallon basis and were therefore selected by our
model for these cases.
5-42
-------
One vendor only submitted information for postreater revamp cost estimates for FCC
naphtha in the 400 - 1,200 ppm sulfur category that produced a 5 ppm sulfur gasoline. In our
refinery-by-refinery model, however, we had multiple refineries with FCC feed sulfur levels in
the 0 - 400 and >1,200 ppm categories that use this vendor's postreating technology. In order to
apply this vendor's cost estimate to cases of low (0 - 400 ppm) and high (>1,200 ppm) sulfur
feed categories we adjusted this vendors 400 - 1,200 ppm postreater revamp cost based on the
cost differentials between the three FCC naphtha sulfur levels in the other vendors' revamp
estimates. We similarly derived a postreater revamp cost estimate to produce a 10 ppm gasoline
for this vendor using cost differentials between the 10 and 5 ppm cases from other vendors. For
refineries currently employing technology by other vendors for which we had no specific cost
information, we used an average of all of the vendors' estimates to represent FCC postreater
revamp costs for refiners using this particular technology in our refinery-by-refiner model.
After we had determined cost estimates for the FCC postreater revamps based on
information from the vendors the next step was to scale these costs based on the size of the FCC
postreating unit present in each refinery. The vendor estimates submitted for revamp costs were
based on various FCC postreater design volumes ranging from 10,000 BPSD to 45,000 BPSD
depending on the base unit size used by the vendor. To determine how to apply these vendor
costs to each refinery, we first calculated each refinery's maximum FCC naphtha production.
The maximum production was derived by assuming each refiner runs their FCC unit at its
maximum nameplate throughput capacity (barrels per stream day) with the FCC naphtha yield
rates discussed in Section 5.1.3.1.2. After sizing the FCC postreater that would be required for
each refinery we then scaled the costs given by the vendors using the six-tenths rule as shown in
Equation 5-4. This is a "rule of thumb" cost estimating tool commonly used for cost estimating
by the refining and petrochemical industries for estimating the cost of a process unit based on a
similar unit of differing size.
Cost to Revamp an FCC Unit= A * (B/C)0'6
Where:
A = Cost Estimate Received from the Vendor
B = Size of the FCC Unit in the Refinery
C = Size of the FCC Unit in the Vendor's Estimate
Equation 5-4 Six-Tenths Rule for Estimating Capital Cost
We also adjusted the costs submitted by the technology providers based on the location of
each refinery. We assumed that each vendor's estimate was based on revamping an FCC
postreater in PADD 3 (Gulf Coast), which is the lowest cost region for installing new capital in
refineries. The cost for refineries that are not located in PADD 3 were adjusted upwards based
on a ratio of the cost of refinery capital projects in the PADD in which they are located relative
to PADD 3. An additional factor was applied to account for the "offsite" costs that are incurred
when installing new capital in refineries. When vendors provide a cost estimate for their
technology, this estimated cost is called the inside battery limits (ISBL) cost and it is solely for
the immediate unit of interest. However, refiners may need to install peripheral capital to
support the new unit, such as electrical switchgear, a control room, storage for feed, intermediate
or unit products, and longer than anticipated pipeline runs - these costs are usually considered
Outside battery limit (OSBL) costs, or offsite costs. In some cases, OSBL costs may include
5-43
-------
hydrogen and sulfur plant costs, although, for our analysis, we separately estimated the cost for
providing additional hydrogen and for processing the removed sulfur and included this cost in
our cost analysis. Based on feedback from our peer reviewers, for the FRM analysis, we
increased the grassroots and revamp offsite cost factors for desulfurization units significantly to
1.35.
To reflect the capital costs in the per-gallon costs, we amortize the capital costs over the
volume of gasoline produced. The Office of Management and Budget (OMB) set guidelines
how this is to be done. The capital costs are to be amortized assuming a before-tax 7 percent
return on investment (ROI), which OMB believes reflects the societal cost of complying with
environmental regulations. Using this criterion, we derived a capital amortization factor which is
0.11. Thus, the aggregate capital costs are multiplied times the capital amortization factor to
derive a yearly capital cost charge. Also in the OMB guidelines, OMB stated that the program
costs should be estimated based on how the industry would assess costs to achieve a payback on
capital invested. For a capital payback cost analysis, we assessed costs assuming an after-tax 10
percent ROI. For the capital payback cost analysis, we derived a capital amortization factor
which is 0.16. The other relevant factors we used in deriving the capital cost amortization
factors are: a 10 year depreciation life, a 15 year economic life and 39 percent federal and tax
rate.
These cost factors, as well as the utility prices that we used in our refinery-by-refinery
cost model, are shown in Table 5-24:
Table 5-24 Cost Factors for Various PADDs
Capital Cost Factor
Natural Gas ($/MMBTU)
Electricity (e7kW-hr)
Steam ($/l, 000 Ib)
Offsite Capital Cost Factor -
New Units
Offsite Capital Cost Factor -
Unit Revamps
Capital Amortization Factor
Before Tax 7% ROI
Capital Amortization Factor
After Tax 10% ROI
PADD 1
1.5
8.91
8.66
20.56
1.35
1.35
0.11
0.16
PADD 2
1.3
7.77
6.17
17.89
1.35
1.35
0.11
0.16
PADD 3
1.0
6.31
5.87
14.23
1.35
1.35
0.11
0.16
PADD 4
1.4
6.78
5.36
17.03
1.35
1.35
0.11
0.16
PADD5a
1.2
7.69
8.83
17.24
1.35
1.35
0.11
0.16
a Excluding California
The volume-weighted cost estimates for revamping FCC postreaters across the entire
refining industry as calculated by our refinery-by-refinery model are shown in Table 5-25 and
Table 5-26. These costs are aggregated cost estimates for the FCC revamp costs used in our
refinery-by-refinery model. In our model, we paired vendor cost data with refineries that are
already using that particular vendor's technology for their FCC postreating units. We further
tailored the information provided by the vendors to match the specific refinery configuration to
the extent possible. We assumed that the data provided by the vendors includes the cost for
5-44
-------
complying with the applicable stationary emission standards and that any incidental costs for
permitting, if necessary, are negligible and covered by the offsite factor. The information that
we received from the vendors and the individual refinery capital costs, however, cannot be
shown due to CBI concerns.
Table 5-25 Revamp Cost for a 30,000 BPSD FCC Postreater for 10-ppm Standard
FCC Feed Sulfur Level*
Capital Cost ($/B ISBL)
Hydrogen (scf/bbl)
Fuel Gas (kBTU/bbl)
Electricity (kWh/bbl)
Octane Loss (R+M)/2
Olefm Decrease (vol%)
Catalyst Cost ($/B)
Steam (Ib/bbl)
0 - 400 ppm
294
46.3
8.5
0.10
0.68
2.44
0.02
6.23
400 -1,200 ppm
214
51.3
7.6
0.12
1.00
2.57
0.03
6.96
>l,200ppm
592
58.5
3.1
0.51
0.93
1.31
0.02
19.86
Volume Weighted
Average
334
42.0
5.3
0.17
0.77
1.89
0.02
6.25
a $/B = dollars per barrel, scf/bbl = standard cubic feet per barrel; kBTU/bbl = thousand BTU per barrel; kWh/bbl =
kilowatt-hours per barrel; (R+M)/2 = (research octane + motor octane)/2; vol% = volume percent; $/B = dollars per
barrel; Ib/bbl = pounds of steam per barrel of feed.
b Of the refineries that are expected to revamp their FCC naphtha hydrotreater for the no ABT case, 28 have FCC
naphtha sulfur levels in the 0 - 400 ppm range, 30 have FCC naphtha sulfur levels in the 400 - 1200 ppm range and
17 have FCC naphtha sulfur levels greater than 1200 ppm.
Table 5-26 Revamp Cost for a 30,000 BPSD FCC Postreater for 5 ppm Standard"
FCC Feed Sulfur Level
Capital Cost ($/B ISBL)
Hydrogen (scf/bbl)
Fuel Gas (KBTU/bbl)
Electricity (kWh/bbl)
Octane Loss (R+M)/2
Olefm Decrease
Catalyst Cost ($/B)
Steam (Ib/bbl)
0 - 400 ppm
470
39.3
13.36
0.23
0.97
1.45
0.02
29.17
400- 1,200 ppm
530
64.9
3.07
0.47
1.10
2.46
0.03
6.96
>1,200 ppm
652
55.3
4.45
0.51
1.04
2.12
0.02
19.86
Volume Weighted
Average
487
45.69
5.53
0.33
0.94
1.82
0.02
12.71
a Assumes every refinery is complying with a 5 ppm gasoline sulfur standard
We also found that there were 16 refineries that had an existing FCC postreaters that
were not sized large enough to process their maximum FCC naphtha production volume. For
these refineries we assigned additional capital costs to debottleneck the existing first stage
reactor in order to increase the postreater capacity so that it could accommodate maximum FCC
naphtha production. For each refinery with an existing unit that could not process more than 70
percent of our estimate of a refiner's maximum FCC naphtha production we added capital costs
to revamp and expand the first stage to increase its capacity to allow the postreater to process
100 percent of its maximum FCC naphtha rate. For the capital costs for this debottlenecking we
used 35 percent of the cost of a new grass roots unit (discussed below) for the volume of the
expansion. We once again used the six-tenths rule to adjust the capital cost for the volume
5-45
-------
expansion needed versus the cost for the 30,000 BPSD grass roots treater used for technology
vendor estimates.
5.1.3.5.2 Cost for Grassroots FCC Postr eaters
While all refineries that already have FCC postreaters should be able to meet the Tier 3
standards by revamping their existing postreaters, refineries that do not currently have an FCC
postreater would have to add a grass roots FCC postreater. To determine the cost of building
grass roots FCC postreating units at a refinery we similarly requested cost estimates from
vendors. Only one of the vendors that supplied FCC postreating equipment submitted
information on the cost of a grass roots FCC postreating unit for desulfurizing FCC naphtha with
a feed sulfur content of 100 ppm. Based on the calculation methodology shown in Equation 5-2,
we estimate that refineries that require a grass roots postreater will already have an FCC feed
sulfur level that averages between 85 and 100 ppm as these refineries already have FCC feed
pretreaters. The other grassroots vendor estimate we received, as well as those we received for
Tier 2, represented a grass roots postreater with an FCC feed sulfur content of about 800 ppm.
These estimates were deemed to be not representative of the costs to refineries that would be
installing grass roots postreating units as the capital, hydrogen, and other operating costs would
be much higher for an FCC feed sulfur of 800 ppm vs. 100 ppm. We did not consider this other
vendor's cost estimate for a grass roots postreater and therefore relied on a single vendor's cost
estimate for grass roots FCC postreating units for the Tier 3 program. In our FRM analysis we
expanded our analysis for the vendor's cost estimate of a new standalone unit by incorporating
additional cost information from other vendors and literature sources on new units. Based on
this other information, the cost estimate we obtained for a new Tier 3 unit, seemed reasonable
relative to the other cost data that we have for higher levels of desulfurization.
The vendor estimate submitted for a grass roots postreater was based on a postreater with
a capacity of 30,000 BPSD capable of producing an FCC naphtha with a sulfur level of 10 ppm,
corresponding to a gasoline sulfur level of 5 ppm. To scale the cost submitted by the technology
vendor to be applicable to a specific refinery, we used a similar methodology to that which was
used for postreater revamps. We first determined the appropriate size for each unit based on
each refiners maximum FCC naphtha production rate. We then used the six-tenths rule
(Equation 5-4) to scale the cost reported by the vendor up or down as appropriate based on the
relative volume of the grass roots unit required by the refinery and the size on which the
vendor's cost estimate was based. We once again assumed that the capital cost from the
technology vendor was representative of a refinery in PADD 3 complying with the applicable
stationary emission standards. We then adjusted the cost based on the cost of refinery capital
projects in the PADD in which they are located relative to PADD 3. Finally, we used a new unit
offsite adjustment factor listed in Table 5-24 to estimate the final cost of a grass roots FCC
postreater for each refinery. The costs to produce FCC naphtha with a sulfur level of 25 ppm
(corresponding to a 10-ppm gasoline) were estimated based on the grass roots postreater unit that
makes FCC naphtha for the 5-ppm standard. These costs are summarized in Table 5-27 and
Table 5-28 below.
5-46
-------
Table 5-27
Cost for a 30,000 BPSD Grass Roots FCC Postreater for 10-ppm Standard
FCC Feed Sulfur Level
Capital Cost ($/B ISBL)
Hydrogen (scf/bbl)
Fuel Gas (KBTU/bbl)
Electricity (kWh/bbl)
Octane Loss (R+M)/2
Olefin Decrease
Catalyst Cost ($/B)
Steam (Ib/bbl)
100 ppm
1500
94.2
7.5
1.06
0.83
4.02
0.04
20.0
Table 5-28
Cost for a 30,000 BPSD Grass Roots FCC Postreater for 5-ppm Standard
FCC Feed Sulfur Level
Capital Cost ($/B ISBL)
Hydrogen (scf/bbl)
Fuel Gas (kBTU/bbl)
Electricity (kWh/bbl)
Octane Loss (R+M)/2
Olefin Decrease
Catalyst Cost ($/B)
Steam (Ib/bbl)
100 ppm
1500
113
9
1.06
1.00
5.15
0.04
20.0
5.1.3.5.3 Adjustments to Vendor Costs for A typical Levels of
Desulfurization
The information that we obtained from the vendors estimated the desulfurization cost for
a typical refinery which is reducing its gasoline sulfur from 30 ppm to 10 ppm, or 5 ppm. Since
the refiner would need to reduce the sulfur in its FCC naphtha, we estimated that for a typical
refinery, the refiner would need to reduce its FCC naphtha by 50 ppm for 10 ppm gasoline, or
reduce its FCC naphtha by 65 ppm for 5 ppm gasoline. However, even if every refinery were
reducing its gasoline sulfur from 30 ppm to 10 or 5 ppm, because FCC naphtha comprises a
different fraction of the gasoline pool, refineries would need to desulfurize their FCC naphtha
different amounts. Added to this, under Tier 2, US refineries produce gasoline which ranges
from well under 10 ppm and up to the cap at 80 ppm. Thus, to best model the cost for each
refinery, we needed to develop a strategy for adjusting the desuflurization cost to capture the cost
for desulfurizing FCC naphtha different increments than the 50 or 65 ppm that the vendor data is
based on.
We reviewed desulfurization curves from desulfurization vendors for desulfurizing FCC
naphtha. The curves show that the octane loss for desulfurizing FCC naphtha is fairly linear
when the desulfurization severity is below a certain point. Thus, if a particular vendor estimated
a 0.5 octane number loss for a 50 ppm sulfur reduction in the FCC naphtha, we assumed that if a
particular refinery would require a 100 ppm sulfur reduction, then the octane loss would be 1
5-47
-------
octane number. Similarly, if that refinery would only require a 25 ppm reduction in the FCC
naphtha, then the octane number loss would be 0.25 octane numbers. Since hydrogen is
consumed to saturate the olefms that leads to the octane loss, we scaled hydrogen in the same
manner. We, however, didn't scale utility demands, but rather assumed that the estimated utility
demands are the same regardless of the change in sulfur required.
Above a certain desulfurization severity point, the octane loss seems to increase in an
exponential manner, and if we did not account for this we would be underestimating the
desulfurization cost for some refneries. The point at which that occurs seems to depend on the
capital investment made to hydrotreat the FCC naphtha. For example, if a refinery has a one
stage reactor and it uses distillation to separate the light FCC naphtha, the cost curves suggest
that the octane loss for this hydrotreater would begin to increase exponentially when the percent
desulfurization exceeds 93 - 96%. For complying with Tier 3, we assume that many revamps
fall into this category. Alternatively, FCC postreaters which are two stage units do not seem to
reach the point where the octane loss begins to increase exponentially until the percent
desulfurization is in the high 90s, such as 98% or 99%. We asked one of the
hydrodesulfurization vendors which licenses an FCC postreating technology where to draw the
line between linear and nonlinear octane loss, and the vendor suggested at 99% desulfurization
and we used that value in our cost model.
5.1.3.5.4 Light Straight Run and Natural Gas Liquids
Desulfurization Costs
Another action refiners may need to take to reduce the sulfur content of their gasoline is
to desulfurize their light straight run naphtha (LSR) and natural gas liquids (NGL) blendstocks.
While these blendstocks have lower sulfur contents than the FCC naphtha, in some cases, the
sulfur levels may still be too high and it would be cheaper to desulfurize than FCC naphtha for
refineries that are not already treating these streams. Many refineries have been desulfurizing
some or all of these blendstocks using extraction desulfurization technologies, such as Merox or
Merichem, due to the very strong odor of the mercaptans. The extraction desulfurization
technologies can usually reduce the sulfur levels of LSR and NGL by about 2/3rds by removing
most of the mercaptans. One third of the sulfur is not removable by these technologies either
because the sulfur compounds are thiophenes or are heavier mercaptans, which cannot be
extracted using these technologies. If refiners find that they need to resort to hydrotreating
because the sulfur levels are too high, they may be able to use existing excess hydrotreating
capacity in their naphtha hydrotreaters or FCC naphtha postreaters. Additionally, as opposed to
hydrotreating FCC naphtha which contains olefms, the LSR and NGL blendstocks contain no
olefms and therefore, hydrotreating them does not result in octane loss and has a lower hydrogen
consumption. The combination of the potential for using excess capacity in existing units and
low operating costs result in the relatively low desulfurization costs for the LSR and NGL
blendstocks. From our discussions with refiners, several refineries indicated that they would
install new standalone hydrotreaters for processing LSR and NGL blendstocks, though it is
unclear which other refineries will have to add equipment to desulfurize LSR and NGL. To
determine the cost to desulfurize the LSR and NGL blendstocks we first had to determine the
volume of blendstock that requires desulfurization. Our determination of the quantity of LSR
and NGL used as gasoline blendstock at each refinery is discussed in Section 5.1.3.1.5. From
this total we then subtracted the volume of LSR processed in the isomerization unit. Because the
5-48
-------
isomerization units require a low sulfur feedstock we assume that refiners with isomerization
units are treating all the light straight run naphtha in the naphtha hydrotreater. Refiners likely
determine how they desulfurize their naphtha depending on whether they have an isomerization
unit. If a refinery has an isomerization unit, the refiner likely hydrotreats the straight run naphtha
before sending the stream to a splitter, after which the heavy straight run naphtha is sent to the
reformer while the light straight run naphtha is sent to the isomerization unit. If a refinery does
not have an isomerization unit, the refiner likely sends the straight run naphtha to the splitter and
afterwards, the heavy straight run naphtha is sent to the naphtha hydrotreater before it is
reformed, while the light straight run naphtha is blended directly into the gasoline pool.
We also evaluated the likely sulfur level of LSR associated with crude sulfur level. Table
5-29 summarizes the name of the crude oil, the total sulfur level and sulfur level of the LSR from
a crude assay (or in the case of Bakken and Eagle Ford fracked crude oil, data was provided by a
peer reviewer), and the expected sulfur level of the LSR after using extractive desulfurization.
5-49
-------
Table 5-29 Sulfur of LSR associated with Crude Sulfur Level
Crude Oil Name
Alaskan North Slope
Kern River
Ecuador Export
Saudi Heavy
Saudi Light
Saudi Medium
West Texas
Intermediate
Bonny Light
Bow River
Cabinda
Maya
Canadian Interprov.
West Texas Sour
Bakken
Eagle Ford Light
Eagle Ford Heavy
Crude Oil Sulfur
Level (wt%)
1.04
1.1
1.0
2.82
1.8
2.32
0.28
0.11
2.96
0.1
3.04
0.37
1.57
0.15
0.01
0.12
LSR Sulfur Level
(ppm)
26
330
10
6
200
108
0
48
65
5
50
0
1510
4
1
10
LSR Sulfur Level
after Extracting Sulfur
(ppm)
9
109
3
2
66
36
0
16
21
2
17
0
498
1
0
4
Table 5-29 shows that lower sulfur crude oils tend to have very low sulfur LSR when
extractive desulfurization is factored in. In fact, for crude oils with less than 1 percent sulfur,
there is only one crude oil where the LSR contains more than 5 ppm sulfur after extractive
desulfurization is accounted for, and that is Bonny Light. However, imported light crude oils
such as Bonny Light (which is imported from Nigeria) are likely being displaced by domestically
fracked crude oils such as Bakken or Eagle Ford, the LSR of which is very low in sulfur. Based
on this data, we assumed that refiners would solely rely on extractive desulfurization for
5-50
-------
desulfurizing their LSR under Tier 3 if the crude sulfur being refined at their refineries was
below 1 percent sulfur.
The next step in our assessment of the desulfurization costs of the LSR and NGL
blendstocks was to estimate the extent to which higher sulfur LSR (refineries with crude oil
sulfur levels above 1 percent and without an isomerization unit) is already being treated at
refineries to meet the existing Tier 2 sulfur standards. Based on our discussion with refining
consultants, vendors and refiners, it appears that in response to the Tier 2 standards refiners have
altered their operations to use excess capacity in their FCC naphtha postreaters and naphtha
hydrotreaters to reduce the sulfur content of LSR and NGL blendstocks. Since information on
the extent to which these streams are currently being hydrotreated is not publicly available we
estimated these volumes using the capacities of the FCC postreaters and reformer feed
hydrotreaters under normal refiner crude throughputs and yields from the refinery-by-refinery
model.
We evaluated each refinery's capacity to hydrotreat LSR and NGL using existing
equipment by first determining the volume that can be processed in their naphtha hydrotreaters.
We compared the throughput volume of each refinery's naphtha hydrotreater with the calendar-
day capacity of that unit. The difference between the two represented the available capacity of
that unit. If a refinery had insufficient excess capacity in their naphtha hydrotreater to treat all of
the LSR and NGL volumes we next determined if there was excess capacity in that refinery's
FCC postreater. We allowed LSR and NGL to be processed using excess FCC postreating
capacity in refineries where the capacity of the FCC postreater exceeds 120 percent of that
required to process a refinery's maximum FCC naphtha yield as determined by the refinery-by-
refinery model. Several refiners had excess FCC postreating capacity available for the treating
of LSR or NGL feedstocks, as the capacity of most FCC postreaters was less than 120 percent of
the maximum FCC naphtha production rate. We assumed that refiners are currently using any
excess hydrotreating capacity in their naphtha hydrotreating and FCC postreating units to
desulfurize LSR and NGL in response to the Tier 2 sulfur standards.
If a refinery did not have sufficient excess hydrotreating capacity for all of the LSR and
NGL in these units we assumed the refinery would have to either revamp their existing
equipment or add new hydrotreating capacity. If the additional capacity needed at any given
refinery exceeded the existing naphtha hydrotreater capacity by less than 30 percent we assumed
the necessary capacity could be added by revamping the existing unit. If, however, the additional
capacity required exceeded the existing reformer feed hydrotreater capacity by more than 30
percent we assumed the refinery would install a new stand-alone hydrotreater to desulfurize the
excess LSR and NGL. Based on available capacity in our refinery-by-refinery model and 2010
and 2011 refinery unit throughput data, we estimated that refiners are already hydrotreating 66
percent of the LSR and NGL that are directly blended into gasoline (excluding LSR processed in
the isomerization units). The results of this assessment are shown in Table 5-30.
5-51
-------
Table 5-30 Refineries Adding Hydrotreating Capacity for LSR and NGL
Number of
Refineries
New Hydrotreater
(No FCC Unit)
4
New Hydrotreater
(with FCC Unit)
11
Revamped
Hydrotreater
(No FCC Unit)
1
Revamped
Hydrotreater
(with FCC Unit)
7
We conservatively evaluated the capital cost required for hydrotreater revamps and new
units by assuming that refiners will size their hydrotreater equipment needs to treat all production
volumes of LSR and NGL based on each refinery's maximum crude run rate. The operating
costs used in our refinery-by-refinery model, however, are based on LSR and NGL blendstock
volumes from the model's yields based on 2011 operational crude throughputs and crude oil
quality as discussed in Section 5.1.3.1.5. Sizing the equipment this way allows each refiner to
have excess hydrotreater capacity utilization, which is beneficial in the event of process unit
shutdowns and to reprocess blendstocks from abnormal operations.
Our estimate for the cost of adding a new hydrotreater at a refinery was obtained from
Gary and Handework's Petroleum Refining Technology and Economics, page 182-183, Curve C,
Table 9.1, 30,000 BSD unit. The capital cost for a grass roots hydrotreater listed by this source
was for a hydrotreater with a capacity of 30,000 BPSD and was based on 1999 dollars. We
multiplied this cost by 1.534 to determine the equivalent cost in 2010 dollars based on the
relative increase in the Nelson Refining Construction index from 1999 to 2010 (listed as 1497
and 2296 respectively). We used the six-tenths rule to scale the capital cost listed in Petroleum
Refining Technology and Economics to those of differing capacities based on relative size of the
desired unit. We assumed a hydrogen consumption of 40 SCF/Bbl for the processing of LSR
and NGL blendstocks which we obtained from the Jacobs Refining LP modeling database for
naphtha hydrotreating as this information was not presented in the literature source. For
refineries that only required a revamp of existing units we assumed a capital cost equivalent to
40 percent of the cost of a new hydrotreating unit of equal size. We assumed equivalent
operating costs for new hydrotreating units and revamped units. The capital and operating costs
for these hydrotreating units that were used in our model is shown in Table 5-31.
5-52
-------
Table 5-31 Capital and Operating Costs for LSR and NGL Hydrotreaters
Capital ($/BBL ISBL)
Hydrogen (SCF/BBL)
Fuel Gas (kBTU/BBL)
Electricity (kWh/BBL)
Octane Loss (R+M)/2
Change in Olefins (vol%)
Steam (Ib/BBL)
Catalyst Cost ($/BBL)
New Hydrotreating Units
1380
40
100
2.0
0.0
0.0
6.0
0.03
Revamped Hydrotreating Units
550
40
100
2.0
0.0
0.0
6.0
0.03
5.1.3.5.5 Using FCC Pretreaters to Comply with Tier 3
5.1.3.5.5.1 FCC Pretr eater Sulfur Control Cost
We also assessed the cost of utilizing an existing FCC pretreater to comply with Tier 3.
To understand the feasibility and estimate the cost for using FCC pretreaters to comply with Tier
3, we conducted a literature survey and spoke to vendors and refiners. In our literature search
and conversation with vendors, we found that the feasibility and cost for using current FCC
pretreaters for complying with Tier 3 is dependent on many factors. These factors include the
pressure of the FCC pretreater, the current turnaround schedule of the unit, the price of natural
gas used to produce hydrogen for the unit, the crude oil sulfur level which determines the sulfur
level of the gas oil feed, the percent desulfurization that must be achieved to comply with Tier 3,
the refiner's future product marketing plans, and whether the refiner is willing to spend capital
dollars to revamp the existing unit. Since we don't know the response to these issues that we
identified, we found it useful to conduct several different cost assessments to estimate the costs
for a range of situations that different refiners may be in with their FCC pretreaters.
FCC pretreaters hydrotreat the gas oil which boils in the range of 690F to 1000F. FCC
pretreaters enable greater production of FCC naphtha by the FCC unit by saturating aromatic
compounds in the gas oil. These saturated aromatic compounds can be cracked more easily
which allows the FCC unit to produce more FCC naphtha. Even before the Tier 2 gasoline
standards were put in place, many refiners installed FCC pretreaters to increase the production of
FCC naphtha. To comply with Tier 2, more refiners installed FCC pretreaters, not only because
the refinery was able to comply with Tier 2 by hydrotreating the feed to the FCC unit, but also
because refiners were able to realize a return on investment by producing more FCC naphtha
(increase gasoline production). FCC pretreaters also improve the quality of the light cycle oil, a
diesel fuel blendstock, which is produced by the FCC unit.
Some refiners who installed FCC pretreaters for Tier 2 installed repurposed reactors
which were originally installed for other purposes and these can affect a refiner's ability to
comply with Tier 3. For example, refiners installed hydrotreaters to hydrotreat residual fuel, but
found that the hydrotreating catalyst is quickly diminished by the high metal content in residual
fuel. Refiners then reused the very high pressure (2000+ psi) reactors as their FCC pretreaters.
Other refiners repurposed diesel fuel hydrotreaters as FCC hydrotreaters and put in higher
pressure diesel hydrotreaters when the diesel fuel sulfur standards took effect. The repurposed
5-53
-------
diesel fuel hydrotreaters were likely low pressure (i.e. 600 pound per square inch (psi)) units.
While refiners constructing a grassroots FCC pretreater would target a 3 to 4 year cycle length
and a higher pressure unit (i.e. 1400 psi), if a refiner instead repurposed an existing reactor for its
FCC pretreater, the cycle lengths of its FCC pretreater may be shorter, and the unit's pressure
could be lower pressure, than desired. As we discuss in section 4.2.3.1, refiners with low
pressure FCC hydrotreaters are likely limited in their ability to use those units for complying
with Tier 3. If refiners are currently experiencing short cycle lengths with their FCC pretreaters,
the cost and inconvenience of even shorter cycle lengths would likely deter any refiner from
using existing FCC pretreaters to comply with Tier 3 without adding additional reactor volume.
In conducting our cost analysis, we based our cost analysis on a refinery which is solely
using an FCC pretreater today to comply with Tier 2. We assume that this unit currently requires
a turnaround every two years. From EIA data for that refinery, the throughput to the FCC
pretreater is 75,000 bbl/day and the total gasoline volume is a little more than over twice that
volume. Our refinery-by-refinery model estimates that this refinery is achieving about 92%
desulfurization of the FCC feed, which allows the refinery to reduce its FCC naphtha to about 80
ppm. To comply with Tier 3, the refinery would need to achieve about 98% desulfurization of
its FCC feed to comply with Tier 3.
We analyzed three different pathways that the refiner could take if it were to solely rely
on its FCC pretreater to further reduce the sulfur of the feed to its FCC unit to comply with Tier
3. In the first case (Case 1), the refiner solely increases the severity of its FCC pretreater, by
raising the temperature of the reactor, which further reduces the sulfur of the FCC naphtha. In
this case we assume that the unit is capable of further saturating aromatics and creating
additional yield benefits. However, for this first case, the cycle length of the FCC pretreater
shortens from 2 years to 1 year, so the refiner would have to replace the catalyst in the FCC
pretreater every year instead of every two years.
In the second case (Case 2), the refiner revamps the FCC pretreater by adding a second
reactor, which effectively doubles the reactor volume. This allows the refinery to comply with
Tier 3, achieve additional yield benefits and still maintain a two year cycle length with its FCC
pretreater.
In the third case (Case 3), we assume no additional investments (like Case 1), but in this
case the refinery is assumed to have a lower pressure FCC pretreater and the FCC pretreter has
reached its poly-aromatics saturation maximum. Thus, while the FCC pretreater is able to
achieve the desired desulfurization, it does not realize additional yield benefits.
The cost factors for increasing the severity or or revamping an existing FCC pretreater
(for all three cases) are from a National Petrochemical and Refiners Association (NPRA) Annual
Meeting technical report.16 The report includes yield estimates, hydrogen demands and a capital
cost for adding additional reactor volume for 90%, 98% and 99% desulfurization. We estimate
that the refinery we are studying is desulfurizing its FCC naphtha by 92% and we estimate that it
must desulfurize its FCC naphtha at 98% to comply with Tier 3 (produce 10 ppm gasoline).
Thus, we used the differences between 90% HDS and 98% HDS in our cost analysis. Table 5-32
summarizes the changes in yield, hydrogen consumption and capital costs for operating an FCC
pretreater at a higher desulfurization level.
5-54
-------
Table 5-32 Yield Changes and Other Impacts of Higher Severity FCC Pretreating
H2S (wt%)
C2 (wt%)
C3-C4 (wt%)
Full Range FCC naphtha
(wt%)
LCO (wt%)
CSO (wt%)
Coke (wt%)
Total
Naphtha Sulfur (ppm)
LCO sulfur (ppm)
LCO cetane index
Hydrogen Addition (wt%)
Capital Cost
90% HDS
0.1
3.5
17.6
51.5
15.7
6.6
5.0
100
225
3400
25.7
0.51
61.8
98% HDS
0.0
3.2
18.7
52.5
15.0
5.9
4.7
100
55
900
26.4
0.74
70.3
Difference
-0.1
-0.3
1.1
1.0
-0.7
-0.7
-0.3
-
170
-2500
0.7
0.23
8.5
To estimate the cost impacts of operating the FCC pretreater at a higher level of
desulfurization, it was necessary to estimate the costs for the changes in yield, the increased
hydrogen demand, the cost of additional capital. We used the component cost values for ethane,
propane and butane, FCC naphtha, light cycle oil (LCO), and cycle slurry oil (CSO) from our LP
refinery model for the year 2018. The value of hydrogen is from our LP model based on natural
gas which averages $7.1 per million BTU. The hydrotreating catalyst is estimated to cost $244
per barrel per stream day. These cost factors are summarized in Table 5-33.
5-55
-------
Table 5-33 Economic Value of Yield Components, Hydrogen and Catalyst
FCC Product or Reactant
C2 ($/bblFCC)
C3-C4 ($/bblFCC)
Full Range FCC naphtha ($/bbl)
LCO ($/bbl)
CSO ($/bbl)
Hydrogen ($/KSCF) based on natural gas at 7. 1
$/mmbtu
Value
9.2
66.5
115.6
122
87
4.1
The economic value of the FCC unit yield changes for the three cases is estimated by
converting the weight percent percentages to volume percent, multiplying each times the FCC
unit feed, which is 75,000 bbl/day and then multiplying the resulting yield changes in barrels per
day by the respective economic value of each hydrocarbon product in dollars per barrel. The
hydrogen is calculated in a similar way. We estimate that the hydrogen consumption
corresponds to about 300 standard cubic feed per barrel of feed. We multiplied this times $4.1
per thousand standard cubic feet. The capital costs were inflated from 1999 dollars to 2011
dollars using Nelson-Farrar inflation index, which increased the capital costs by a factor of 1.63.
We applied a 20% contingency factor and a 35% offsite factor. Finally, the costs were scaled
using the sixth-tenths rule to estimate the cost for the refinery unit processing 75,000 bbl/day.
For Cases 1 and 3, we estimate that the FCC pretreater must be shutdown every year
instead of every two years. Based on information shared to us by one vendor for a 35,000 barrel
per day unit, the lost production due to the shutdown is estimated to cost $150,000/day for 5
weeks, the catalyst is estimated to cost $5 million to fill the reactor, and an additional $1.5
million cost is added to represent the maintenance cost during the shutdown. To estimate the
cost for a 75,000 bbl/day unit these costs were scaled by a factor of 75,000 over 35,000. The
catalyst cost is also used for Case 2 for filling the added reactor volume.
Table 5-34 contains the results of the cost analysis for each of the three cases analyzed.
5-56
-------
Table 5-34 Cost for Different FCC Pretreater Scenarios for Complying with Tier 3
Yield Cost ($/bbl)
Catalyst and
Turnaround Cost
($/bbl FCC feed)
Hydrogen Cost ($/bbl
FCC feed)
Capital Cost ($/bbl
FCC feed)
Fixed Cost ($/bbl
FCC feed)
Total Cost FCC Feed
($/bbl)
Total Cost All
Gasoline ($/bbl)
Total Cost All
Gasoline (c/gal)
Case 1
Higher Severity
-1.1
0.47
1.2
-
-
0.59
0.27
0.65
Case 2
Revamp
-1.1
0.23
1.2
0.18
0.11
0.64
0.29
0.71
Case 3
Higher Severity at
Aromatics Peak
Saturation
0
0.47
1.2
-
-
1.7
0.77
1.8
Our analyses of the three different cases summarized in Table 5-34 shows that achieving
additional yield gain is an important factor for achieving further sulfur control at a low cost. For
the cases which assume additional yield gain, we estimate a cost of about 0.3 dollars per barrel,
or about 0.7 cents per gallon when the costs are amortized over the refinery's entire gasoline
pool. The estimated total capital cost for the revamp case is $44 million. For the case which
assumes no additional yield gain, our estimated cost is about 1.8 cents per gallon. These costs
may be overstated somewhat because we are not accounting for the further desulfurization and
cetane improvements of the light cycle oil, which is a diesel fuel blendstock. We estimate that if
this refinery were to install a grassroots FCC postreater for complying with Tier 3 instead of its
FCC pretreater, the grassroots postreater is estimated to cost 1.45 c/gal and need to invest $109
million in capital. Thus, our cost analysis estimates that if a refinery were to revamp its FCC
pretreater or simply turn up the severity of its FCC pretreater and expect to achieve additional
yield benefits in addition to lower FCC naphtha sulfur levels, then it would do so based on costs,
as opposed to installing a grassroots FCC postreater.
5-57
-------
5.1.3.5.5.2 Would Refiners use FCC Pretreaters to
Comply with Tier 3?
The decision to rely solely on FCC pretreating to comply with Tier 3 is more complicated
than simple economics. If the refinery solely relies on an FCC pretreater to comply with Tier 3,
the refinery will not gain the operational flexibility of having both an FCC pretreater and
postreater would provide (if one of those units suffers an emergency shutdown, the other unit
would allow the refiner to continue to use the FCC unit to blend up its gasoline and still comply
with the cap standard). If the refiner were to revamp its FCC pretreater, it would roughly
maintain the operational flexibility that it currently experiences under Tier 2, although it would
be somewhat lower due to the higher desulfurization that would have to be maintained and the
more stringent sulfur standard that the refiner would have to comply with. If a refiner chooses to
solely increase the severity of its FCC pretreater for complying with Tier 3, which would reduce
the cycle length of its FCC pretreater, the refiner would have lower operational flexibility since
the unit would need to shutdown more frequently.
Refiners may have sacrificed some of their operational flexibility when complying with
Tier 2 by turning up the severity of its FCC pretreater, and therefore may already be
experiencing shorter turnaround times with its FCC pretreaters than desired. If this is the case,
then when investing for Tier 3 the refiner may choose to add additional reactor volume when
revamping its pretreater to lengthen the FCC pretreater turnaround times. This additional
investment would be justified if the refiner would offset the increased capital costs by lower
turnaround costs. This same investment strategy could occur if the refiner intends on installing a
grassroots postreater. The refiner would design the postreater to take over some of the
pretreater's desulfurization duty which would allow for lengthening the turnaround times for the
FCC pretreater.
Another consideration of refiners contemplating relying solely on FCC pretreating to
comply with Tier 3 is the coke make on the FCC catalyst and how it would affect the FCC unit
operations. As shown in Table 5-32, coke make in the FCC unit decreases as the percent
desulfurization by the FCC pretreater increases. As the amount of coke which forms on FCC
catalyst decreases, there would be less heat created in the FCC unit regenerator. The lower heat
generation in the FCC regenerator can apparently be addressed by increasing the catalyst flow
between the reactor and regenerator, but if the catalyst control valve must be opened too much to
control the FCC temperature balance, an FCC control issue can develop which potentially could
cause operational problems that would force a shutdown of the FCC unit. To counter this, the
refiner may have to operate their FCC pretreater at the poly aromatics saturation maximum
temperature to preserve coke make. However, in operating in maximum poly aromatics
saturation mode, the refinery would not realize the yield gains and would face higher
desulfurization costs.
One more consideration that refiners could have about relying on FCC pretreaters for
complying with Tier 3 is the price of natural gas, which essentially determines the production
cost for producing hydrogen. FCC pretreaters demand a lot of hydrogen and low natural gas
prices support a desulfurization strategy that uses FCC pretreating. We are aware that several
US liquid natural gas (LNG) export projects have been approved to export natural gas to the rest
of the world. Although natural gas prices are projected to stay low, if refiners are concerned that
5-58
-------
increased exports of natural gas from the US could cause US natural gas prices to double or even
triple, which is roughly what the rest of the world pays for natural gas, it would significantly
increase the cost of FCC pretreating compared to FCC postreating. This could cause refiners to
choose to add an FCC postreater instead.
One more strategy that a refiner could pursue would be to convert the FCC pretreater to a
mild hydrocracker. This strategy involves a lot more investment to allow for fractionating the
product after the hydrocracking reactor. For two reasons, the potential level of hydrocracking is
often limited. First, the FCC unit requires on the order of 50 percent of its maximum capacity to
permit the unit to operate. If the FCC pretreater were to be converted over to a high conversion
hydrocracker, there may not be sufficient feed for the FCC unit. Second, many FCC pretreaters
are not sufficiently high in pressure to permit its conversion to a high conversion hydrocracker.
We believe that the increased demand for diesel fuel will already cause refiners to convert their
FCC pretreaters over to mild hydrocrackers, so we did not pursue this as part of a Tier 3 strategy,
although we did include a sensitivity case for which we estimated the compliance cost to Tier 3
assuming lower feed rates to FCC units if FCC pretreaters are converted over to mild
hydrocrackers.
Based on the low estimated cost for Case 1 and Case 2 that we evaluated for relying on
FCC pretreaters to comply with Tier 3, we could have included FCC pretreater modifications in
our cost analysis for Tier 3. For several different reasons we decided not to. First, we assessed
the use of credits under Tier 2 and found that the sulfur credits are freely traded between refining
companies. When we assumed nationwide credit trading, the ABT program effectively
eliminated most of the need to install grassroots postreaters which would be installed at these
refineries currently using FCC pretreating for Tier 2 and are thus the candidates for using FCC
pretreaters to comply with Tier 3. Thus, even if we included options for using existing FCC
pretreaters to comply with Tier 3, it may not actually be the most economical option available to
these refiners given the flexibility of the ABT program.
A second reason why we did not include in our cost estimates the revamp of existing
FCC pretreaters for complying with Tier 3 is the uncertainty of how each pretreater is operated
today and how it would operate when complying with Tier 3. We don't know the current
turnaround schedule for each refinery's FCC pretreater, nor do we know their operating pressure,
thus, we don't know how these units would operate at higher severity, nor if refiners would add
reactor volume or just increase their severity. We also don't know what the coke-make is on the
FCC catalyst at these refineries that rely on FCC pretreaters for Tier 2 and how a refiner would
operate its FCC pretreater if insufficient coke-make would be an issue when operating its FCC
unit when relying on an FCC pretreater to comply with Tier 3. Furthermore, as discussed above,
adding a postreater would provide additional compliance flexibility. For these reasons, we did
not include FCC pretreating as an option for complying with Tier 3 in our cost modeling. While
we did not model any refineries revamping their existing FCC pretreaters to comply with Tier 3,
there may nevertheless be some for whom it may be commercially advantageous. If so, this
would tend to lower the overall costs of the Tier 3 standards compared to our cost analysis
described in sections 5.1 and 5.2.
5-59
-------
5.2 Estimated Tier 3 Sulfur Control Costs
5.2.1 Final Fuel Program Cost Results
We used the refinery-by-refinery cost model to estimate the costs of the 10 ppm average
standard being adopted in this final rule. The Tier 3 fuels program maintains the 80 ppm cap
sulfur control standard that was put in place under the Tier 2 sulfur program. In general, the cost
model indicates that further desulfurizing the FCC naphtha will be the most cost-effective means
for achieving sulfur control. We accounted for additional costs to refiners for desulfurizing their
LSR naphtha, for those refineries where we estimate that the LSR naphtha is not being
desulfurized today and found that it likely needs to be. In addition to analyzing the cost of the
final Tier 3 fuels program, we also assessed the cost of other possible gasoline sulfur standards
that we considered for the proposed rulemaking. The cost information for these other gasoline
sulfur standards that we analyzed is summarized in subsection 5.2.2.
As described in Section V.D of the preamble and Chapter 4, we are also adopting an ABT
program that is designed to ease the overall burden on the industry while still achieving the 10-
ppm annual average sulfur standard for the nation as a whole. Under the ABT program,
refineries that can reduce sulfur below 10 ppm at a relatively low cost can generate credits which
can then be acquired by refiners for the refineries at which the cost of attaining the 10-ppm sulfur
standard would be higher. These credits can be traded among refineries within the same
company, or between refiners and importers nationwide. The net effect of this credit trading
would be to reduce the overall cost of the program. The extent to which the ABT provisions
reduce the cost of the Tier 3 program depends on the extent that the ABT program is used by
refiners. Since we were not sure about the extent that credits were traded under Tier 2 when we
conducted the cost analysis for the NPRM we conservatively assumed for the NPRM cost
analysis that refiners would only volume-average sulfur levels among their refineries, and not
trade credits between refining companies. However, for the final rule cost analysis we evaluated
the credit trading that was occurring under Tier 2. We found that 56% of credits were in fact
being traded between refining companies under Tier 2, with the balance being used within
refining companies. This demonstrated that credit trading was freely occurring between refining
companies, supporting the conclusion that credit trading would occur nationwide among
refineries under Tier 3. We therefore assumed nationwide credit trading for our final rule.
To estimate the impact that the ABT program could have on nationwide average fuel
costs, we began with the refinery-by-refinery costs for sulfur reductions down to either 10 ppm
or 5 ppm. We then determined the lowest cost option among three alternatives for each refinery:
1. The refinery reduces its sulfur to 10 ppm.
2. The refinery reduces its sulfur to 5 ppm and generates credits for the increment
between 10 ppm and 5 ppm.
3. The refinery does not lower sulfur, but instead relies on the purchase of credits to
comply with the 10-ppm standard.
5-60
-------
A fourth category applied to refineries whose average gasoline sulfur levels are already
below 10 ppm (their refineries don't have FCC units). All such refineries were assumed to
generate credits for the increment between 10 ppm and their current sulfur level.
Our methodology was unable to consider a fifth category where a refinery may utilize
less expensive capital and operational changes to reduce their sulfur levels partially below Tier 2
levels and rely on purchasing credits only for the remainder. Such opportunities are likely to
exist at most refineries, but such refinery specific information is not available to us. As a result,
refineries in the third category are modeled to simply remain at Tier 2 sulfur levels and incur no
capital or operating cost.
To simplify the modeling of how an ABT program might operate, we focused on the
circumstances that refineries would face in the longer term, specifically after 2020. This
approach meant that the ABT program modeling did not consider the impact on gasoline sulfur
levels of delayed compliance for small refiners and small volume refineries, nor did it consider
the generation and use of any early sulfur credits. Moreover, our ABT modeling considered only
gasoline sold for use outside of California, and only gasoline produced by domestic refineries
(not importers).
To model credit trading in our cost analysis, we first establish an estimated cost for each
refinery for reducing its gasoline sulfur down to 10 ppm and to 5 ppm. Next we ranked the
sulfur control strategies for all the refineries in order from lowest to highest sulfur control cost
per gallon of gasoline and estimated the impact of their projected sulfur control strategies on
refinery sulfur levels using only one cost (either 10 or 5 ppm) for any one refinery. The model
then follows this ranking, starting with the lowest-cost refineries, and adds refineries and their
associated control technologies one-by-one until the projected national average gasoline sulfur
level reaches 10 ppm. This modeling strategy projects the sulfur control technology that will be
used by each refinery, as well as identifies those refineries that are expected to generate credits
and those that are expected to use credits in lieu of investing in sulfur control. The sum of the
costs of the refineries expected to invest in further sulfur control provides the projected overall
cost of the program.
Based on the results of our cost analysis, we estimate that for the US refining industry to
achieve a 10-ppm average level with the full benefit of nationwide credit trading, the final sulfur
control program would cost on average 0.65 cents per gallon when it is fully phased in, assuming
that capital investments are amortized at a seven percent return on investment before taxes and
expressed in 2011 dollars. Refiners would be expected to make $2.025 billion in capital
investments to achieve this sulfur reduction. These capital investments are expected to be made
over the 6 years that the Tier 3 program is expected to be phased in, which would spread out the
capital costs to average about $330 million per year.
Our cost assessment is likely conservative. The capital cost estimate is based on vendor
data which assumes that refiners are hydrotreating full range FCC naphtha. If refiners are indeed
undercutting their FCC naphtha at many refineries (and more will be doing so in the future),
many refiners would likely not need to make any capital changes. This is because the FCC
postreaters were designed when refiners were maximizing their gasoline production and
hydrotreating full range FCC naphtha. When undercutting the FCC naphtha to the diesel pool,
5-61
-------
refiners are cutting out about 16% of the FCC naphtha volume, but more importantly, they are
cutting out about half of the sulfur. Thus, if a refiner was able to produce 30 ppm gasoline, after
fully undercutting their FCC naphtha into the diesel pool, they would likely be able to produce
15 ppm sulfur gasoline using their existing Tier 2 postreater. They could then use a more active
catalyst which likely would enable the refinery to achieve 10 ppm gasoline without any capital
changes to their FCC postreaters. If all refiners were undercutting their FCC naphtha and are
able to comply with Tier 3 without any capital additions to their FCC postreaters, the cost of the
program would decrease to about 0.4 c/gal.
Another way that our modeling could be conservative is that refiners are slowly
converting their FCC pretreaters over to mild hydrocrackers to produce more diesel fuel, which
is in higher demand. We don't know the extent that this is happening, and our current analysis
assumes that none of the FCC pretreaters have been converted over to mild hydrocrackers.
However, a cost sensitivity analysis that we conducted with our refinery model estimates that if
all the FCC pretreaters were converted over to mild hydrocrackers, costs of the Tier 3 program
would decrease to 0.55 c/gal, assuming nationwide credit trading. If we combined the cost
reduction of undercutting with the mild hydrocracking, the Tier 3 costs would be lower than
either of two cost sensitivities which were conducted independently.
We also received some comments by API and two of the peer reviewers about our octane
costs. We will not include all the discussion here about octane costs because we do so in detail
in the response to peer review comments. While we are comfortable with the octane costs that
we used, we did conduct a sensitivity at a higher octane cost ($0.5/per octane number barrel
instead of $0.3 I/octane number-barrel that we used). At the higher octane cost of $0.5/octane
number barrel, the Tier 3 sulfur control costs increases from 0.65 c/gal to 0.73 c/gal.
We also estimated annual aggregate costs, including the amortized capital costs,
associated with the new fuel standard.0 When the 10-ppm gasoline sulfur standard is fully
phased in 2020, we estimate that the sulfur standard would cost $790 million in that year. Figure
5-1 shows the distribution of refinery costs over the accumulated gasoline volume for the fully
phased in 10 ppm sulfur standard.
D The aggregate annual costs is the estimated per-gallon cost multiplied by the total projected gasoline volume in
that year.
5-62
-------
Tier 3 Refinery Sulfur Control Costs
Assumes Nationwide Trading
40 60
Percent of Total
80
100
Figure 5-1 Tier 3 Sulfur Control Costs
Figure 5-1 shows that for almost 20 percent of the gasoline pool, refineries will not incur
any cost under Tier 3, either because these refineries are already very low in sulfur because they
don't have FCC units, or because the refineries are purchasing credits.E For another 10 percent
of the gasoline pool, the refinery costs are in the 0 - 0.5 cent/gal range. For the next 55 percent
of the gasoline pool, the refinery costs are in the 0.5-1.0 c/gal range. For the last 15 percent of
the gasoline pool, the refinery costs range from 1.0 to 2.1 c/gal for revamps, with the exception
of one refinery at 2.8 c/gal representing the cost for the sole grassroots unit which our modeling
estimates would need to be installed. All other refiners that may otherwise need to install a
grassroots hydrotreater were able to do so more cheaply through the purchase of credits.
In addition to assessing the Tier 3 program costs on a societal cost basis which amortizes
the capital costs on a before-tax 7 percent ROI, we also assessed the program costs based on how
industry would assess costs to achieve a payback on the capital invested, which amortizes capital
costs on an after-tax 10 percent ROI. When the capital costs are amortized assuming an after-tax
10 percent ROI, the per-gallon costs of the final Tier 3 fuels program increases to 0.75 cents per
gallon.
Figure 5-2 summarizes our estimated US gasoline sulfur levels over the accumulated
gasoline volume post Tier 3.
E Refineries purchasing credits will incur a cost for the purchase of the credit, but since we don't know what the
price of a credit will be, we allocate all the cost for complying with Tier 3 solely on the refineries adding capital and
incurring operating costs to comply with Tier 3.
5-63
-------
Estimated Gasoline Sulfur Levels Post Tier3
(Nationwide Credit Trading)
°n
"S 70
a. /u
ttfio
01
> en
oj 5MJ
>- d.n
Ł4°
l/l :?n
01
-------
taxes and expressed in 2011 dollars. Refiners would be expected to make $2.190 billion in
capital investments to achieve this sulfur reduction.
We also estimated annual aggregate costs, including the amortized capital costs,
associated with the new fuel standard. When the 10 ppm gasoline sulfur standard is fully phased
in 2020, we estimate that the sulfur standard would cost $910 million in that year. Figure 5-3
shows the distribution of refinery costs over the accumulated gasoline volume for the fully
phased-in 10 ppm sulfur standard which assumes that sulfur credits were solely traded within
refining companies.
Alternative Sulfur Control Costs
10 ppm Average Assuming
Intracompany Credit Trading
20
40 60
Percent of Total
so
100
Figure 5-3 Alternative Sulfur Control Costs - 10 ppm Average Assuming Intracompany
Credit Trading
Figure 5-3 shows that for about 15 percent of the gasoline pool, refineries would not
incur any cost, either because these refineries are already very low in sulfur because they don't
have FCC units, or because the refineries are purchasing credits. For the next 65 percent of the
gasoline pool, the refinery costs are in the 0-1.0 c/gal range. For the last 20 percent of the
gasoline pool, the refinery costs range from 1.0 to 5 c/gal. While the average cost increased by
just 0.1 cents per gallon compared to the cost based on nationwide trading, the upper end of this
range is nearly twice that of the maximum cost for nationwide trading. This emphasizes the
benefits to the program that result from nationwide trading.
Figure 5-4 summarizes our estimated US gasoline sulfur levels over the accumulated
gasoline volume post Tier 3.
5-65
-------
Alternative Refinery Sulfur Levels
10 ppm Average Assuming
Intra-company Credit Trading
70
60
0 30
u
20
1 n
1U
o '
0
I
20 40 60 80
100
Percent of Total
Figure 5-4 Alternative US Gasoline Sulfur Levels - 10 ppm Average Assuming
Intracompany Credit Trading
Figure 5-4 shows that for about 90 percent of the gasoline pool would predominately be
either 5 or 10 ppm representing the two sulfur levels to which we assumed that refiners would
desulfurize their gasoline pool. For the rest of the gasoline pool, the refineries are clearly
purchasing credits and their sulfur levels range from 10 to nearly 70 ppm. As discussed earlier,
lacking more detailed refinery-specific information, for these refineries we assumed that they
take no action to reduce their gasoline sulfur below their Tier 2 levels. In reality these refineries
would likely to take some very cost-effective steps to partially reduce their gasoline sulfur and
not rely solely on credits to demonstrate compliance with Tier 3.
5.2.2.2 10 ppm Average Assuming No ABT program
Based on the results of our cost analysis, we estimate that for the US refining industry to
achieve a 10-ppm average level but with no benefit of credit trading, the final sulfur control
program would cost on average 0.87 cents per gallon when it is fully phased-in, and with capital
investments at a seven percent return on investment before taxes and expressed in 2011 dollars.
Refiners would be expected to make $2.990 billion in capital investments to achieve this sulfur
reduction.
We also estimated annual aggregate costs, including the amortized capital costs,
associated with the new fuel standard. When the 10 ppm gasoline sulfur standard is fully phased
in 2020, we estimate that the sulfur standard would cost $1060 million in that year. Figure 5-5
shows the distribution of refinery costs over the accumulated gasoline volume for the fully
phased in 10 ppm sulfur standard.
5-66
-------
Cost (c/gal)
Alternative Refinery Sulfur Control Costs
10 ppm Average Assuming No ABT Program
m
n
0
7
C
A
•3
•>
~^S_
,-r
^^^^^ ^^^^^
~
0 10 20 30 40 50 60 70 80 90 100
Percent of Total
Figure 5-5 Tier 3 Sulfur Control Costs
Figure 5-5 shows that for a small fraction of the gasoline pool, refineries would not incur
any cost under Tier 3, because these refineries are already very low in sulfur because they don't
have FCC units. For about 80 percent of the gasoline pool, the refinery costs are in the 0 - 1
cent/gal range. For the next 55 percent of the gasoline pool, the refinery costs are in the 0.5 - 1.0
c/gal range. For the last 20 percent of the gasoline pool, the refinery costs range from 1.0 to 8.5
c/gal. All other refiners that may otherwise need to install a grassroots hydrotreater were able to
do so more cheaply through the purchase of credits. While the average cost increased by just
0.22 cents per gallon compared to the cost based on nationwide trading, the upper end of this
range is about three times higher than that of the maximum cost for nationwide trading. This
emphasizes the benefits to the program that result from nationwide trading. This cost scenario
most closely parallels that conducted by Baker & O'Brien for API as discussed below in Section
5.3, given the stringent per-gallon cap they assumed. This serves to emphasize the cost benefits
provided by a flexible nationwide ABT program.
5.2.2.3 5 ppm Average Assuming No ABT program
Based on the results of our cost analysis, we estimate that for the US refining industry to
achieve a 10-ppm average level with the full benefit of nationwide credit trading, the final sulfur
control program would cost on average 1.28 cents per gallon when it is fully phased in, assuming
that capital investments are amortized at a seven percent return on investment before taxes and
expressed in 2011 dollars. Refiners would be expected to make $3810 million in capital
investments to achieve this sulfur reduction.
We also estimated annual aggregate costs, including the amortized capital costs,
associated with the new fuel standard. When the 10-ppm gasoline sulfur standard would be
fully phased in 2020, we estimate that the sulfur standard would cost $1555 million in that year.
5-67
-------
Figure 5-6 shows the distribution of refinery costs over the accumulated gasoline volume for the
fully phased in 10 ppm sulfur standard.
Alternative Refinery Sulfur Control Costs
5 ppm Average and No ABT Program
20
30
40 50 GO
Percent of Total
70
80
90
100
Figure 5-6 Alternative Sulfur Control Costs - 5 ppm Average and no ABT Program
Figure 5-6 shows that for a small fraction of the gasoline pool, refineries would not incur
any cost, because these refineries would already be very low in sulfur because they don't have
FCC units. For another 80 percent of the gasoline pool, the refinery costs are in the 0-1.0
cent/gal range. For the last 20 percent of the gasoline pool, the refinery costs range from 1.0 to
13 c/gal.
5.2.2.4 Estimated Cost of a 50 ppm Sulfur Cap Standard
For the NPRM, we co-proposed a 50 ppm sulfur cap along with the current 80 ppm cap
already in place under the Tier 2 regulations. In doing so, we assumed, based on feedback from
some refiners, that a 50 ppm cap would not increase the cost of compliance with Tier 3.
However, we received comments on this assumption from API and other commenters suggesting
that lowering the sulfur cap might considerably increase the cost of Tier 3 gasoline sulfur
control. To support their comments, they provided a study they commissioned by Turner, Mason
and Company (TMC) as discussed further below.
In response to the comments provided, for the FRM we estimated the cost of a potential
50 ppm refinery gate sulfur cap standard relative to the current 80 ppm cap standard. We
identified two potential cost impacts of lowering the cap standard to 50 ppm: 1) some refineries
that could maintain their sulfur levels at higher levels under Tier 3 with a 80 ppm cap standard
would have to reduce their gasoline sulfur levels under a 50 ppm cap, and 2) some refineries may
need to invest in additional tankage to store high sulfur FCC naphtha if the FCC pretreater or
postreater has operational problems or needs to shutdown.
5-68
-------
If the refinery gate cap standard were to be reduced from 80 ppm sulfur to 50 ppm, we
estimate that refineries which otherwise might avoid capital investment by purchasing credits
and continuing to produce gasoline above an average of 35 ppm under Tier 3 with an 80 ppm cap
would need to make a capital investment to avoid having batches of gasoline exceed the 50 ppm
cap on a regular basis. Refiners cannot produce gasoline which averages the same as the cap
standard because of the variability in gasoline blendstock and gasoline batch-to-batch sulfur
content (see Section 4.42). Based on our evaluation of refinery gasoline sulfur levels under the
Tier 2 program, we found that the average sulfur level for refineries is at or below 70% of the
cap standard, with the exception of only two refineries. This appears to be about the level
necessary for most refineries to operate with adequate flexibility. Seventy percent of a 50 ppm
cap would translate into a maximum average sulfur production level of 35 ppm for refineries. Of
the 40 refineries which average greater than 35 ppm under Tier 2, our refinery modeling for Tier
3 estimates that 8 refineries would continue to average above 35 ppm sulfur when the Tier 3
sulfur standard is fully phased in under the 80 ppm cap standard. For the purpose of estimating
the cost of the 50 ppm cap standard we assume that these 8 refineries would then need to make
capital investments to reduce their gasoline sulfur levels. Because our refinery modeling could
not model partial sulfur reductions in gasoline (i.e., reducing gasoline sulfur from 55 to 35 ppm),
we estimated that these 8 refineries would reduce their gasoline sulfur levels to 10 ppm. To
avoid overcomplying with Tier 3, other refineries would then not invest as much which would
maintain the industry average of 10 ppm sulfur. After modeling these changes in our refinery
model, we estimate that complying with Tier 3 with a more stringent cap of 50 ppm increases the
average cost by 0.02 c/gal and increases the overall capital costs at refineries by $135MM.
The second factor we identified that would increase refinery costs is the potential need to
add storage at refineries to store high sulfur FCC naphtha, or other high sulfur naphtha, in the
event that a refinery's gasoline hydrotreater (pretreater or postreater) were to require emergency
maintenance. This volume of storage is not intended to store all the FCC naphtha to enable a
full turnaround of the FCC pretreater or postreater. Instead, it would: 1) provide short-term
storage for a modest fix to either of these units, 2) permit the refiner to make arrangements for
storing or selling the FCC naphtha elsewhere until the inoperable hydrotreating unit could be
made operational, or 3) allow for an orderly shutdown the FCC unit. Based on data that TMC
collected on behalf of API from a survey of refiners, refiners today have on average 4.6 days
worth of storage for unhydrotreated FCC naphtha to comply with the current Tier 2 standards (an
average of 30 ppm and a cap of 80 ppm). The affect of a lower cap with Tier 3 would be slight.
As we describe in Section 5.1.3.2, a Tier 3 refinery's FCC naphtha is typically 21 ppm assuming
that FCC naphtha comprises 35% of the refinery's gasoline pool. However if a refinery's FCC
pretreater or postreater needs to be taken offline, the refinery's FCC naphtha sulfur level could
increase to hundreds or even thousands of ppm. Assuming that the refinery's FCC naphtha is
1000 ppm when the pretreater or posttreater is down, the refiner would only be able to blend less
than 1% FCC naphtha in its gasoline under either an 80 ppm or a 50 ppm cap, so the cap should
not fundamentally change the refinery's situation when the cap standard is at 80 or 50 ppm.
Either way the refiner needs to find a home for the high sulfur FCC naphtha and they are
currently doing so with an average of 4.6 days worth of storage. Nevertheless, it is possible that
some refineries may need to increase their storage capacity for unhydrotreated FCC naphtha in
response to a lower sulfur cap.
5-69
-------
We conservatively estimated potential storage need increases with a 50 ppm cap by
evaluating which refineries might experience lower ratio of cap to average sulfur level under Tier
3 compared to Tier 2, and therefore have a need for increases storage capacity. To estimate
which refineries might need more storage we used our modeling results to look at each refinery's
situation before and after Tier 3. For example, one refinery's average sulfur level is 29 ppm
under Tier 2 and our modeling assumed that this refinery would remain at 29 ppm and purchase
credits under Tier 3. Under Tier 3 the refinery's cap to average sulfur level would be 1.6
(compared to 2.6 under Tier 2). This refinery could experience a greater need for storing high
sulfur FCC naphtha batches than under Tier 2. On the other hand, if a refinery reduces its
gasoline sulfur under Tier 3 from 30 ppm to 10 or 5 ppm, the refinery's cap to average sulfur
level would be higher under Tier 3 than Tier 2 (assuming a 50 ppm cap standard) and the
refinery is less likely to need FCC naphtha storage under Tier 3 than under Tier 2. We project
that after complying with Tier 3 and with a 50 ppm refinery gate standard, 17 refineries would be
projected to experience lower ratio of cap to average sulfur level under Tier 3 than under Tier 2
To estimate the potential tankage costs, we assumed that these 17 refineries have no FCC
storage capacity today and would need to add the full 4.6 days worth. For each of the 17
refineries, we calculated the volume of FCC naphtha that may be needed to be stored based on
the maximum FCC naphtha volume capable of being produced (at maximum FCC throughput
rates) with the FCC unit operating full time over those 4.6 days. We estimated the cost of
petroleum storage to be $25/bbl for a 250,000 barrel storage tank, which is based on a cost
estimate we obtained from the Independent Fuel Terminal Association for a gasoline storage
tank. We applied a 20% contingency factor and a 20% offsite factor to estimate a total installed
cost. We adjusted this cost to reflect the cost for the appropriate-sized storage tank for each
refinery using the six-tenths rule (see equation 5-4). Based on the capital cost for adding the
storage capacity at each of these refineries, we estimated the fixed costs assuming that they
comprise 6.7% of the capital costs. The total of capital and fixed costs, averaged over the entire
US gasoline production volume is 0.02 cents per gallon, and the capital costs for this storage are
$95 million.
Combining the above cost estimates, we estimate that reducing the cap standard from 80
to 50 ppm could increase the total compliance cost for Tier 3 compliance cost by 0.04 cents per
gallon, thus, increasing Tier 3 estimated costs from 0.65 to 0.69 cents per gallon. The estimated
total capital cost increase for a 50 ppm cap standard is $230 million, thus increasing Tier 3
estimated capital costs from $2.025 billion to $2.26 billion.
In their comments on our proposal, the American Petroleum Institute provided a report
which discussed the costs and other impacts that might result from a more stringent gasoline
sulfur cap standard. The comments were in the form of a report analyzed and written by the
Turner, Mason and Company (TMC) wherein they principally report the results of a survey of
refiners that they conducted to ascertain their compliance approach and potential compliance
costs with more stringent gasoline sulfur caps of 60, 40, 30, and 20 ppm as compared to the
current 80 ppm cap. The survey was comprised of a series of multiple choice questions. While
the TMC study did not look specifically at the implications of the proposed 50 ppm cap standard,
one could look at the responses for the 40 and 60 ppm cap standard and interpolate to get a good
sense of what the study would show for a 50 ppm cap standard. Apparently, 6 refiners
responded to the survey and TMC estimated that these refiners operate 30% of the operating
5-70
-------
refineries and produce 44% of the gasoline produced in the U.S. Thus, the respondents were
primarily large refining companies which own multiple refineries. TMC then extrapolated the
results from these 6 refiners to the remaining 38 U.S. refiners to estimate the overall impact. A
summary of the API estimated cost and other impacts is summarized in Table 5-35. We
calculated a per-gallon cost by amortizing the capital costs over the volume of gasoline that API
used in its cost study assuming a before tax 7% return on investment (specified by the Office of
Management and Budget). The estimated per-gallon costs are also summarized in Table 5-35.
Table 5-35 API Estimated Cost and Supply Impacts
Capital Cost ($MM)
Per-Gallon Cost
(c/gal)
Gasoline Supply Loss
(Kbbl/day)
80
-
-
12
60
1768
0.17
43
40
3390
0.33
63
30
5533
0.54
108
20
6090
0.60
129
The TMC survey results can be placed into five different categories associated with lower
cap standards: 1) how refiners would comply with the Tier 3 average sulfur standard, 2) how
refiners would react to an outage of a FCC naphtha desulfurization unit (FCC pretreater or
postreater), 3) how refiners would react to a loss of low sulfur blendstocks, 4) what the projected
increase in capital costs would be for tighter cap standards, and 5) what the loss in gasoline
production volume would be. Many of the survey results are presented in the report as a series
of plusses (from 1 to 4) or a single minus instead of presenting the raw survey results.
Presumably, one plus indicates a modest tendency to undertake an action and increasing numbers
of plusses indicates a greater tendency to undertake an action, while a single minus sign indicates
very little or no tendency to undertake an action.
With respect to the first survey category, when complying with the 10 ppm average
standard, the TMC refiner survey found that as the cap standard was decreased refiners would
rely less and less on FCC pretreating as the sole means for desulfurizing FCC naphtha, and
therefore would rely on both combined pre and postreating (at a 20 ppm cap, the report seems to
conclude that every refinery which solely relies on FCC pretreating today to lower its gasoline
sulfur would install an FCC postreater). This survey result in TMC's study for API is
inconsistent with the cost study Baker & O'Brien performed for API. In that study, which
assumed a 20 ppm cap standard, refiners were assumed to rely on revamped FCC pretreaters for
complying with the 10 ppm average standard. In our cost analysis discussed in section 5.1.3.5,
we conservatively assumed that the only option considered by refineries which are complying
with Tier 2 by only using FCC pretreating is to put in a grassroots FCC postreater to comply with
Tier 3. Thus, our cost analysis of the 10 ppm average and 80 ppm cap standard may already
include some of the costs of a tighter cap standard. Furthermore, there are several California
refineries, already governed by a 20 ppm cap standard, that solely use FCC pretreating to comply
with the California gasoline sulfur standard. Thus, it does not appear that decreasing the cap
would necessarily force FCC postreating versus FCC pretreating.
5-71
-------
With respect to the first survey category, the survey also found that refiners are more
likely to reduce the endpoint of FCC naphtha and add treatment of other blendstocks as the cap
standard is lowered. Reducing the endpoint of gasoline is likely an important factor driving the
gasoline supply results from the TMC study. However, our LP refinery modeling projects that
refineries will already be extensively undercutting FCC naphtha when the Tier 3 program is
implemented in order to maximize distillate production and profits. Thus, we believe that this
will be occurring in the baseline with or without Tier 3.
With respect to the second survey category, (how refiners would react to an outage of a
FCC naphtha desulfurization unit with lower cap standards), the refiner survey found that
refiners would predominantly idle or significantly reduce feed rate to the FCC unit, regardless of
the stringency of the cap standard. The second most important strategy that refiners would use
would be to store high sulfur components, and that this strategy would increase as the cap
tightens. Other strategies that increased as the cap was lowered included purchasing low-sulfur
components, selling high sulfur components and/or exporting high-sulfur gasoline, reducing the
endpoint of FCC naphtha, and reducing crude oil feed rate. One other strategy identified for that
subset of refineries with both a pretreater and a postreater or multiple postreaters today was to
simply operate the nonshutdown hydrotreating unit more severely. The refiner survey results
supports the analysis that we conducted above, which is to add storage capacity to allow
additional storing of high sulfur FCC naphtha. The refiner then has time to deal with the very
high sulfur FCC naphtha stream through reprocessing, selling, blending, exporting, etc.
With respect to the third survey category (loss of low sulfur blendstocks), the refiner
survey found that refiners would take a number of actions. The most common response was to
store high sulfur components, although idling or significantly reducing the feed rate to the FCC
unit was also common. Other responses were to increase the severity of the FCC pre or
postreater, sell high sulfur components or export high sulfur gasoline. We believe that this issue
is predominantly about an emergency shutdown of the naphtha hydrotreater which facilitates the
production of reformate, isomerate, and low-sulfur, light straight run naphtha. Both the reformer
and isomerization units require near zero sulfur gasoline to function, and if the naphtha
hydrotreater undergoes an emergency shutdown, then the feedstock to the reformer and
isomerization units must be stopped and those units would need to be shutdown as well. Since
straight-run naphtha is high in sulfur, it likely cannot be blended straight to gasoline without
exceeding the sulfur cap standard. This is a situation that exists today, and due to the very high
sulfur of untreated FCC naphtha, this situation is little impacted by the level of the cap. Perhaps
in limited situations if a refinery was refining a sweet crude oil, the refinery may be able to
blend some of the heavy straight run naphtha into its gasoline pool and still meet the sulfur cap
without hydrotreating it if the cap was not too stringent. However, this would be further limited
by the need to meet the octane requirements for the gasoline. Provided that the refiner put in
emergency storage for FCC naphtha, this storage could also be used for heavy straight run
naphtha since it is unlikely that the naphtha hydrotreater would need to be shutdown at the same
time that an FCC pre or postreater would need to be shutdown. As a result, there should really
be no additional impact or cost associated with this issue that isn't already captured under survey
category 2.
With respect to the fourth survey category (estimated capital costs for lower caps), the
survey results, TMC extrapolated to the industry as a whole, estimated the impact of the cap
5-72
-------
standards would be $1.8 billion for a 60 ppm cap standard, $3.4 billion for a 40 ppm cap
standard, $5.5 billion for a 30 ppm cap standard, and $6.1 billion for a 20 ppm cap standard.
Thus, for complying with a 50 ppm cap standard, the TMC refiner survey estimates a capital cost
of about $2.5 billion, which is about one order of magnitude higher than our estimate. Because
these values are solely responses to a survey questionnaire, and no justification or support for
these values was provided in the report, it is difficult to identify the basis for the reported costs.
It is difficult to reconcile these results with our own. There is no connection made in the report
between the responses to the other survey questions discussed above, and the potential capital
costs. As discussed above, there are several other survey results which would have an impact on
capital costs that we believe are inappropriate to attribute to a declining cap standard, but the
report doesn't provide any means to assess this in the context of capital costs. Part of the
discrepancy between the survey results and our own may simply be due to confusion with the
refiner responses to the survey. The survey asked refiners what the costs would be for
complying with the various hypothetical cap standards, but the survey did not specify whether
the refiner should report their capital cost estimate on a refinery basis or company basis. Thus, it
is possible that TMC assumed that the refiner responses for a capital cost estimate were on a
refinery basis and multiplied the values by the number of refineries that each refiner has, when in
fact the refiner may have responded on a company basis. No discussion is provided to help the
reader understand how this potential confusion may have been resolved by TMC.
The last survey category discussed in the TMC report is the potential gasoline production
loss during turnarounds associated with tighter cap standards. The TMC report estimated that
the 80 ppm cap standard would already be responsible for 12 thousand barrels per day (kbbl/day)
of gasoline production loss. As the cap is lowered to 60 ppm the report estimates this gasoline
production loss increases to 43 kbbl/day, if the cap is lowered to 40 ppm the projected loss
increases to 63 kbbl/day, at a 30 ppm cap standard the loss increases to 108 kbbl/day and a 20
ppm cap standard is projected to cause a loss of 129 kbbl/day of gasoline production.
Interpolating between 60 and 40 ppm, the projected gasoline production loss for a 50 ppm cap is
53 kbbl/day. Thus, decreasing the cap standard from 80 to 50 ppm is projected by TMC to
increase the gasoline production loss by 41 kbbl/day. Of the 8452 kbbl/day of gasoline
consumed in the US, this projected gasoline production loss represents 0.5%. Similar to our
observation of estimated capital costs, there is no connection made in the report between the
responses to the other survey questions discussed above and the potential gasoline production
loss. Furthermore, as discussed above, there are several other survey results which would have
an impact on gasoline production that we believe are inappropriate to attribute to a declining cap
standard, but the report doesn't provide any means to assess this in the context of gasoline
production impacts. A portion of this projected gasoline production loss is a short term loss and
once the affected units are restarted, high sulfur intermediate refinery streams (such as FCC
naphtha) would be hydrotreated with the excess capacity available in the respective hydrotreating
units and would eventually be blended into the gasoline pool. Thus, the projected net gasoline
production loss would be much less than the 0.5% reported by TMC for a 50 ppm cap standard.
5.3 Other Cost Studies
Other cost studies were recently conducted to estimate the cost of additional reduction in
gasoline sulfur. We evaluated each of these studies and compare them to our own cost analysis.
5-73
-------
5.3.1 International Council for Clean Transportation Cost Study
The International Council for Clean Transportation (ICCT) retained Mathpro in October
2011 to study the cost of a 10 ppm average gasoline sulfur standard as well as a 1 psi reduction
in RVP.17 Since the lower RVP standard was modeled as a separate step from the low sulfur
standard, we were able to isolate the gasoline sulfur reduction costs from the low RVP costs.
ICCT's estimated cost for a 10 ppm average gasoline sulfur standard is 0.8 cents per
gallon which reflects the capital costs amortized assuming a before-tax 7 percent rate of return
on investment (to be consistent with our analysis). This cost reflects an assumption that the
capital cost for revamps of FCC postreaters is 30 percent of the capital costs for a grassroots
FCC postreater. Mathpro also analyzed costs assuming that the capital cost for revamps of FCC
postreaters are 50 percent of a grassroots FCC postreaters, which is 1.1 cents per gallon. ICCT's
cost estimate for complying with a 10 ppm average gasoline sulfur standard is very close to ours.
5.3.2 The Alliance Cost Study
In 2008 The Alliance retained Mathpro to use its LP refinery cost model to estimate the
costs of what they termed National Clean Gasoline (NCG) in PADDs 1, 2 and 3 (generally
1 &
speaking, this is the part of the U.S. east of the Rocky Mountains). Achieving NCG would
entail reducing gasoline sulfur to 5 ppm under a 10-ppm cap standard and the reduction of
gasoline RVP to 7 psi. For the low-RVP standard, a 1-psi waiver was allowed for conventional
gasoline, but not for current low-RVP areas. The study also evaluated two sensitivity cases
which increases the stringency of the distillation index (DI) from 1250 to 1200. The Alliance
study also evaluated crude oil price as a second sensitivity case, evaluating crude oil prices at
$51/bbland$125/bbl.
The Alliance studied three different cases. The first case applied the 10 ppm sulfur cap to
RFG. The second case applied the 10 ppm sulfur cap and the 7.0-psi low-RVP standard to RFG
as well as 7.0- and 7.8-psi low-RVP gasoline. The third case applies the 10 ppm sulfur cap and
7.0-psi RVP standard to all RFG and CG. Of these three cases, the first case is most relevant
because applying the fuels changes to RFG solely applies the 10 ppm sulfur cap to RFG and does
not involve any changes in RVP. However, the 10 ppm sulfur cap standard studied by the
Alliance is still 5 ppm more stringent than the 10 ppm average standard that we are proposing.
The Alliance cost estimate for Case 1 is 1.6 cents per gallon for RFG in PADDs 1, 2 and
3. This cost estimate is based on amortizing the capital costs on a 10 percent after-tax return on
investment (ROI). We adjusted the cost estimate to amortize the capital costs based on a before
tax 7 percent ROI and adjusted the costs to 2010 dollars which increases the costs to 1.75 cents
per gallon. The 1.75 eVgal cost estimate is based on a crude oil price of $51/bbl. The Alliance
estimated the cost of a 10 ppm sulfur cap standard on RFG assuming that crude oil is priced at
$125/bbl. At the $125/bbl crude oil price, the Alliance study estimates that it costs 2.50 eVgal to
require that RFG comply with a 10-ppm sulfur cap standard. Adjusting the Alliance costs to
reflect a 7 percent before tax ROI and 2010 dollars increases the Alliance costs based on a
$125/bbl crude oil price to 2.69 e7gal.
5-74
-------
For our cost analysis we analyzed the cost of sulfur control assuming that crude oil is
priced at $91.8/bbl. We can interpolate between the Alliance costs based on $51 and $125 per
barrel crude oil prices, which results in a single cost which is 2.3 cents per gallon. We also
estimated a cost for refiners lowering their gasoline sulfur to 5 ppm using the refinery-by-
refmery cost model and our cost is 1.28 eVgal.
5.3.3 API Cost Study
In response to the Alliance study, API retained Baker & O'Brien (BOB) in 2010 to study
the cost of additional gasoline sulfur control and RVP control using a refinery-by-refinery cost
approach with BOB's Prism model.19 The Prism model is largely a spreadsheet cost model with
blending optimization. The primary case analyzed by the API study is the cost of reducing
gasoline sulfur to an average of 10 ppm and reducing gasoline RVP to 7.0 psi without a 1-psi
waiver for blending 10 percent ethanol. The study also analyzed three other sensitivity cases: 1)
a 5-ppm average gasoline sulfur standard with 7 psi RVP limit on conventional gasoline without
a 1-psi waiver; 2) a 10 ppm average gasoline sulfur standard and a 7.8-psi RVP limit on
conventional gasoline without a 1-psi waiver; and 3) a 10 ppm average gasoline sulfur standard
with a 7.8-psi RVP limit on all conventional gasoline with a 1-psi waiver.
In an addendum to its fuels study report released in 2011, API contracted with Baker &
O'Brien to study a sensitivity case 4, which is a sulfur only case, using its PRISM refinery
model. From our understanding of the study, the study parameters seemed to be about the same
as the original study, except that API solely studied a 10 ppm average gasoline sulfur standard
(not including any RVP control), the same sulfur standard that we are proposing. However, API
also assumed that a 20 ppm cap standard would also be in place which would effectively not
allow the application of an averaging, banking and trading (ABT) program to optimize refinery
investments and minimize overall costs.
API made a series of conclusions based on the study. Perhaps the most important
difference with the original study is that API concluded that not a single refinery would shut
down as a result of the proposed 10 ppm gasoline sulfur control standard, even though API did
not study the flexibilities of an ABT program and used excessively high capital costs for a
grassroots FCC postreater (see below). Like the original study, API did not report average costs,
but reported only the marginal costs for the cost study. Marginal costs reflect the cost of the
program to the refinery or refineries which would incur the highest costs, assuming that the
highest cost refineries would set the price (or in this case, the price increase) of gasoline. The
report concluded that marginal costs after the imposition of a 10 ppm gasoline sulfur program
would increase the price of gasoline by 6 to 9 cents per gallon in most markets. API did not
define how its statement "in most markets" would apply to the US gasoline supply. API also did
not provide any justification why it assumed that the refineries that would experience the highest
desulfurization cost under Tier 3 would also be the same refineries which set the gasoline price
in their gasoline markets today.
Although API did not provide an average gasoline desulfurization cost in its report, we
could calculate an average cost based on the gasoline volume and total annual costs provided.
The total cost reported in the report for the 10 ppm average gasoline sulfur standard is
$2390MM/yr and the non-California gasoline volume is 7343 thousand barrels per day. This
5-75
-------
results in an average per-gallon desulfurization cost of $0.89/bbl or 2.12 c/gal. The difference
between the average cost and marginal cost (price increase) that API is projecting is profit. API
is projecting that the oil industry would profit from 10 ppm low sulfur standard by the roughly 4
to 7 cents per gallon difference between the average cost and the two marginal price values.
That per-gallon profit translates into $4 to $8 billion dollars per year in profit.
The average cost of the 10 ppm average gasoline sulfur standard was calculated using
API's methodology for amortizing capital investments. To facilitate a fairer comparison
between the API cost study and our cost study, we adjusted the API costs to be on a similar basis
as our costs. We adjusted the API costs to reflect a before-tax 7 percent return on investment
(ROI) for capital invested for the hydrotreaters and hydrogen plants instead of the after-tax 10
percent ROI used by API. This lowered the API estimated costs from 2.12 c/gal to 1.58 c/gal.
API's 1.58 cents per gallon cost is still higher than our 0.89 c/gal cost with an ABT program that
assumes intercompany trading of credits, and higher than our 0.97 c/gal for the case which
assumes no ABT program. Thus comparing "apples-to-apples" to the to the extent possible,
API's 1.58 c/gal estimated cost for complying with a 10 ppm average gasoline sulfur standard
compares very favorably with our own cost estimates, and not at all near the 25 c/gal value that
was sometimes quoted from the first API study. The remaining cost difference between our
estimated costs and those by API are the capital cost assumptions that API used, as discussed
below. While little detail is provided by API about what hardware comprises their
desulfurization units, the inside battery limits (ISBL) and total capital costs for the FCC
postreaters and FCC pretreaters are provided in API's report. API's FCC pretreaters capital
costs are consistent with the capital costs that we have used for this unit. However, the FCC
postreater costs used by API are much higher than what we used and have been used in the past
by others. API's capital cost for a grassroots FCC postreater is $228 million for a 35,000 bbl/day
unit, or $6540 per/bbl per day. API's capital cost includes the outside battery limit (OSBL)
costs. In contrast, the ISBL capital cost that we used for a grassroots FCC postreater is
$1500/bbl-day for a 30,000 bbl/day grassroots unit, which increases to $2440/bbl/day when the
offsite costs and a 20% contingency are added on. Thus, the API capital costs are more than 2 /^
times higher than the capital costs that we are using for a grassroots FCC postreater. To check
our capital costs, we found other capital cost estimates to which we could compare our costs.
Table 5-36 contains a cost comparison of ISBL and OSBL FCC postreater capital costs for a
grassroots unit.
Table 5-36 Capital Cost Comparison
Technology
ISBL Capital
Cost ($/bbl/day)
ISBL and OSBL
Capital Cost
($/bbl/day)
EPA
(Tier 3)
1500
2430
Mathpro
(ICCT)
-
1800
Jacobs
2440
3538
API
-
6540
Table 5-36 shows that, compared to the average of the rest of the capital cost estimates,
the API capital cost for FCC postreater is about 2 /^ times higher. Compared to the next highest
5-76
-------
cost estimate, which is the FCC postreater capital cost from the Jacobs data base in the Haverly
refinery cost model that we use,F the API capital costs are almost two times higher.
An important distinction must be made with respect to the severity of desulfurization for
the capital cost comparison made in 5-32. For complying with the Tier 2 gasoline sulfur
standard (Jacobs), a typical refinery would have installed an FCC postreater to desulfurize the
FCC naphtha from about 800 ppm down to about 75 ppm, a 725 ppm, or a 91 percent sulfur
reduction. In the case of a grassroots postreater that would be installed for Tier 3, the postreater
would treat FCC naphtha already low in sulfur due to the pretreater installed before the FCC unit
(these refineries are currently complying with Tier 2 using an FCC pretreater). Thus, the new
grassroots FCC postreater would only have to reduce the FCC naphtha from 100 ppm to 25 ppm,
a much smaller 75 ppm or 75 percent sulfur reduction. A grassroots FCC postreater installed for
Tier 2 would typically remove 10 times more sulfur than one installed for Tier 3. This is
important because a significant portion of the FCC postreater capital cost is devoted to avoiding
the recombination reactions which occur when hydrogen sulfide concentrations are high and
react with the olefms contained in the FCC naphtha. Thus, a grassroots FCC postreater installed
for Tier 3 would be expected to be significantly lower in capital cost compared to a Tier 2 FCC
postreater. API's costs are based solely on Tier 2 compliance costs, which is one reason why
their costs are so high. API obtained either estimated installation costs or actual installation costs
(API did not specify) for FCC postreaters for installation in 5 different refineries for complying
with the Tier 230 ppm gasoline sulfur standard. The postreater capital cost information which
reflected cost information from the years 2003 to 2005 was adjusted upward to reflect mid-year
2009 capital costs using the Nelson-Farrar index and normalized to reflect a 35 thousand barrel
per day unit. This resulted in an average ISBL cost of $144.5 million for installing a Tier 2
compliant FCC postreater. After discussing this capital cost estimate with several refiners who
built several of the units in recent years, those refiners felt that the estimated capital costs that
API had calculated were too low, and one refiner thought that the estimated capital costs should
be doubled. Based on the information provided by that one refiner, API doubled its estimated
capital costs for a 35K bbl/day FCC postreater to $228.8 million.
Another way to assess the API capital cost for the FCC postreaters is to compare it to the
FCC pretreater cost that API is using. FCC pretreaters are much higher pressure units and use
more expensive metallurgy than FCC postreaters and, for these two reasons, are much more
expensive than FCC postreaters on a per-barrel basis. However, API's FCC postreater capital
costs are about 50 percent more expensive than its FCC pretreater capital costs, which is
inconsistent with the design requirements of the units. API acknowledged this inconsistency, but
did not take steps to correct it.
API's estimated range of capital cost for revamping an FCC postreater is also higher than
our range of capital cost for revamping an FCC postreater, when assessing the revamped costs as
a percentage of the capital cost for a grassroots unit. API estimates that revamping an FCC
F The installed capital cost for an FCC postreater from the Jacobs data base was adjusted to current year dollars.
This estimated installed capital cost is several years old and may not represent Jacobs current cost estimate for a
FCC postreater.
5-77
-------
postreater would cost 30 to 70 percent of the capital cost for a grassroots FCC unit. Our capital
cost estimate for revamping FCC naphtha postreaters range from 17 to 50 percent of the capital
cost for a grassroots FCC postreater, however, most of the revamps are estimated to cost at the
lower end of that range.
As discussed above, an important reason why API's projected capital costs for complying
with Tier 3 are so high is that API assumed a 20 ppm cap standard in addition to the 10 ppm
average standard that it studied. The 20 ppm cap standard eliminates the possibility of realizing
the cost savings of an ABT program. After we proposed the Tier 3 rule, API presented to EPA,
in its comments on the proposed rulemaking, its estimate for the cost of finalizing a more
stringent cap standard. The study, which was contracted to the Turner, Mason & Company,
estimated that a 20 ppm cap standard would increase the capital cost of complying with a 10 ppm
average standard by $6.1 billion. If we subtract the $6.1 billion in capital costs attributed to the
20 ppm cap standard from the $9.8 billion in total capital costs from API's Addendum report
which estimated the cost of complying with Tier 3, and adjust the fixed operating costs
accordingly, the API estimated average cost (not marginal cost) for complying with Tier 3
decreases to 0.97 c/gal. In addition to the questionable capital costs assumed for FCC postreaters
as discussed above, this information from API on its estimated cost of complying with a 20 ppm
cap standard helps to answer an important question of why the API estimated average cost was
still higher than the other studies after other cost adjustments were made. This final adjustment
to the API costs makes the estimated API costs for complying with Tier 3 right in line with the
other cost studies. This adjusted API cost, however, still does not include the cost saving aspects
of credit averaging and trading since the API analysis assumed that each refinery meets the 10
ppm average sulfur standard. Thus, to compare this most recent cost adjustment of API costs to
our cost study, our 0.87 c/gal cost for no ABT program would be the most appropriate cost for
comparison (see section 5.2.2 for alternative costs). The adjusted API cost and our cost are only
0.1 c/gal different.
5.3.4 Emissions Control Technology Association Cost Study
The Emissions Control Technology Association (ECTA) retained personnel within
Navigant Economics to study the costs of a 10 ppm average gasoline sulfur standard and assess
the ICCT and API cost studies.20 The authors made a number of conclusions. After reviewing
both the ICCT and API studies, the authors found that a primary difference in estimated costs
between the two studies was the capital costs. The authors contacted vendor companies that
license FCC postreater technologies and surveyed the companies to find out what the capital
costs are for a FCC postreater. As a result of the survey, the report authors concluded that API's
capital costs were too high, and those used in the ICCT study were about right. The authors
found that Baker & O'Brien has a history of exaggerating the economic impacts of EPA rules,
citing the costs and other impacts of its analysis of the 2007 on-highway heavy-duty proposed
rulemaking. The authors concluded that the impact of a 10 ppm gasoline sulfur standard on the
average refining cost would likely be closer to the 1 cent per gallon estimate by the ICCT study.
Furthermore, the report's authors also pointed out that the marginal cost analysis conducted by
API did not consider the proposed averaging banking and trading (ABT) program that we were
expected to propose, which would reduce the marginal costs of the Tier 3 proposed rule.
5-78
-------
5.4 Projected Energy Impacts and Impacts on Permitting
Our refinery-by-refinery model was also used to determine the impact the Tier 3
standards would have on the energy related CO2 emissions and permitting of existing refineries.
While the Tier 3 proposal will reduce emissions from vehicles, the addition of grass roots units
and revamping of existing units which we project will happen as a result of the Tier 3 sulfur
standards are likely to result in some increased emissions of regulated air pollutants at refineries.
Refinery projects designed to meet the new fuel standards could trigger preconstruction air
permitting requirements under the Clean Air Act and EPA's New Source Review (NSR)
regulations. To address this concern, we used our refinery-by-refinery model to estimate the
likely process and equipment changes that may be required to meet the Tier 3 gasoline standards.
This information was submitted to EPA's Office of Air Quality Planning and Standards
(OAQPS) to provide the inputs that are necessary for the modeling and analysis of the refinery
emissions and permitting impacts of the Tier 3 fuel standards.
Using our refinery-by-refinery model we generated refinery-specific estimates of the
increased energy and hydrogen demands that we estimate will result from the proposed Tier 3
standards. We also estimated the increase in sulfur plant recovery unit (SRU) loading/operations
for the 110 U.S. refineries that we modeled in our analysis. Energy demand includes fuel that is
needed to generate refinery process heat, steam and electricity. Hydrogen demand is associated
with increased hydrotreating of Fluid Catalytic Cracking (FCC) naphtha and light straight run
(LSR) streams. Increased SRU loading results from the increased fuel desulfurization and
associated H2S generation. All of these incremental demands will be referred to as "demands"
in the following sections. We used our refinery-by-refinery model to calculate the increase in
these various demands for several scenarios where sulfur averaging, banking, and trading (ABT)
was not allowed and each refinery had to meet the lOppm standard, as well as scenarios that
allowed ABT between refineries owned by the same parent company to minimize the cost of
compliance with the Tier 3 standards.
5.4.1 Emissions Impacts of Different Production Volumes
In addition to considering scenarios with and without ABT we also considered the
impacts on emissions and permitting of different gasoline production volumes for each refinery.
In the first case, called the normal case, we considered the incremental demands for each refinery
assuming no change in gasoline production volume. We also considered a case, called the
maximum demand case, where each refinery maximized gasoline production based on currently
existing refinery capacity and equipment.
5.4.1.1 Normal Case
The normal case was estimated using each refinery's predicted yields of FCC naphtha
and LSR from our refinery-by-refinery model, along with each refinery's total gasoline
production volume from EPA's RFG database. For each refinery the refinery-by-refinery model
generated specific Tier 3 demands for hydrogen, steam, fuel gas and electricity based on the
desulfurization technology used by each refinery for any FCC postreating and LSR
hydrotreating. To determine the FCC postreating demands the model considers each refinery's
volume of FCC naphtha under normal operations, the FCC naphtha sulfur level at the refinery
5-79
-------
prior to postreating, and the process requirements of the FCC postreater technology used by that
refinery. The demands are calculated by multiplying the FCC naphtha volume by the demands
from the use of the associated FCC postreating technology. Table 5-20 through Table 5-23 show
the FCC postreater technology demand averages as applied to refineries on a national basis for
the 5 and 10 ppm gasoline sulfur standards. Note that the demands vary significantly with the
FCC naphtha sulfur level prior to postreating.
Similarly, the normal case demands for any LSR blendstocks that require additional
hydrotreating as a result of the Tier 3 standards were determined based on each refinery's yield
of LSR blendstock under normal operations and the demands for the additional LSR
hydrotreating. These demands are discussed in section 5.1.3.1.2. The normal case demands for
FCC postreating and LSR hydrotreating were then summed to determine the increase in energy
and hydrogen demand. To determine the additional sulfur removed from gasoline we first
calculated the difference between the current gasoline sulfur level of the gasoline produced at
each refinery according to their compliance reports to EPA and the proposed Tier 3 standard.
This difference was multiplied by the refinery's gasoline production volume and divided by the
number of days of operation to calculate the additional sulfur removal level at each refinery.
This sulfur removal information was then used to determine the increase in SRU loading on a
fractional basis by dividing the additional sulfur removal as a result of the Tier 3 standards (in
tons of sulfur per day) by the refineries SRU process capacity.
5.4.1.2 Maximum Case
We also considered a second demand case, called the maximum case, in which we
calculated the demands that result from the Tier 3 standards if each refinery maximizes gasoline
production based on currently existing refinery capacity. For this case we first determined each
refiners FCC unit process capacity utilization rate in the normal case. The annual FCC unit
feedstock charge rate for each refinery as reported in the 2011 EIA data was divided by the FCC
unit design capacity as reported in the Oil and Gas Journal (OGJ) to calculate the capacity
utilization rate for the normal case. These normal capacity utilization rates were then scaled up
to reflect maximum capacity utilization rates, and further adjusted using an overdesign factor.
For refineries projected to meet the proposed Tier 3 standards by revamping existing
FCC postreating units we assumed that their maximum gasoline production rate was equal to the
rate produced running the FCC unit at 92% of the refinery's maximum FCC design capacity.
There were several refineries that are currently operating their FCC unit greater than a 92%
capacity utilization rate. We assumed that these refineries were already operating at their
maximum annual capacity utilization rate. For the refinery projected to install a new FCC
postreater, we similarly assumed that the new unit would be scaled to process the output of the
FCC unit operating with a 92% utilization rate. For the new FCC postreater, however, we
increased the results by 15% as an overdesign factor and adjusted the results accordingly. A
similar sizing approach was taken for refineries we projected would revamp or add new LSR
hydrotreating capacity to comply with the proposed Tier 3 standards.
The results represent a "maximum" annual gasoline production case for each refinery
under the Tier 3 standards based on each refiners FCC unit design capacity. These cases
represent a scenario where each refinery's emissions based on the maximum achievable annual
5-80
-------
production rate for their existing processing units. These cases reflect each refiner's potential
emissions impacts as a result of the proposed Tier 3 standards, relative to the existing Tier 2
standards, when operating at maximum FCC rates as opposed to normal operation more
indicative of national gasoline demand.
5.4.2 Refinery Demand Sourcing
After determining the increased demands for each refinery as a result of the proposed
Tier 3 regulations we next developed cases for each refinery demand scenario that represented
different options for sourcing these demands. Some refiners may choose to produce all of the
required hydrogen and electricity. Others may choose to purchase some or all of the hydrogen
and electricity that they would need to comply with the Tier 3 standards from external suppliers.
These decisions have a significant impact on the emissions and permitting impacts of the
proposed Tier 3 regulations. In order to bound all possible scenarios we considered both high
and low impact cases for each refinery demand scenario. In the high impact scenarios we
assume that each refinery produces all of the required hydrogen, and electricity needs. In the low
impact scenarios we assume that all the necessary hydrogen and electricity are purchased from
an external supplier.
In both cases we assumed that fuel gas demands would increase to meet the increased
thermal demands at the refinery. In the high impact scenarios the refinery's fuel gas needs
would be further increased to produce the needed hydrogen and electricity. We consulted
literature sources to determine the conversion factors from MBTU fuel gas to 1,000 standard
cubic feet (scf) of hydrogen21 and 1,000 pounds of steam22 that are typical for refineries. We
also assumed a standard conversion efficiency from fuel gas to electricity for our modeling. For
hydrogen needs, in our emissions analysis, we presumed that fuel gas used to make hydrogen
would only generate C02 emissions, as we presumed that the conversion process was efficient.
Fuel gas used as process energy needs in the hydrogen production process, however, we
projected would emit the full range of emissions. All of the fuel gas demand estimates are
shown in Table 5-37.
Table 5-37 Fuel Gas Required to Produce Hydrotreater Utilities
Utility
Hydrogen Process
Energy (1,000 SCF)
Hydrogen Process Feed
Needs (1,000 SCF)
Steam (1, 000 Ibs)
Electricity (1 kWh)
Fuel Gas Required (M BTU)
140
248
1530
5.1
We assumed that refiners would not need any additional energy needs to make up for lost
octane, because our LP refinery modeling estimates that the utility demands associated with
recovering the octane debit from Tier 3 was inconsequential. Refiners are expected to use
reformers and other high octane processes, such as alkylation units, to cover any octane debit in
FCC naphtha, by making minor shifts in operations which apparently does not impact utility
5-81
-------
demands. Also, in some cases lost octane could be recovered by purchasing high octane
blendstocks which would result in no fuel gas or electricity demand increases at the refinery.
5.4.3 Refinery Energy Demand Impacts
Using our refinery-by-refinery model, along with the technology vendor data for new and
revamped FCC postreating and LSR hydrotreating data (shown in Table 5-25 through Table 5-28
and Table 5-31), we were able to calculate the increases in refinery energy demands as a result of
our proposed Tier 3 regulations for each of the various scenarios outlined in the previous
sections. This information is summarized in Table 5-38 below. The refineries have been
identified by randomly assigned numbers to protect confidential business information (CBI).
This information was submitted to our sister office, OAQPS, to serve as the basis for
their emissions and permitting analysis of the Tier 3 regulations. Based on the updated refinery
analysis, EPA determined that under the final Tier 3 gasoline sulfur standard, which includes
national ABT, 9 refineries are projected to trigger New Source Review (NSR). This equates to
approximately 8% of the 108 refineries. Of these 9 refineries, only 3 are projected to trigger
permitting for both a NAAQS-related pollutants and GHGs while an additional 6 refineries are
projected to require PSD permits addressing only GHG emissions. A technical memorandum
describing the OAQPS analysis and results is in the public docket for this final rule.23
5-82
-------
Table 5-38 Tier 3 Maximum Refinery Energy, Hydrogen and Sulfur Plant Demand
Increases for ABT Case
Refinery
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Demand Estimates
Fuel
Gas Demands
(Million BTU/Yr)
Low Case
55,766
494,692
0
50,404
0
125,277
140,263
1,152,149
0
498,770
0
512,053
32,553
181,259
High Case
239,291
579,106
0
125,517
0
344,410
215,618
1,403,863
0
1,088,552
0
618,649
113,349
259,280
High Case Plus
CO2 for
Hydrogen
555,121
690,241
0
252,794
0
561,788
345,295
1,743,848
0
1,989,386
0
766,760
210,888
370,586
Sulfur Plant Production
Sulfur
Production
Increase
(Tons
Sulfur/Day)
0.50
0.44
0.00
0.04
0.00
0.88
0.30
0.86
0.00
0.08
0.00
0.68
0.25
0.45
Sulfur Plant
Capacity
Increase
(Percent of
Existing
Facility)
0.07
0.04
0.00
0.08
0.00
No Data
0.09
0.12
0.00
0.06
0.00
0.18
0.21
0.10
Hydrogen
Hydrogen
Demand
Increase
(million
scf/year)
1529
462
0
501
0
867
524
1686
0
3638
0
595
394
450
5-83
-------
Refinery
Number
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Demand Estimates
Fuel
Gas Demands
(Million BTU/Yr)
0
184,692
0
592,842
579,988
2,169,093
0
29,522
127,534
53,024
0
0
19,972
1,040,992
0
0
0
279,786
0
464,226
721,733
2,506,050
0
115,563
358,933
129,662
0
0
81,398
1,309,416
0
0
0
443,434
0
485,068
918,750
2,863,334
0
169,530
594,542
259,460
0
0
123,448
1,716,105
0
0
Sulfur Plant Production
Sulfur
0.00
0.05
0.00
0.27
0.26
0.78
0.00
0.56
2.20
0.41
0.00
0.00
0.25
0.30
0.00
0.00
Sulfur Plant
0.00
0.03
0.00
0.50
0.26
0.09
0.00
0.04
0.00
0.28
0.00
0.00
0.11
0.06
0.00
0.00
Hydrogen
Hydrogen
0
661
0
88
811
1441
0
217
0
536
0
0
167
1633
0
0
5-84
-------
Refinery
Number
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
Demand Estimates
Fuel
Gas Demands
(Million BTU/Yr)
0
0
0
24,527
0
0
0
50,152
76,080
90,268
25,536
22,695
724,372
597,780
0
21,128
0
0
0
64,717
0
0
200,906
158,367
142,703
53,123
71,902
929,979
803,621
0
32,402
0
0
0
102,608
0
0
396,393
202,974
232,938
99,593
127,645
1,225,219
1,157,854
0
51,803
Sulfur Plant Production
Sulfur
0.00
0.00
0.00
0.10
0.00
0.00
0.00
0.88
0.28
0.12
0.01
0.27
0.40
0.60
0.00
0.07
Sulfur Plant
0.00
0.00
0.00
0.06
0.00
0.00
0.00
0.41
0.07
0.09
0.05
0.22
0.04
0.04
0.00
0.69
Hydrogen
Hydrogen
0
0
0
153
0
0
786
180
364
188
186
1187
1307
0
78
5-85
-------
Refinery
Number
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
Demand Estimates
Fuel
Gas Demands
(Million BTU/Yr)
1,421,831
99,075
0
43,263
0
33,128
0
0
0
1,433,373
124,488
0
868,124
62,323
0
634,584
1,530,970
113,173
0
81,700
0
118,783
0
0
0
1,661,638
152,511
0
1,123,388
101,795
0
760,395
1,643,335
122,179
0
146,144
0
223,948
0
0
0
1,955,563
192,146
0
1,492,456
169,723
0
933,119
Sulfur Plant Production
Sulfur
0.47
0.56
0.00
0.31
0.00
0.31
0.00
0.00
0.00
0.77
0.16
0.00
0.59
0.14
0.00
0.34
Sulfur Plant
0.12
0.19
0.04
0.03
0.00
0.09
0.00
0.00
0.00
0.04
No Data
0.00
0.09
0.70
0.00
0.14
Hydrogen
Hydrogen
454
0
0
249
0
425
0
0
0
1172
165
0
1366
331
0
605
5-86
-------
Refinery
Number
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
Demand Estimates
Fuel
Gas Demands
(Million BTU/Yr)
0
177,667
666,474
0
276,071
173,066
0
0
358,243
18,780
396,525
71,768
96,393
0
0
0
0
249,774
766,007
0
504,649
242,782
0
0
508,962
78,287
470,283
210,552
185,026
0
0
0
0
359,493
891,308
0
834,842
362,756
0
0
768,333
180,693
569,854
357,881
325,957
0
0
0
Sulfur Plant Production
Sulfur
0.00
0.27
0.52
0.00
0.37
0.37
0.00
0.00
0.63
0.13
0.42
0.18
0.24
0.00
0.00
0.00
Sulfur Plant
0.00
0.63
0.05
0.00
0.09
0.91
0.00
0.00
0.07
0.10
0.26
0.01
0.05
0.04
0.00
0.00
Hydrogen
Hydrogen
0
443
502
0
1170
435
0
0
976
0
400
595
297
0
0
0
5-87
-------
Refinery
Number
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
Demand Estimates
Fuel
Gas Demands
(Million BTU/Yr)
581,482
34,842
70,820
0
144,344
30,605
0
48,840
0
0
454,267
0
221,368
48,475
0
60,159
734,018
87,838
90,336
0
489,552
51,089
0
109,034
0
0
689,362
0
386,207
98,562
0
130,128
956,397
177,666
119,036
0
899,581
86,339
0
210,697
0
0
1,093,938
0
669,879
182,847
0
248,169
Sulfur Plant Production
Sulfur
0.73
0.19
0.15
0.00
0.56
0.11
0.00
0.36
0.00
0.00
0.23
0.00
0.34
0.18
0.00
0.28
Sulfur Plant
0.09
0.06
0.09
0.00
0.07
0.90
0.00
0.03
0.00
0.00
0.06
0.00
No Data
0.02
0.00
0.15
Hydrogen
Hydrogen
895
663
113
0
1653
142
0
381
0
0
1708
0
1001
310
0
476
5-S
-------
Refinery
Number
95
96
97
98
99
100
101
102
103
104
105
106
107
108
Demand Estimates
Fuel
Gas Demands
(Million BTU/Yr)
22,121
31,063
0
0
64,964
16,472
145,657
0
0
37,609
57,350
96,148
23,486
18,721
71,369
182,172
0
0
99,342
33,132
151,780
0
0
122,658
100,794
164,991
86,923
63,065
127,917
442,216
0
0
158,505
61,154
158,099
0
0
221,067
173,298
283,464
195,167
115,508
Sulfur Plant Production
Sulfur
0.14
0.39
0.00
0.00
0.09
0.08
0.16
0.00
0.00
0.17
0.15
0.17
0.19
0.10
Sulfur Plant
0.22
0.73
0.00
0.00
0.05
No Data
4.32
0.00
0.00
0.13
0.01
No Data
0.16
0.21
Hydrogen
Hydrogen
237
822
0
0
238
119
25
0
0
397
0
403
437
212
" The refinery did not have published information on the capacity of the existing sulfur plant.
5-89
-------
5.5 Fuel Quality Requirements for Denatured Fuel Ethanol
5.5.1 Costs for Producers and Importers of Denatured Fuel Ethanol
The State of California's has a long-standing 10 ppm sulfur requirement for denatured
fuel ethanol (DFE). Do to logistical issues associated with maintaining a separate DFE pool that
complies with California requirements, all DFE manufacturers currently produce California
compliant DFE. Therefore, we expect that there will be only minimal additional costs to DFE
producers as a result of the Tier 3 program associated with new recordkeeping and product
transfer document requirements as discussed in the Information Collection Request (ICR) for the
Tier 3 rule.
5.5.2 Costs for Producers and Distributors of Ethanol Denaturants
Suppliers of denaturants for use in manufacturing denatured fuel ethanol (DFE) are
already accustomed to providing product with a sulfur content that is consistent with a 10 ppm
sulfur cap for DFE as discussed in Chapter 5.4.1. Denaturant manufacturers will be required to
maintain per batch test records on the denaturants they produce to demonstrate that the sulfur
content does not exceed 200 ppm. As discussed in Section V of the preamble to the Tier 3 final
rule, that addition of denaturant with sulfur content of 200 ppm would result in DFE with a sulfur
content of 10 ppm when blended at 5 volume percent into neat ethanol. As is current practice
today, we anticipate that ethanol manufacturers will negotiate what specific sulfur level they
require from denaturant manufacturers to facilitate with compliance with the 10 ppm sulfur cap
for DFE considering what level of compliance margin a given manufacturer feels is necessary.
We understand that ethanol manufacturers often currently require denaturant manufacturers to
provide a product with a sulfur content of 120 ppm or less in order to ensure that DFE that
contains 5 volume percent denaturant can comply with California's 10 ppm sulfur cap for DFE.
Therefore, we expect that there will be no additional processing costs to manufacture denaturant
for use in the manufacture of DFE meting the Tier 3 program 10 ppm sulfur cap.
As discussed in the Information Collection Request (ICR) for the Tier 3 rule, there will
only be minimal additional costs to ethanol denaturant producers and importers to comply with
the Tier 3 requirements associated with the registration, maintenance of denaturant testing
records, and generation of product transfer documents (PTDs). There will also be minimal
additional costs for distributors of ethanol denaturants associated with maintaining ethanol
denaturant product transfer documents.
5.6 Gasoline Additives
The Tier 3 rule requires that manufacturers of gasoline additives used downstream of the
refinery at less than 1 volume percent must limit the sulfur contribution to the finished gasoline
from the use of their additive to less than 3 ppm when the additive is used at the maximum
recommended treatment rate. All current gasoline additives contribute less than 3 ppm to the
sulfur content of the finished fuel when used at the maximum recommended treatment rate (with
3 ppm being the extreme). Normal additive production quality control practices already have
had to consider the sulfur contribution of the additive to finished gasoline as a result of the Tier 2
gasoline sulfur program. Therefore, the Tier 3 requirements will not necessitate the
5-90
-------
reformulation of any gasoline additives or additional additive production quality testing. The
maximum recommended treatment rate is already stated on product transfer document or
packaging for the additive. Hence, no changes will be needed to the product transfer documents
for gasoline additives. The only additional burden for additive manufacturers will be to retain
production quality control records for 5 years and make these available to EPA upon request.
Therefore, the Tier 3 program requirements will not result in significant additional costs to
gasoline additive manufacturers.
5.7 Downstream Pentane Blending
The Tier 3 program provisions that will facilitate the blending of pentane into previously
certified gasoline at terminals will provide additional flexibility to industry. The associated
requirements are described in Section VIA. of the preamble to the Tier 3 final rule. We expect
that the cost savings in reduced gasoline costs from downstream pentane blending will
substantially outweigh the associated compliance costs. Industry will only take advantage of the
pentane blending provisions, and be subject to the associated compliance costs, the extent that
there is a substantial cost motive to do so. Therefore, we are not assessing any costs associated
with the Tier 3 pentane blending provisions. In addition, pentane blending will likely be used in
place of the butane blending already allowed. Consequently, there would be no new increase in
the amount of testing and recordkeeping.
5-91
-------
Appendix A
Linear Program Refinery Modeling Tables
5-92
-------
Table 5-39 - Summary of LP Refinery Modeling Ouput Information and the Resulting
Cost of Octane Used in the Refinery-by-Refinery Cost Model
(Cost of Recovering 1 Octane Number Loss in FCC Naphtha, no El 5 in Gasoline Pool and Low
NGL Prices)
Control Case
FCC naphtha volume
(thousand bbls)
Objective Function
^ous^nd_do\\ars)____
Change in Obj Funct
Summer
Winter
Annual
Reference Case
Control Case
US Total
374463
425601
800,064
-118884419
-118634411
-250.008
PADD1
32827
__53282
PADD2
93577
_jje>j34
PADD3
221108
_J21J3249
PADD 4/5
26951
__37936
Cost 0.312 $/octane-bbl
0.744 c/octane-gal
Table 5-40 - Summary of LP Refinery Model Estimated Regular and Premium Marginal
Production Costs and the Calculated Cost of Octane - Not Used in Refinery Cost Analysis
(No El5 in Gasoline Pool and Low NGL Prices)
Summer
Winter
Annual Average
US
PADD1
PADD 2
PADD 3
PADD 4/5 OC
PADD1
PADD 2
PADD 3
PADD 4/5 OC
PADD1
PADD 2
PADD 3
PADD 4/5 OC
Octane Number
(R+M)/2
93
93
93
91.5
93
93
93
91.5
Marginal Price
Premium
264.36
261.71
262.31
259.72
248.17
221.55
249.58
243.63
Regular
254.27
253.49
252.23
251 .99
243.06
218.09
244.47
240.07
Cost Difference
c/gal
10.09
8.22
10.08
7.73
5.11
3.46
5.11
3.56
$/bbl
4.2378
3.4524
4.2336
3.2466
2.1462
1.4532
2.1462
1.4952
3.192
2.4528
3.1899
2.3709
$/ON-bbl
0.532
0.4088
0.53165
0.526867
0.495089
c/ON-gal
1.266667
0.973333
1.265833
1.254444
1.178783
5-93
-------
Table 5-41 - PADDs 1 and 2 - Modeled Refinery Purchase Volumes for Reference and Low
Octane LP Refinery Modeling Cases (thousand bbl/day)
PADD Crude
GIL Naphtha
GTL Diesel
VGO HS
VGOLS
HS AR (A960)
LS AR (Alg)
Normal Butane
Isobutane
MTBE
Ethanol-E10
Ethanol -E15
Ethanol - E85
Reformer Feed
Methanol
Natural Gas (FOE)
Hydrogen (MSCF)
Pentanes Plus
Import CBOB 10% Reg
Import CBOB 10% Prem
Import RBOB 10% Reg
Import RBOB 10% Prem
Import Alkylate
Import Raffinate
Import Reformate
Import FCC Naphtha
Import Lt Naphtha
Import Hvy Naph
Transfer Lt Naphtha
Transfer Reformate
Transfer Alkylate
Transfer FCC Naphtha
Transfer Raffinate
Transfer CBOB 10% Reg
Transfer CBOB 10% Prem
Transfer RBOB 10% Reg
Transfer RBOB 10% Prem
Isooctane
Isooctene
PADD1
Summer
Ref
1,058
0
0
0
115
0
47
7
12
0
311
0
21
0
0
73
0
0
0
0
0
0
0
25
0
0
0
81
0
0
0
0
0
0
0
0
0
0
0
Control
1,022
0
0
0
170
0
66
3
12
0
310
0
21
0
0
74
0
0
4
0
1
0
0
25
0
0
0
53
22
68
23
0
0
1,508
197
260
104
0
0
Dif.
-36
0
0
0
54
0
19
-4
0
0
0
0
0
0
0
1
0
0
4
0
1
0
0
0
0
0
0
-28
22
68
23
0
0
1,508
197
260
104
0
0
Winter
Ref
1,728
0
0
0
76
0
0
166
16
0
295
0
20
0
0
100
0
0
0
0
0
0
0
50
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Control
1,728
0
0
0
76
0
0
165
16
0
295
0
20
0
0
99
0
0
2
0
2
0
0
50
0
0
0
0
0
0
0
0
0
1,254
144
275
87
0
0
Dif.
0
0
0
0
0
0
0
-1
0
0
0
0
0
0
0
-1
0
0
2
0
2
0
0
0
0
0
0
0
0
0
0
0
0
1,254
144
275
87
0
0
PADD 2
Summer
Ref
2,903
0
0
0
90
0
0
23
57
0
238
0
16
0
0
201
0
238
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Control
2,924
0
0
0
90
0
0
31
53
0
238
0
16
0
0
198
0
214
6
0
0
0
0
0
0
0
0
0
22
0
0
29
88
116
0
0
0
0
0
Dif.
21
0
0
0
0
0
0
8
-4
0
0
0
0
0
0
-3
0
-24
6
0
0
0
0
0
0
0
0
0
22
0
0
29
88
116
0
0
0
0
0
Winter
Ref
4,245
0
0
0
1
0
0
200
56
0
231
0
15
0
0
219
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Control
4,246
0
0
0
1
0
0
200
56
0
231
0
15
0
0
220
0
0
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Dif.
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
5-94
-------
Table 5-42 - PADDs 3, 4 and 5 non-California - Modeled Refinery Purchase Volumes for
Reference and Low Octane LP Refinery Modeling Cases (thousand bbl/day)
PADD Crude
GTL Naphtha
GTL Diesel
VGOHS
VGOLS
HS AR (A960)
LS AR (Alg)
Normal Butane
Isobutane
MTBE
Ethanol -E10
Ethanol -E15
Ethanol - E85
Reformer Feed
Methanol
Natural Gas (FOE)
Hydrogen (MSCF)
Pentanes Plus
Import CBOB 10% Reg
Import CBOB 10% Prem
Import RBOB 10% Reg
Import RBOB 10% Prem
Import Alkylate
Import Raffinate
Import Reformate
Import FCC Naphtha
Import Lt Naphtha
Import Hvy Naph
Transfer Lt Naphtha
Transfer Reformate
Transfer Alkylate
Transfer FCC Naphtha
Transfer Raffinate
Transfer CBOB 10% Reg
Transfer CBOB 10% Prem
Transfer RBOB 10% Reg
Transfer RBOB 10% Prem
Isooctane
Isooctene
PADDS
Summer
Ref
7,408
0
0
0
0
0
750
37
0
0
142
0
10
0
0
627
0
250
0
0
0
0
0
0
0
0
0
100
0
0
0
0
0
0
0
0
0
0
0
Control
7,409
0
0
0
0
0
750
34
0
0
142
0
10
0
0
626
0
250
0
0
0
0
0
0
0
0
0
100
0
0
0
0
0
0
0
0
0
0
0
Dif.
1
0
0
0
0
0
0
-2
0
0
0
0
0
0
0
-1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Winter
Ref
6,368
0
0
0
0
0
750
216
0
0
137
0
g
0
0
607
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Control
6,377
0
0
0
0
0
750
216
0
0
137
0
9
0
0
606
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Dif.
c
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
PADD 4/5 nonCA
Summer
Ref
1,437
0
0
0
0
0
0
24
6
0
80
0
5
0
0
96
0
60
2
0
0
0
0
10
0
0
0
0
0
0
0
0
0
50
0
0
0
0
0
Control
1,425
0
0
0
0
0
0
24
7
0
80
0
5
0
0
95
0
60
2
0
0
0
0
10
0
0
0
0
0
0
0
0
0
60
0
0
0
0
0
Dif.
-12
0
0
0
0
0
0
0
1
0
0
0
0
0
0
-1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
11
0
0
0
0
0
Winter
Ref
1,452
0
0
0
0
0
0
77
3
0
76
0
C
0
0
100
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Control
1,441
0
0
0
0
0
0
75
3
0
76
0
C
0
0
99
0
13
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Dif.
-12
0
0
0
0
0
0
-2
1
0
0
0
0
0
0
-2
0
13
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
US Total
Summer
Ref
12,806
0
0
0
205
0
797
90
75
0
771
0
51
0
0
996
0
548
2
0
0
0
0
35
0
0
0
181
0
0
0
0
0
50
0
0
0
0
0
Control
12,781
0
0
0
260
0
816
92
72
0
770
0
51
0
0
992
0
524
12
0
1
0
0
35
0
0
0
153
44
68
23
29
88
1684
197
260
104
0
0
Dif.
-25
0
0
0
54
0
19
2
-3
0
0
0
0
0
0
-4
0
-24
10
0
1
0
0
0
0
0
0
-28
44
68
23
29
88
' 1634
' 197
' 260
' 104
0
0
Winter
Ref
13,794
0
0
0
77
0
750
659
75
0
739
0
49
0
0
1026
0
0
1
0
0
0
0
50
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Control
13,792
0
0
0
77
0
750
656
75
0
739
0
49
0
0
1024
0
13
7
0
2
0
0
50
0
0
0
0
0
0
0
0
0
1254
144
275
87
0
0
Dif.
-2
0
0
0
0
0
0
-3
1
0
0
0
0
0
0
-2
0
13
6
0
2
0
0
0
0
0
0
0
0
0
0
0
0
1254
144
275
87
0
0
5-95
-------
Table 5-43 - PADDs 1 and 2 - Modeled Refinery Sale Volumes for Reference and Low
Octane Cases (thousand bbl/day)
Propane
Propylene
Normal Butane
Isobutane
PC Naphtha
PC Gasoil
CG Reg
CG Prem
CGEIOReg
CG E10 Prem
RFGEIOReg
RFC E10 Prem
CG E15 Reg
RFC E1 5 Reg
E85 to CG
E85 to RFC
Transfer CBOB 10% Reg
Transfer CBOB 10% Prem
Transfer RBOB 10% Reg
Transfer RBOB 10% Prem
Export CBOB 10% Reg
Export RBOB 10% Reg
Jet/Kero A (450ppm)
X-Fer Diesel Rundown to St
HSDGr76(0.2%)
LSDGr74(.05%)
ULSD(15ppm)
CARB Diesel
X-Fer C5's to Storage
1 % Residual Fuel
Residual Fuel
Slurry
Asphalt & Wax
Gasoil
Lubes
Benzene
Toluene
Xylenes
Cumene
Cyclohexane
Transfer Raffinate
Transfer Alkylate
Transfer Reformate
Transfer FCC naphtha
Transfer Lt Naphtha
Transfer Blendstock
Sulfur (STons)
Coke (STon)
PADD1
Summer
Ref
13
25
0
0
25
0
0
0
1,669
219
1,072
115
0
0
17
11
0
0
0
0
0
0
70
0
0
0
469
0
0
0
30
0
118
0
19
2
0
0
0
0
0
0
0
0
0
0
5
6
Control
14
25
0
0
25
0
0
0
1,669
217
1,072
115
0
0
17
11
0
0
0
0
0
0
70
0
0
0
469
0
0
0
30
0
118
0
19
2
0
0
0
0
0
0
0
0
0
0
5
5
Dif.
2
0
0
0
0
0
0
0
0
-2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-1
Winter
Ref
21
25
0
0
16
0
0
0
1,562
168
1,057
139
0
0
16
11
0
0
0
0
0
0
70
0
0
0
820
0
0
0
50
40
73
0
19
2
0
0
0
0
0
0
0
0
0
0
8
42
Control
22
25
0
0
16
0
0
0
1,562
168
1,057
139
0
0
16
11
0
0
0
0
0
0
70
0
0
0
820
0
0
0
50
40
73
0
19
2
0
0
0
0
0
0
0
0
0
0
8
42
Dif.
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
PADD2
Summer
Ref
49
40
0
0
50
0
0
0
1,785
227
302
38
0
0
18
3
0
0
0
0
0
0
189
0
0
0
767
0
0
0
39
80
210
10
8
19
0
0
0
0
0
0
0
0
0
0
20
41
Control
50
40
0
0
50
0
0
0
1,785
227
302
38
0
0
18
3
0
0
0
0
0
0
189
0
0
0
752
0
0
0
39
78
210
10
8
19
0
0
0
0
0
0
0
0
0
0
21
41
Dif.
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-15
0
0
0
0
-2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-1
Winter
Ref
58
40
0
0
40
1,031
0
0
1,721
219
305
39
0
0
17
3
0
0
0
0
0
0
185
0
0
0
680
0
0
0
39
97
210
8
8
22
0
0
0
0
0
0
0
0
0
0
26
95
Control
58
40
0
0
40
1,031
0
0
1,721
219
305
39
0
0
17
3
0
0
0
0
0
0
185
0
0
0
680
0
0
0
39
97
210
8
8
22
0
0
0
0
0
0
0
0
0
0
26
95
Dif.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
5-96
-------
Table 5-44 - PADDs 3, 4 and 5 non-California - Modeled Refinery Sale Volumes for
Reference and Low Octane Cases (thousand bbl/day)
Propane
Propylene
Normal Butane
Isobutane
PC Naphtha
PC Gasoil
CGReg
CG Prem
CGEIOReg
CG ElOPrem
RFGEIOReg
RFG ElOPrem
CGE15Reg
RFG E 15 Reg
E85 to CG
E85 to RFG
Transfer CBOB 10% Reg
Transfer CBOB 10% Prem
Transfer RBOB 10% Reg
Transfer RBOB 10% Prem
Export CBOB 10% Reg
Export RBOB 10% Reg
Jet/Kero A (450ppm)
X-Fer Diesel Rundown to S
HSDGr76(0.2%)
LSD Gr 74 (.05%)
ULSD(15ppm)
CARS Diesel
X-Fer C5's to Storage
1% Residual Fuel
Residual Fuel
Slurry
Asphalt & Wax
Gasoil
Lubes
Benzene
Toluene
Xylenes
Cumene
Cyclohexane
Transfer Raffinate
Transfer Alkylate
Transfer Reformate
Transfer FCC naphtha
Transfer Lt Naphtha
Transfer Blendstock
Sulfur (STons)
Coke (STon)
PADDS
Summer
Ref
101
250
0
4
550
158
0
0
968
74
334
33
0
0
10
3
1,718
198
259
104
7
1
895
0
0
0
2,119
0
0
0
125
157
250
0
158
51
34
8
0
0
90
120
63
0
5
0
64
237
Control
103
250
0
0
550
158
0
0
968
74
334
33
0
0
10
3
1,684
197
260
104
12
1
895
0
0
0
2,135
0
0
0
125
157
250
0
158
51
34
8
0
0
89
83
68
29
22
0
64
237
Dif.
2
0
0
-4
0
0
0
0
0
0
0
0
0
0
0
0
-33
-2
1
-1
4
0
0
0
0
0
15
0
0
0
0
0
0
0
0
0
0
0
0
0
-1
-37
5
29
17
0
0
0
Winter
Ref
69
250
0
0
433
158
0
0
930
71
325
31
0
0
g
•3
1,251
144
242
120
7
2
819
0
0
0
2,091
0
0
0
125
102
250
0
158
50
34
8
0
0
1
1
0
0
0
0
55
173
Control
70
250
0
0
433
158
0
0
930
71
325
32
0
0
9
3
1,254
144
275
87
7
2
819
0
0
0
2,091
0
0
0
125
102
250
0
158
50
35
8
0
0
1
1
0
0
0
0
55
174
Dif.
1
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
4
0
34
-34
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
PADD 4/5 nonCA
Summer
Ref
17
15
0
0
0
0
0
0
686
107
0
0
0
0
7
0
0
0
0
0
0
0
233
0
0
0
518
0
0
0
39
20
30
0
0
0
0
0
0
0
0
0
0
0
0
0
10
44
Control
17
15
0
0
0
0
0
0
686
107
0
0
0
0
7
0
0
0
0
0
0
0
233
0
0
0
517
0
0
0
37
20
30
0
0
0
0
0
0
0
0
0
0
0
0
0
10
43
Dif.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-1
0
0
0
-1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Winter
Ref
17
15
0
0
0
0
0
0
650
101
0
0
0
0
6
0
0
0
0
0
0
0
236
0
0
0
513
0
0
0
33
20
30
0
0
0
0
0
0
0
0
0
0
0
0
0
11
44
Control
17
15
0
0
0
0
0
0
650
101
0
0
0
0
6
0
0
0
0
0
0
0
236
0
0
0
513
0
0
0
32
20
30
0
0
0
0
0
0
0
0
0
0
0
0
0
11
43
Dif.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
US Total
Summer
Ref
179
330
0
4
625
158
0
0
5108
627
1708
187
0
0
51
17
1718
198
259
104
7
1
1387
0
0
0
3874
0
0
0
233
257
608
10
184
72
34
8
0
0
90
120
63
0
5
0
99
328
Control
184
330
0
0
625
158
0
0
5108
625
1708
187
0
0
51
17
1684
197
260
104
12
1
1387
0
0
0
3873
0
0
0
231
256
608
10
184
72
34
8
0
0
89
83
68
29
22
0
100
326
Dif.
4
0
0
-4
0
0
0
0
0
-2
0
0
0
0
0
0
-33
-2
1
-1
4
0
0
0
0
0
0
0
0
0
-1
-2
0
0
0
0
0
0
0
0
-1
-37
' K
29
17
0
1
-2
Winter
Ref
164
330
0
0
489
1189
0
0
4862
559
1687
208
0
0
49
17
1251
144
242
120
7
2
1310
0
0
0
4104
0
0
0
247
259
563
8
184
74
34
8
0
0
1
1
0
0
0
0
99
354
Control
167
330
0
0
489
1189
0
0
4862
559
1687
210
0
0
49
17
1254
144
275
87
7
2
1310
0
0
0
4104
0
0
0
246
260
563
8
184
74
35
8
0
0
1
1
0
0
0
0
99
354
Dif.
2
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
4
0
34
-34
0
0
0
0
0
0
0
0
0
0
-1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
5-97
-------
Table 5-45 - PADD 1 Modeled Refinery Unit Throughput Volumes for Reference and Low
Octane Cases (Thousand bbl/day)
PADD1
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Isooctane
DHT- Total
D HT 2nd RCT- Total
DHT Arom Saturation
NHT - Total Fd
CGH - Generic
CGH - Olefin Sat'n
FCCU Fd HOT
LSR Splitter
LSR Bz Saturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant- Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2012
1,250
560
0
0
0
494
0
44
0
82
0
0
239
0
0
18
9
49
34
0
6
0
585
0
0
371
211
211
44
0
0
0
0
33
0
0
113
58
1
0
0
4
9
0
0
0
2018 Refcase
Summer
1,058
454
52
123
0
467
257
44
0
25
0
3
102
0
11
16
9
44
51
0
1
0
454
416
0
175
152
0
40
49
0
6
19
5
0
0
493
53
1
0
0
4
9
0
0
0
Winter
1,728
731
64
151
0
550
298
49
0
172
0
21
124
0
11
16
0
44
65
0
6
0
807
746
0
316
152
0
40
87
0
1
3
15
0
0
695
53
1
0
0
4
9
0
0
0
Control Case Minus 1
ON
Summer
1,022
442
51
142
0
535
296
44
0
21
0
3
101
0
11
16
9
44
58
0
6
0
454
416
0
170
152
0
40
47
0
6
19
5
0
0
519
53
1
0
0
4
9
0
0
0
Winter
1,728
731
66
151
0
550
298
49
0
172
0
21
162
0
11
16
0
44
66
0
6
0
807
746
0
316
152
0
40
87
0
1
3
15
0
0
673
53
1
0
0
4
9
0
0
0
Control Case Relative
to Ref Case
Summer
-36
-11
0
19
0
68
39
0
0
-4
0
0
-1
0
0
0
0
0
7
0
5
0
0
0
0
-5
0
0
0
-2
0
0
0
0
0
0
25
0
0
0
0
0
0
0
0
0
Winter
0
0
2
0
0
0
0
0
0
0
0
0
39
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-23
0
0
0
0
0
0
0
0
0
5-98
-------
Table 5-46 - PADD 2 Modeled Refinery Unit Throughput Volumes for Reference and Low
Octane Cases (Thousand bbl/day)
PADD 2
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Isooctane
DHT- Total
D HT 2nd RCT- Total
DHT Arom Saturation
NHT - Total Fd
CGH - Generic
CGH - Olefin Sat'n
FCCU Fd HOT
LSR Splitter
LSR Bz Saturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant- Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2012
4,005
1,707
0
0
0
1,314
0
307
0
508
0
0
859
0
0
28
178
175
98
3
5
0
1,526
0
6
1,324
511
511
576
0
0
0
0
54
0
0
956
25
8
0
0
43
80
0
0
0
2018 Refcase
Summer
2,903
1,229
138
301
0
1,066
616
276
0
168
0
20
428
0
100
25
153
157
104
0
0
0
789
588
0
620
318
0
518
57
0
1
3
54
0
0
1,170
23
4
0
0
37
80
0
0
0
Winter
4,245
1,840
144
330
0
1,134
649
276
0
388
0
47
550
0
127
25
118
150
104
3
8
0
633
441
0
965
232
0
518
85
0
1
3
54
0
0
1,171
23
5
0
0
43
80
0
0
0
Control Case Minus 1
ON
Summer
2,924
1,221
141
296
0
1,047
603
276
0
166
0
20
483
0
100
25
156
157
97
0
0
0
770
572
0
677
312
0
518
59
0
1
3
54
0
0
1,108
23
4
0
0
37
80
0
0
0
Winter
4,246
1,841
144
330
0
1,134
649
276
0
388
0
47
550
0
127
25
150
157
97
3
9
0
634
441
0
966
232
0
518
86
0
1
3
54
0
0
1,176
23
5
0
0
43
80
0
0
0
Control Case Relative
to RefCase
Summer
21
-8
3
-6
0
-19
-12
0
0
-3
0
0
55
0
0
0
2
0
-6
0
0
0
-19
-16
0
57
-6
0
0
3
0
0
0
0
0
0
-62
0
0
0
0
0
0
0
0
0
Winter
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
32
7
-6
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
5
0
0
0
0
0
0
0
0
0
5-99
-------
Table 5-47 - PADD 3 Modeled Throughput Volumes for Refinery Units for Reference and
Low Octane Cases (Thousand bbl/day)
PADD 3
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Isooctane
DHT- Total
D HT 2nd RCT- Total
DHT Arom Saturation
NHT - Total Fd
CGH - Generic
CGH - Olefin Sat'n
FCCU Fd HOT
LSR Splitter
LSR Bz Saturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant- Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2012
9,567
4,778
0
0
0
3,120
0
1,029
0
1,604
51
0
1,665
0
0
95
323
297
311
19
22
0
4,339
0
4
2,677
1,613
1,613
1,822
0
0
0
0
433
28
3
2,004
530
27
0
0
93
204
0
0
0
2018 Refcase
Summer
7,408
3,382
424
843
0
2,736
1,515
1,037
110
859
46
104
909
0
256
86
0
267
280
0
0
0
2,230
1,643
0
1,078
690
0
1,640
1
0
10
30
389
0
0
4,287
440
13
0
0
93
204
0
0
0
Winter
6,368
2,882
366
736
0
2,089
1,109
1,037
110
629
46
77
725
0
256
70
0
190
280
0
0
0
2,107
1,643
0
888
396
0
1,640
1
0
1
3
389
0
0
4,301
440
11
0
0
93
204
0
0
0
Control Case Minus 1
ON
Summer
7,409
3,383
428
848
0
2,736
1,511
1,037
110
859
46
104
945
0
256
86
13
267
287
0
0
0
2,246
1,659
0
1,080
673
0
1,640
1
0
10
30
389
0
0
4,282
440
13
0
0
93
204
0
0
0
Winter
6,377
2,885
369
738
0
2,097
1,114
1,037
110
631
46
77
750
0
256
71
0
186
287
0
0
0
2,106
1,643
0
909
398
0
1,640
1
0
1
3
389
0
0
4,286
440
11
0
0
93
204
0
0
0
Control Case Relative
to RefCase
Summer
1
1
4
5
0
0
-4
0
0
0
0
0
36
0
0
0
13
0
7
0
0
0
17
16
0
2
-18
0
0
0
0
0
0
0
0
0
-5
0
0
0
0
0
0
0
0
0
Winter
9
4
3
3
0
9
4
0
0
2
0
0
25
0
0
1
0
-4
7
0
0
0
0
0
0
20
1
0
0
0
0
0
0
0
0
0
-15
0
0
0
0
0
0
0
0
0
5-100
-------
Table 5-48 - PADD 4 and 5 non-California Modeled Throughput Volumes for Refinery
Units for Reference and Low Octane Cases (Thousand bbl/day)
PADD 4 and 5 nonCA
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Isooctane
DHT- Total
D HT 2nd RCT- Total
DHT Arom Saturation
NHT - Total Fd
CGH - Generic
CGH - Olefin Sat'n
FCCU Fd HOT
LSR Splitter
LSR Bz Saturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant- Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2012
1,791
637
0
0
0
368
0
128
0
173
56
0
0
273
0
26
45
39
47
1
10
0
584
0
0
391
132
132
121
0
0
0
0
58
0
0
518
0
2
0
0
0
0
0
0
0
2018 Refcase
Summer
1,437
647
64
104
0
376
146
130
0
154
0
18
92
0
0
23
41
39
48
0
0
0
557
404
0
297
107
0
109
100
0
1
3
58
0
0
679
0
2
0
0
0
0
0
0
0
Winter
1,452
650
68
105
0
376
147
146
0
155
0
18
95
0
0
23
0
39
48
0
0
0
559
404
0
300
107
0
109
100
0
10
30
58
0
0
722
0
2
0
0
0
0
0
0
0
Control Case Minus 1
ON
Summer
1,425
642
62
103
0
374
143
130
0
152
0
18
98
0
0
23
41
39
47
0
0
0
551
398
0
295
107
0
109
100
0
1
3
58
0
0
665
0
2
0
0
0
0
0
0
0
Winter
1,441
645
67
104
0
374
145
146
0
153
0
18
106
0
0
23
0
39
47
0
0
0
553
398
0
297
107
0
109
100
0
8
24
58
0
0
704
0
2
0
0
0
0
0
0
0
Control Case Relative
to Ref Case
Summer
-12
-5
-1
-1
0
-3
-2
0
0
-2
0
0
6
0
0
0
0
0
-1
0
0
0
-6
-6
0
-3
0
0
0
0
0
0
0
0
0
0
-14
0
0
0
0
0
0
0
0
0
Winter
-12
-5
-1
-1
0
-3
-2
0
0
-2
0
0
11
0
0
0
0
0
-1
0
0
0
-6
-6
0
-3
0
0
0
0
0
-2
-6
0
0
0
-17
0
0
0
0
0
0
0
0
0
5-101
-------
Table 5-49 - PADDs 1-4 and 5 non-California Modeled Throughput Volumes for Refinery
Units for Reference and Low Octane Cases (Thousand bbl/day)
PADDs 1 -4 and 5 nonCA
Refinery Units
Crude Tower
Vacuum Tower
Sats Gas Plant
Unsats Gas Plant
FCC DeC5 Tower
FCC
FCC Splitter
Hydrocracker
H-Oil Unit
Delayed Coker
Visbreaker
Thermal Naphtha Splitter
CRU Reformer
SRU Reformer
BTX Reformer
C4 Isomerization
C5/C6 Isomerization
HF Alkylation
H2SO4 Alkylation
Dimersol
Cat Poly
Isooctane
DHT- Total
D HT 2nd RCT- Total
DHT Arom Saturation
NHT - Total Fd
CGH - Generic
CGH - Olefin Sat'n
FCCU Fd HOT
LSR Splitter
LSR Bz Saturator
Reformate Saturator
Reformate Splitter
SDA
MTBE
TAME
Hydrogen Plant- Total MSCF
Lube Unit
Sulfur Plant
Merox Jet
Merox Diesel
BTX Reformer - Tower feed
BTX Reformer - Extract feed
Toluene Dealkyation
Cumene
Cyclohexane
Actual
Capacity
in 2012
16,613
7,683
0
0
0
5,296
0
1,508
0
2,366
107
0
2,763
273
0
168
556
560
490
23
43
0
7,034
0
10
4,763
2,467
2,467
2,564
0
0
0
0
577
28
3
3,591
614
38
0
0
140
293
0
0
0
2018 Refcase
Summer
12,806
5,712
677
1,371
0
4,646
2,533
1,488
110
1,207
46
145
1,532
0
367
151
203
508
483
0
2
0
4,030
3,052
0
2,170
1,267
0
2,308
206
0
18
55
506
0
0
6,630
515
20
0
0
135
293
0
0
0
Winter
13,794
6,103
642
1,321
0
4,149
2,203
1,508
110
1,344
46
162
1,494
0
394
135
118
424
497
4
14
0
4,106
3,234
0
2,469
887
0
2,308
273
0
13
39
516
0
0
6,889
515
20
0
0
140
293
0
0
0
Control Case Minus 1
ON
Summer
12,781
5,688
682
1,389
0
4,692
2,554
1,488
110
1,198
46
144
1,627
0
368
151
218
508
490
0
6
0
4,021
3,045
0
2,222
1,244
0
2,308
207
0
18
55
506
0
0
6,574
515
20
0
0
135
293
0
0
0
Winter
13,792
6,102
647
1,323
0
4,155
2,206
1,508
110
1,344
46
162
1,568
0
394
136
150
426
498
4
15
0
4,100
3,228
0
2,487
888
0
2,308
274
0
11
33
516
0
0
6,839
515
20
0
0
140
293
0
0
0
Control Case Relative
to Ref Case
Summer
-25
-24
5
18
0
46
21
0
0
-9
0
-1
95
0
0
0
15
0
7
0
5
0
-9
-6
0
52
-24
0
0
1
0
0
0
0
0
0
-56
0
0
0
0
0
0
0
0
0
Winter
-2
-1
4
2
0
6
2
0
0
0
0
0
75
0
0
1
32
3
0
0
1
0
-6
-6
0
18
2
0
0
1
0
-2
-6
0
0
0
-50
0
0
0
0
0
0
0
0
0
5-102
-------
Table 5-50 - PADD 1 - Gasoline Qualities Estimated by LP Refinery Modeling for
Reference and Octane Recovery Cases
PADDl
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur(ppm)
% at 200
% at 300
RVP(psi)
T10(F)
T50(F)
T90(F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volume (Kbbl/day)
2018 Reference Case
Summer
CG
5.00
259
26
57
83
9.1
125
186
328
1067
132
21.8
0.61
10.8
11.5
4.27
1,864,595
RFC
5.00
261
26
54
86
7.0
139
192
316
1093
144
18.7
0.59
8.7
11.6
4.29
1,241,809
Pool
5.00
260
26
56
84
8.3
131
188
323
1077
137
20.6
0.60
9.9
11.5
4.28
3,106,404
Winter
CG
4.92
255
15
60
87
13.0
100
179
310
991
110
17.6
0.61
10.7
11.5
4.34
1,745,324
RFC
4.87
254
18
62
90
13.0
101
176
300
974
110
16.1
0.63
9.6
11.5
4.35
1,181,174
Pool
4.90
255
16
61
88
13.0
101
178
306
984
110
17.0
0.62
10.2
11.5
4.34
2,926,498
2018 minus ION in FCC naphtha
Summer
CG
5.01
260
25
56
83
9.1
125
187
329
1072
132
23.2
0.61
10.1
11.5
4.27
1,864,595
RFC
4.99
261
25
54
86
7.0
139
191
316
1092
144
18.9
0.61
8.9
11.6
4.29
1,241,809
Pool
5.00
260
25
56
84
8.3
131
189
324
1080
137
21.5
0.61
9.6
11.5
4.28
3,106,404
Winter
CG
4.92
255
15
60
87
13.0
100
179
310
991
110
17.6
0.61
10.7
11.5
4.34
1,745,324
RFC
4.87
254
18
62
90
13.0
101
176
300
974
110
16.1
0.63
9.6
11.5
4.35
1,181,174
Pool
4.90
255
16
61
88
13.0
101
178
306
984
110
17.0
0.62
10.2
11.5
4.34
2,926,498
Table 5-51 - PADD 2 - Gasoline Qualities Estimated by LP Refinery Modeling for
Reference and Octane Recovery Cases
PADD 2
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur(ppm)
% at 200
% at 300
RVP(psi)
T10(F)
T50(F)
T90(F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volume (Kbbl/day)
2018 Reference Case
Summer
CG
4.93
259
26
60
86
9.2
125
178
317
1033
130
21.5
0.64
8.1
11.6
4.33
2,068,931
RFC
4.99
260
14
57
85
7.0
139
186
321
1081
143
20.6
0.68
4.8
11.6
4.31
349,978
Pool
4.94
259
24
60
86
8.9
127
180
318
1040
132
21.4
0.64
7.6
11.6
4.32
2,418,909
Winter
CG
4.86
257
23
62
87
12.7
103
176
311
986
110
19.9
0.55
9.2
11.6
4.35
1,959,358
RFC
4.87
257
26
62
88
12.7
102
175
309
982
110
18.7
0.67
10.6
11.6
4.35
346,825
Pool
4.86
257
23
62
87
12.7
103
176
311
985
110
19.8
0.57
9.4
11.6
4.35
2,306,183
2018 minus ION in FCC naphtha
Summer
CG
4.93
259
26
61
86
9.2
125
178
317
1032
130
21.6
0.61
8.5
11.6
4.33
2,068,931
RFC
4.99
259
12
57
85
7.0
139
186
321
1079
143
20.3
0.67
4.9
11.6
4.32
349,978
Pool
4.94
259
24
60
86
8.9
127
179
318
1039
132
21.4
0.62
7.9
11.6
4.33
2,418,909
Winter
CG
4.86
257
23
62
87
12.7
103
176
311
986
110
19.9
0.55
9.2
11.6
4.35
1,959,358
RFC
4.87
257
26
62
88
12.7
102
175
309
982
110
18.7
0.67
10.6
11.6
4.35
346,825
Pool
4.86
257
23
62
87
12.7
103
176
311
985
110
19.8
0.57
9.4
11.6
4.35
2,306,183
5-103
-------
Table 5-52 - PADD 3 - Gasoline Qualities Estimated by LP Refinery Modeling for
Reference and Octane Recovery Cases
PADD 3
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur(ppm)
% at 200
% at 300
RVP(psi)
T10(F)
T50(F)
T90(F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volume (Kbbl/day)
2018 Reference Case
Summer
CG
5.01
259
26
58
83
9.0
125
183
330
1061
132
22.2
0.61
10.8
11.5
4.27
1,051,761
RFC
5.07
260
11
55
83
7.0
139
190
330
1100
143
19.3
0.51
4.4
11.6
4.32
378,078
Pool
5.02
260
22
57
83
8.5
129
185
330
1072
135
21.5
0.59
9.1
11.5
4.28
1,429,839
Winter
CG
4.98
257
15
59
85
11.1
112
181
318
1025
120
19.0
0.62
11.2
11.5
4.30
1,010,516
RFC
5.05
257
9
52
80
11.2
112
195
341
1087
121
18.0
0.36
1.7
11.5
4.31
352,882
Pool
5.00
257
14
57
84
11.1
112
185
324
1041
120
18.7
0.55
8.7
11.5
4.30
1,363,398
2018 minus ION in FCC naphtha
Summer
CG
5.02
260
25
58
83
9.0
126
184
331
1065
132
23.5
0.62
10.2
11.5
4.26
1,051,761
RFC
5.03
260
20
57
84
7.0
139
186
326
1086
143
19.8
0.67
6.6
11.6
4.32
378,078
Pool
5.02
260
24
58
83
8.5
129
184
330
1071
135
22.5
0.63
9.2
11.5
4.27
1,429,839
Winter
CG
4.98
257
15
59
85
11.1
112
181
318
1025
120
19.0
0.62
11.2
11.5
4.30
1,010,516
RFC
5.05
257
9
52
80
11.2
112
195
341
1087
121
18.0
0.36
1.7
11.5
4.31
352,882
Pool
5.00
257
14
57
84
11.1
112
185
324
1041
120
18.7
0.55
8.7
11.5
4.30
1,363,398
Table 5-53 - PADD 4 and 5 non-California - Gasoline Qualities Estimated by LP Refinery
Modeling for Reference and Octane Recovery Cases
PADD 4/5 nonCA
Energy (MMBTU/bbl)
Density (Ib/bbl)
Sulfur(ppm)
% at 200
% at 300
RVP(psi)
T10(F)
T50(F)
T90(F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins (vol%)
Alcohol (vol%)
Oxygen (wt%)
Volume (Kbbl/day)
2018 Reference Case
Summer
CG
4.90
259
22
65
93
9.3
124
170
286
978
129
16.1
1.16
7.5
11.6
4.33
801,247
RFC
0.00
0
0
0
0
0.0
0
0
0
0
0
0.0
0.00
0.0
0.0
0.00
0
Pool
4.90
259
22
65
93
9.3
124
170
286
978
129
16.1
1.16
7.5
11.6
4.33
801,247
Winter
CG
4.87
256
28
67
93
12.6
102
165
282
926
110
14.4
0.43
7.7
11.6
4.37
758,461
RFC
0.00
0
0
0
0
0.0
0
0
0
0
0
0.0
0.00
0.0
0.0
0.00
0
Pool
4.87
256
28
67
93
12.6
102
165
282
926
110
14.4
0.43
7.7
11.6
4.37
758,461
2018 minus ION in FCC naphtha
Summer
CG
4.90
259
20
65
93
9.3
123
170
286
978
129
15.9
1.11
7.6
11.6
4.33
801,247
RFC
0.00
0
0
0
0
0.0
0
0
0
0
0
0.0
0.00
0.0
0.0
0.00
0
Pool
4.90
259
20
65
93
9.3
123
170
286
978
129
15.9
1.11
7.6
11.6
4.33
801,247
Winter
CG
4.87
256
28
67
93
12.6
102
165
282
926
110
14.4
0.43
7.7
11.6
4.37
758,461
RFC
0.00
0
0
0
0
0.0
0
0
0
0
0
0.0
0.00
0.0
0.0
0.00
0
Pool
4.87
256
28
67
93
12.6
102
165
282
926
110
14.4
0.43
7.7
11.6
4.37
758,461
5-104
-------
Table 5-54 - PADD 1 Gasoline Qualities Estimated by LP Refinery Modeling for Reference,
E10 and E15 Cases
Gasoline Qualities and
Volume for PADD 1
Energy (MMBTU/bbl]
Density (Ib/bbl)
Sulfur (ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins |yol%]
Alcohol |yol%]
Oxygen (wt%)
Volumes (kbbl/day)
CG
4.90
252.86
28.00
62.51
95.61
8.40
128.15
174.28
272.12
987.19
135.42
10.11
0.72
19.67
0.00
0.00
21,942
Summer
RFG
5.06
260.41
23.67
53.59
85.52
7.00
139.22
192.49
317.99
1097.88
144.26
18.99
0.58
8.00
10.00
3.70
1,111,857
20051
Pool
5.10
259.87
23.47
48.71
84.11
7.76
134.25
202.45
324.41
1126.11
141.40
23.77
0.67
14.36
4.16
1.54
2,671,322
n2030
CG
4.99
253.26
2.91
45.27
80.77
13.95
98.33
209.47
339.58
1104.02
109.71
24.31
0.92
4.67
0.00
0.00
160,860
Winter
RFG
4.96
255.50
14.71
61.47
88.89
12.96
101.74
176.41
302.70
977.92
109.71
18.93
0.59
10.28
10.00
3.77
1,121,901
Pool
5.02
255.32
14.35
54.10
85.91
12.38
105.33
191.45
316.22
1041.25
114.91
22.55
0.65
13.74
4.36
1.64
2,574,564
CG
5.07
258.49
10.57
56.09
85.30
9.40
123.48
187.38
318.99
1059.49
129.82
20.75
0.52
10.02
10.00
3.73
1,681,574
Summer
RFG
5.05
261.05
23.33
54.08
86.18
7.00
138.75
191.49
315.01
1092.27
144.30
18.79
0.58
11.98
10.00
3.69
1,134,711
E
Pool
5.06
259.52
15.71
55.28
85.65
8.43
129.63
189.04
317.39
1072.69
135.65
19.96
0.54
10.81
10.00
3.71
2,816,285
0
CG
5.02
255.89
2.49
55.40
83.29
13.17
99.94
188.80
328.14
1036.80
109.71
20.23
0.54
7.51
10.00
3.77
1,504,982
Winter
RFG
4.94
256.26
15.98
62.22
89.53
12.96
101.49
174.87
299.79
970.86
109.71
19.34
0.59
14.52
10.00
3.76
1,116,727
Pool
4.98
256.05
8.24
58.30
85.94
13.08
100.60
182.86
316.06
1008.71
109.71
19.85
0.56
10.50
10.00
3.76
2,621,709
CG
4.99
260.93
20.99
63.01
86.70
8.40
129.78
173.26
312.62
1020.79
133.09
20.25
0.76
11.33
15.00
5.54
1,754,900
Summer
RFG
4.92
259.32
27.59
62.44
87.83
7.00
138.64
174.42
307.52
1034.48
141.91
14.09
0.90
6.33
15.00
5.58
1,229,826
E
Pool
4.96
260.27
23.71
62.77
87.17
7.82
133.43
173.73
310.52
1026.43
136.72
17.71
0.81
9.27
15.00
5.56
2,984,726
5
CG
4.92
256.64
25.86
64.96
86.82
12.68
102.09
169.28
312.10
968.09
109.71
17.13
0.50
10.23
15.00
5.63
1,570,606
Winter
RFG
4.89
255.49
25.40
65.14
87.94
12.85
102.18
168.91
307.01
962.01
109.71
14.46
0.45
5.57
15.00
5.66
1,165,422
5-105
-------
Table 5-55 - PADD 2 Gasoline Qualities Estimated by LP Refinery Modeling for Reference,
E10 and E15 Cases
Gasoline Qualities and
Volume for PADD 2
Energy (MMBTU/bbl]
Density (Ib/bbl)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F]
Driveability
Vapor Lock
Aromatics (vol%)
Olefins |yol%]
Alcohol |yol%]
Volumes (kbbl/day]
CG
4.91
256.73
25 00
60.99
89.09
9 45
122.71
177.38
301.79
1012.88
128.45
17.74
054
8.92
9.59
3 60
1,396,446
Summer
RFG
5.02
260.54
19 56
54.39
85.63
700
138.63
190.84
317.48
1091.45
143.42
20.00
058
7.09
10.00
3 70
259,727
20051
Pool
4.93
257.33
24 15
59.95
88.54
9 07
125.21
179.49
304.25
1025.20
130.80
18.09
055
8.63
9.65
3 62
1,656,173
n2030
CG
5.02
256.45
25 31
49.26
82.83
13 31
99.77
201.32
330.22
1073.68
110.29
27.75
051
11.80
0.00
000
1,327,820
Winter
RFG
4.75
252.44
2484
71.10
94.01
12 27
104.71
156.74
279.39
902.57
110.29
14.77
058
9.37
10.00
3 82
257,387
Pool
4.97
255.79
25 24
52.81
84.65
13 14
100.57
194.08
321.97
1045.90
110.29
25.65
052
11.40
1.62
062
1,585,207
CG
4.92
257.04
2487
60.45
88.51
9 49
122.52
178.48
304.39
1018.23
128.34
18.28
052
8.89
10.00
3 75
1,357,780
Summer
RFG
5.02
260.49
21 47
54.40
85.71
700
138.57
190.82
317.12
1091.08
143.45
20.00
058
7.39
10.00
3 70
269,081
E
Pool
4.94
257.61
2431
59.45
88.05
9 08
125.18
180.52
306.50
1030.28
130.84
18.56
053
8.64
10.00
3 74
1,626,861
0
CG
4.89
258.44
23 87
57.80
85.60
12 92
101.33
183.88
317.64
1014.05
110.29
21.33
051
11.07
10.00
3 73
1,375,643
Winter
RFG
4.85
256.94
22 21
60.38
87.22
12 79
102.02
178.63
310.25
992.49
110.29
19.82
055
10.30
10.00
3 75
266,657
Pool
4.89
258.19
23 60
58.22
85.86
12 89
101.44
183.03
316.44
1010.55
110.29
21.08
051
10.94
10.00
3 73
1,642,300
CG
4.92
259.31
2800
64.60
87.92
860
128.27
170.00
307.11
1003.99
131.67
18.08
055
8.37
15.00
5 58
1,509,824
Summer
RFG
4.91
259.23
2800
64.60
89.19
700
138.10
170.00
301.30
1014.21
140.71
15.28
075
7.31
15.00
5 58
280,814
E
Pool
4.92
259.30
2800
64.60
88.12
835
129.81
170.00
306.20
1005.59
133.09
17.64
058
8.20
15.00
5 58
1,790,638
5
CG
4.87
260.25
2800
63.37
84.75
12 54
103.86
172.52
321.48
987.23
110.29
22.09
064
11.03
15.00
5 56
1,435,628
Winter
RFG
4.77
256.17
25 11
70.07
89.69
12 28
104.79
158.84
299.07
927.90
110.29
16.01
077
11.17
15.00
5 64
278,286
5-106
-------
Table 5-56 - PADD 3 Gasoline Qualities Estimated by LP Refinery Modeling for Reference,
E10 and E15 Cases
Gasoline Qualities and
Volume for PADD 3
Energy (MMBTU/bbl]
Density (Ib/bbl)
Sulfur (ppm)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F)
Driveability
Vapor Lock
Aromatics (vol%)
Benzene (vol%)
Olefins |yol%]
Alcohol |yol%]
Oxygen (wt%)
Volumes (kbbl/day)
CG
5.11
258.93
16.95
49.43
83.78
8.69
128.38
200.97
325.92
1113.37
135.62
25.49
0.65
13.68
3.92
1.46
2,528,524
Summer
RFG
5.11
259.59
15.87
50.15
82.02
7.00
139.25
199.51
333.93
1132.80
144.29
20.00
0.47
1.02
10.00
3.71
378,805
20051
Pool
5.09
258.38
9.56
54.62
84.24
8.66
128.72
190.37
323.80
1079.32
133.75
21.81
0.51
4.34
10.00
3.72
1,369,806
n2030
CG
5.09
255.87
10.97
49.84
83.31
11.56
110.21
200.14
328.03
1085.66
120.07
24.00
0.61
12.32
3.18
1.20
2,378,941
Winter
RFG
5.14
256.01
3.52
54.67
83.96
11.26
112.14
190.27
325.07
1056.08
120.07
18.98
0.44
2.95
10.00
3.77
385,854
Pool
5.10
255.89
9.93
50.51
83.40
11.52
110.48
198.76
327.61
1081.53
120.07
23.30
0.58
11.01
4.13
1.56
1,472,992
CG
5.08
258.77
10.43
55.79
85.08
9.30
124.17
188.00
320.01
1063.17
130.38
21.19
0.53
9.94
10.00
3.73
960,985
Summer
RFG
5.08
260.62
22.45
51.98
83.50
7.00
138.64
195.78
327.19
1115.94
144.24
19.79
0.58
8.68
10.00
3.70
392,446
E
Pool
5.08
259.30
13.92
54.68
84.62
8.63
128.37
190.25
322.09
1078.47
134.40
20.78
0.54
9.58
10.00
3.72
1,353,431
0
CG
5.07
257.06
2.76
53.98
83.18
11.32
111.20
191.68
328.64
1063.44
120.07
19.75
0.48
7.88
10.00
3.75
1,126,292
Winter
RFG
5.03
256.33
3.63
57.96
86.25
11.20
111.55
183.56
314.67
1027.57
120.07
18.28
0.58
11.95
10.00
3.76
384,074
Pool
5.06
256.88
2.98
54.99
83.96
11.29
111.29
189.61
325.08
1054.32
120.07
19.38
0.50
8.91
10.00
3.75
1,510,366
CG
5.00
261.39
19.78
61.63
85.50
8.30
130.62
176.08
318.09
1035.20
133.83
21.42
0.77
10.15
15.00
5.53
1,097,387
Summer
RFG
5.06
260.17
2.97
58.43
84.06
7.00
139.16
182.61
324.65
1073.07
141.55
18.70
0.56
1.52
15.00
5.56
409,559
E
Pool
5.02
261.06
15.21
60.76
85.11
7.95
132.94
177.85
319.87
1045.49
135.93
20.68
0.71
7.81
15.00
5.54
1,506,946
5
CG
4.98
258.99
26.77
61.19
84.40
10.93
113.41
176.97
323.09
1018.21
120.07
19.00
0.53
10.03
15.00
5.58
1,175,404
Winter
RFG
4.96
257.57
23.50
58.11
82.11
11.02
113.18
183.26
333.51
1045.60
120.07
17.29
0.52
1.29
15.00
5.61
400,822
5-107
-------
Table 5-57 - PADD 4 and 5 non-California - Gasoline Qualities Estimated by LP Refinery
Modeling for Reference, E10 and E15 Cases
Gasoline Qualities and
Volume for PADDs4&
S nonCA
Energy (MMBTU/bbl]
Density (Ib/bbl)
% at 200
% at 300
RVP (psi)
T10 (F)
T50 (F)
T90 (F]
Driveability
Aromatics (vol%)
Olefins (vol%)
Alcohol |yol%]
Volumes (kbbl/day]
CG
5.14
256.88
9 72
57.05
95 31
8.60
130.48
185 42
273.51
1015.94
133 90
22.25
1 30
790
0.00
000
544,754
Summer
RFG
0.00
0.00
000
0.00
000
0.00
0.00
000
0.00
0.00
000
0.00
000
000
0.00
000
0
20051
Pool
5.14
256.88
9 72
57.05
95 31
8.60
130.48
185 42
273.51
1015.94
133 90
22.25
1 30
790
0.00
000
544,754
n2030
CG
5.01
253.37
1069
59.49
95 16
12.79
102.91
18044
274.16
961.79
11058
20.98
094
788
0.00
000
578,168
Winter
RFG
0.00
0.00
000
0.00
000
0.00
0.00
000
0.00
0.00
000
0.00
000
000
0.00
000
0
Pool
5.01
253.37
1069
59.49
95 16
12.79
102.91
18044
274.16
961.79
11058
20.98
094
788
0.00
000
578,168
CG
5.03
258.60
9 89
62.45
93 21
9.49
122.47
17440
283.04
985.09
12789
18.78
034
636
10.00
3 73
529,671
Summer
RFG
0.00
0.00
000
0.00
000
0.00
0.00
000
0.00
0.00
000
0.00
000
000
0.00
000
0
E
Pool
5.03
258.60
9 89
62.45
93 21
9.49
122.47
17440
283.04
985.09
12789
18.78
034
636
10.00
3 73
529,671
0
CG
4.92
255.69
791
66.88
95 94
12.49
103.82
165 36
270.62
917.05
11029
16.27
065
672
10.00
3 77
598,991
Winter
RFG
0.00
0.00
000
0.00
000
0.00
0.00
000
0.00
0.00
000
0.00
000
000
0.00
000
0
Pool
4.92
255.69
791
66.88
95 94
12.49
103.82
165 36
270.62
917.05
11029
16.27
065
672
10.00
3 77
598,991
CG
4.95
260.83
22 12
64.60
90 13
8.60
127.64
17000
297.06
994.89
132 21
17.63
042
641
15.00
5 54
552,767
Summer
RFG
0.00
0.00
000
0.00
000
0.00
0.00
000
0.00
0.00
000
0.00
000
000
0.00
000
0
E
Pool
4.95
260.83
22 12
64.60
90 13
8.60
127.64
17000
297.06
994.89
132 21
17.63
042
641
15.00
5 54
552,767
5
CG
4.85
256.55
2471
72.87
9624
12.22
104.41
153 13
269.29
882.72
11029
13.40
027
626
15.00
5 64
625,110
Winter
RFG
0.00
0.00
000
0.00
000
0.00
0.00
000
0.00
0.00
000
0.00
000
000
0.00
000
0
5-108
-------
References
1 Ahrens, James, Stancil and Company, Peer Review of Refining Industry Cost Model, September 27, 2013.
2 Lappinen, Mauri and Higgins, Terry, Hart Energy, Review of EPA Proposed Tier 3 Motor Gasoline Refinery Cost
Model, September 2013.
3 Freyman, David, Peer Review of EPA Refinery Tier 3 Cost Model, September 2013.
4 Memorandum from Lester Wyborny to Docket EPA-HQ-OAR-2011-0135, Refinery-by-Refinery Gasoline Sulfur
Cost Model - Response to Peer Review comments for the Final Tier 3 Rulemaking.
5 Conversations with representatives of Foster Wheeler and Bechtel at the American Fuel and Petrochemical
Manufacturers Annual Meeting; March 2012 and March 2013.
6 Annual Energy Outlook 2006, Energy Information Administration, Department of Energy.
7 Lesemann, M, et al, Grace Davison, Increasing FCC Propylene, PTQ Ql 2006.
8 Wagner, Kristen, Grace Davison, Applying ZSM-5 Additives to Meet FCC Propylene and Octane Objectives,
Refinery Operations Vol 3 ISS 02, January 18, 2012.
9 Amalraj, Stephen, Albermarle, Albemarle's AFX Lifts Propylene to new Levels, MEDW 2011.
10 Miller, Rik, KBR and Bhore, Nazeer, Mobil Technology Company, et at, Maxofin: A Novel FCC Process for
Maximizing Light Olefms using a New Generation ZSM-5 Additive, technical paper AM-98-18 presented at the
1998 National Petrochemical and Refiners Association Annual Meeting, March 1998.
1: Peer Review of Refining Industry Cost Model, Stancil and Co., September 27, 2013.
12 Stratiev, D, Lukoil Neftochim, Investigate processing near-zero-sulfur gasoline, Hydrocarbon Processing,
September 2011.
13 Rock, Kerry, CDTech, Long Term Reliability, Hydrocarbon Engineering, September 2004.
14 The Challenges & Opportunities of 10 ppm Sulfur Gasoline. Jay Ross, Delphine Largeteau, Marc Laborde, Larry
Wisdom. NPRA Annual Meeting, AM-11-57, March, 2011, page 6.
15 The Benefits of Cat Feed Hydrotreating and the Impact of Feed Nitrogen on Catalyst Stability. Brian Moyse at
Haldor Topso. NPRA Annual Meeting, AM-10-167, March 2010, Page 2.
16 Shorey, Scott et al, UOP; Exploiting Synergy between FCC and Feed Pretreating Units to Improve Refinery
Margins and Produce Low-Sulfur Fuels; National Petrochemical and Refiners Association Annual Meeting technical
paper AM-99-55, March 1999.
17 Refining Economics of a National Low Sulfur, Low RVP Gasoline Standard, Performed for The International
Council for Clean Transportation by Mathpro, October 25, 2011.
18 Refining Economics of a National Clean Gasoline Standard for PADDs 1 -3; for The Alliance of Automobile
Manufacturers by Mathpro, June 27, 2008.
19 Potential Supply and Cost Impacts of Lower Sulfur, Lower RVP Gasoline; prepared for The American Petroleum
Institute by Baker and O'Brien; July 2011.
5-109
-------
20 Schink, George R., Singer, Hal J., Economic Analysis of the Implications of Implementing EPA's Tier 3 Rules,
prepared for the Emissions Control Technology Association, June 14, 2012.
21 Gary, J., & Handewerk, G. (2001). Petroleum Refining: Technology and Economics. (4th ed.). CRC Press, p 256.
22 Gary, J., & Handewerk, G. (2001). Petroleum Refining: Technology and Economics. (4th ed.). CRC Press, p 337.
23 Keller, P. (2013, February). New Source Review Permitting Impact Analysis for Proposed Tier 3 Gasoline
Program. Memorandum to the docket.
5-110
-------
Chapter 6 Health and Environmental Effects Associated with
Exposure to Criteria and Toxic Pollutants
6.1 Health Effects of Criteria and Toxic Pollutants
Motor vehicles emit pollutants that contribute to ambient concentrations of ozone, PM,
NO2, 862, CO, and air toxics. A discussion of the health effects of these pollutants is included in
this section of the RIA. Children may be more vulnerable to air pollution and other
environmental exposures than adults because their bodily systems are still developing, they
breathe more in proportion to their body size than adults and their behavior can expose them
more to chemicals than adults. Early lifestages (e.g., children) are thought to be more
susceptible to tumor development than adults when exposed to carcinogenic chemicals that act
through a mutagenic mode of action.1
6.1.1 Particulate Matter
6.1.1.1 Background
Particulate matter (PM) is a highly complex mixture of solid particles and liquid droplets
distributed among numerous atmospheric gases which interact with solid and liquid phases.
Particles range in size from those smaller than 1 nanometer (10~9 meter) to over 100 micrometer
(|im, or 10"6 meter) in diameter (for reference, a typical strand of human hair is 70 jim in
diameter and a grain of salt is about 100 jam). Atmospheric particles can be grouped into several
classes according to their aerodynamic and physical sizes, including ultrafine particles (<0.1
|im), accumulation mode or 'fine' particles (< 1 to 3 jim), and coarse particles (>1 to 3 jim). For
regulatory purposes, fine particles are measured as PM2.5 and inhalable or thoracic coarse
particles are measured as PMio-2.5, corresponding to their size (diameter) range in micrometers.
The EPA currently has standards that measure PM2.s and PMi0.2
Particles span many sizes and shapes and may consist of hundreds of different chemicals.
Particles are emitted directly from sources and are also formed through atmospheric chemical
reactions; the former are often referred to as "primary" particles, and the latter as "secondary"
particles. Particle concentration and composition varies by time of year and location, and in
addition to differences in source emissions, is affected by several weather-related factors, such as
temperature, clouds, humidity, and wind. A further layer of complexity comes from particles'
ability to shift between solid/liquid and gaseous phases, which is influenced by concentration and
meteorology, especially temperature.
Fine particles are produced primarily by combustion processes and by transformations of
gaseous emissions (e.g., sulfur oxides (SOx), nitrogen oxides (NOx) and volatile organic
compounds (VOCs)) in the atmosphere. The chemical and physical properties of PM2.5 may vary
greatly with time, region, meteorology and source category. Thus, PM2.s may include a complex
mixture of different components including sulfates, nitrates, organic compounds, elemental
carbon and metal compounds. These particles can remain in the atmosphere for days to weeks
and travel through the atmosphere hundreds to thousands of kilometers.3
6-1
-------
6.1.1.2 Health Effects of PM
Scientific studies show ambient PM is associated with a broad range of health effects.
These health effects are discussed in detail in the December 2009 Integrated Science Assessment
for Paniculate Matter (PM ISA).4 The PM ISA summarizes health effects evidence associated
with both short- and long-term exposures to PM2.5, PMio-2.5, and ultrafine particles. The PM ISA
concludes that human exposures to ambient PM2.5 concentrations are associated with a number
of adverse health effects and characterizes the weight of evidence for these health outcomes.5
The discussion below highlights the PM ISA's conclusions pertaining to health effects associated
with both short- and long-term PM exposures. Further discussion of health effects associated
with PM2.5 can also be found in the rulemaking documents for the most recent review of the PM
NAAQS completed in 2012.6'7
The EPA concludes that a causal relationship exists between both long- and short-term
exposures to PM2.5 and premature mortality and cardiovascular effects and a likely causal
relationship exists between long- and short-term PM2.5 exposures and respiratory effects.
Further, there is evidence suggestive of a causal relationship between long-term PM2.s exposures
and other health effects, including developmental and reproductive effects (e.g., low birth
weight, infant mortality) and carcinogenic, mutagenic, and genotoxic effects (e.g., lung cancer
o
mortality).
As summarized in the Final PM NAAQS rule, and discussed extensively in the 2009 PM
ISA, the scientific evidence available since the completion of the 2006 PM NAAQS review
significantly strengthens the link between long- and short-term exposure to PM2.5 and premature
mortality, while providing indications that the magnitude of the PM2.5- mortality association with
long-term exposures may be larger than previously estimated.9'10 The strongest evidence comes
from recent studies investigating long-term exposure to PM2.5 and cardiovascular-related
mortality. The evidence supporting a causal relationship between long-term PM2.5 exposure and
mortality also includes consideration of new studies that demonstrated an improvement in
community health following reductions in ambient fine particles.11
Several studies evaluated in the 2009 PM ISA have examined the association between
cardiovascular effects and long-term PM2.5 exposures in multi-city studies conducted in the U.S.
and Europe. While studies were not available in the 2006 PM NAAQS review with regard to
long-term exposure and cardiovascular-related morbidity, studies published since then have
provided new evidence linking long-term exposure to PM2.5 with an array of cardiovascular
effects such as heart attacks, congestive heart failure, stroke, and mortality. This evidence is
coherent with studies of short-term exposure to PM2.5 that have observed associations with a
continuum of effects ranging from subtle changes in indicators of cardiovascular health to
serious clinical events, such as increased hospitalizations and emergency department visits due to
cardiovascular disease and cardiovascular mortality.12
As detailed in the 2009 PM ISA, extended analyses of studies available in the 2006 PM
NAAQS review as well as epidemiological studies conducted in the U.S. and abroad published
since then provide stronger evidence of respiratory-related morbidity effects associated with
long-term PM2.5 exposure. The strongest evidence for respiratory-related effects is from studies
that evaluated decrements in lung function growth (in children), increased respiratory symptoms,
6-2
-------
and asthma development. The strongest evidence from short-term PM2.5 exposure studies has
been observed for increased respiratory-related emergency department visits and hospital
admissions for chronic obstructive pulmonary disease (COPD) and respiratory infections.13
The body of scientific evidence detailed in the 2009 PM ISA is still limited with respect
to associations between long-term PM2.5 exposures and developmental and reproductive effects
as well as cancer, mutagenic, and genotoxic effects, but is somewhat expanded from the 2006
review The strongest evidence for an association between PM2.5 and developmental and
reproductive effects comes from epidemiological studies of low birth weight and infant
mortality, especially due to respiratory causes during the post-neonatal period (i.e., 1 month to 12
months of age). With regard to cancer effects, "[mjultiple epidemiologic studies have shown a
consistent positive association between PM2.5 and lung cancer mortality, but studies have
generally not reported associations between PIVb.s and lung cancer incidence."14'15
Specific groups within the general population are at increased risk for experiencing
adverse health effects related to PM exposures.16'17'18'19 The evidence detailed in the 2009 PM
ISA expands our understanding of previously identified at-risk populations and lifestages (i.e.,
children, older adults, and individuals with pre-existing heart and lung disease) and supports the
identification of additional at-risk populations (e.g., persons with lower socioeconomic status,
genetic differences). Additionally, there is emerging, though still limited, evidence for additional
potentially at-risk populations and lifestages, such as those with diabetes, people who are obese,
pregnant women, and the developing fetus.20
For PMio-2.5, the 2009 PM ISA concluded that available evidence was suggestive of a
causal relationship between short-term exposures to PMio-2.5 and cardiovascular effects (e.g.,
hospital admissions and ED visits, changes in cardiovascular function), respiratory effects (e.g,
ED visits and hospital admissions, increase in markers of pulmonary inflammation), and
premature mortality. Data were inadequate to draw conclusions regarding the relationships
between long-term exposure to PMio-2.5 and various health effects. 21'22'23
For ultrafine particles, the 2009 PM ISA concluded that the evidence was suggestive of a
causal relationship between short-term exposures and cardiovascular effects, including changes
in heart rhythm and vasomotor function (the ability of blood vessels to expand and contract). It
also concluded that there was evidence suggestive of a causal relationship between short-term
exposure to ultrafine particles and respiratory effects, including lung function and pulmonary
inflammation, with limited and inconsistent evidence for increases in ED visits and hospital
admissions. Data were inadequate to draw conclusions regarding the relationship between short-
term exposure to ultrafine particle and additional health effects including premature mortality as
well as long-term exposure to ultrafine particles and all health outcomes evaluated.24'25
6.1.2 Ozone
6.1.2.1 B ackground
Ground-level ozone pollution is typically formed through reactions involving VOCs and
NOx in the lower atmosphere in the presence of sunlight. These pollutants, often referred to as
ozone precursors, are emitted by many types of pollution sources such as highway and nonroad
6-3
-------
motor vehicles and engines, power plants, chemical plants, refineries, makers of consumer and
commercial products, industrial facilities, and smaller area sources.
The science of ozone formation, transport, and accumulation is complex. Ground-level
ozone is produced and destroyed in a cyclical set of chemical reactions, many of which are
sensitive to temperature and sunlight. When ambient temperatures and sunlight levels remain
high for several days and the air is relatively stagnant, ozone and its precursors can build up and
result in more ozone than typically occurs on a single high-temperature day. Ozone and its
precursors can be transported hundreds of miles downwind of precursor emissions, resulting in
elevated ozone levels even in areas with low VOC or NOx emissions.
The highest levels of ozone are produced when both VOC and NOx emissions are present
in significant quantities on clear summer days. Relatively small amounts of NOx enable ozone
to form rapidly when VOC levels are relatively high, but ozone production is quickly limited by
removal of the NOx. Under these conditions NOx reductions are highly effective in reducing
ozone while VOC reductions have little effect. Such conditions are called "NOx-limited."
Because the contribution of VOC emissions from biogenic (natural) sources to local ambient
ozone concentrations can be significant, even some areas where man-made VOC emissions are
relatively low can be NOx-limited.
Ozone concentrations in an area also can be lowered by the reaction of nitric oxide (NO)
with ozone, forming nitrogen dioxide (NO2). As the air moves downwind and the cycle
continues, the NO2 forms additional ozone. The importance of this reaction depends, in part, on
the relative concentrations of NOx, VOC, and ozone, all of which change with time and location.
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 "VOC-limited."
Under these conditions, VOC reductions are effective in reducing ozone, but NOx reductions can
actually increase local ozone under certain circumstances. Even in VOC-limited urban areas,
NOx reductions are not expected to increase ozone levels if the NOx reductions are sufficiently
large. Rural areas are usually NOx-limited, due to the relatively large amounts of biogenic VOC
emissions in such areas. Urban areas can be either VOC- or NOx-limited, or a mixture of both,
in which ozone levels exhibit moderate sensitivity to changes in either pollutant.
6.1.2.2 Health Effects of Ozone
This section provides a summary of the health effects associated with exposure to
ambient concentrations of ozone.26 The information in this section is based on the information
and conclusions in the February 2013 Integrated Science Assessment for Ozone (Ozone ISA)
prepared by EPA's Office of Research and Development (ORD).27 The Ozone ISA concludes
that human exposures to ambient concentrations of ozone are associated with a number of
/^o
adverse health effects and characterizes the weight of evidence for these health effects. The
discussion below highlights the Ozone ISA's conclusions pertaining to health effects associated
with both short-term and long-term periods of exposure to ozone.
For short-term exposure to ozone, the Ozone ISA concludes that respiratory effects,
including lung function decrements, pulmonary inflammation, exacerbation of asthma,
respiratory-related hospital admissions, and mortality, are causally associated with ozone
6-4
-------
exposure. It also concludes that cardiovascular effects, including decreased cardiac function and
increased vascular disease, and total mortality are likely to be causally associated with short-term
exposure to ozone and that evidence is suggestive of a causal relationship between central
nervous system effects and short-term exposure to ozone.
For long-term exposure to ozone, the Ozone ISA concludes that respiratory effects,
including new onset asthma, pulmonary inflammation and injury, are likely to be a causally
related with ozone exposure. The Ozone ISA characterizes the evidence as suggestive of a
causal relationship for associations between long-term ozone exposure and cardiovascular
effects, reproductive and developmental effects, central nervous system effects and total
mortality. The evidence is inadequate to infer a causal relationship between chronic ozone
exposure and increased risk of lung cancer.
Finally, interindividual variation in human responses to ozone exposure can result in
some groups being at increased risk for detrimental effects in response to exposure. The Ozone
ISA identified several groups that are at increased risk for ozone-related health effects. These
groups are people with asthma, children and older adults, individuals with reduced intake of
certain nutrients (i.e., Vitamins C and E), outdoor workers, and individuals having certain
genetic variants related to oxidative metabolism or inflammation. Ozone exposure during
childhood can have lasting effects through adulthood. Such effects include altered function of
the respiratory and immune systems. Children absorb higher doses (normalized to lung surface
area) of ambient ozone, compared to adults, due to their increased time spent outdoors, higher
ventilation rates relative to body size, and a tendency to breathe a greater fraction of air through
the mouth. Children also have a higher asthma prevalence compared to adults. Additional
children's vulnerability and susceptibility factors are listed in Section XII.G of the preamble.
6.1.3 Nitrogen Oxides and Sulfur Oxides
6.1.3.1 B ackground
Sulfur dioxide (802), a member of the sulfur oxide (SOx) family of gases, is formed from
burning fuels containing sulfur (e.g., coal or oil), extracting gasoline from oil, or extracting
metals from ore. Nitrogen dioxide (NO2) is a member of the nitrogen oxide (NOx) family of
gases. Most NO2 is formed in the air through the oxidation of nitric oxide (NO) emitted when
fuel is burned at a high temperature. SO2 andNO2 and their gas phase oxidation products can
dissolve in water droplets and further oxidize to form sulfuric and nitric acid which react with
ammonia to form sulfates and nitrates, both of which are important components of ambient PM.
The health effects of ambient PM are discussed in Section 6.1.1.2. NOx along with VOCs are
the two major precursors of ozone. The health effects of ozone are covered in Section 6.1.2.2.
6.1.3.2 Health Effects of Sulfur Oxides
This section provides an overview of the health effects associated with SO2. Additional
information on the health effects of SO2 can be found in the 2008 Integrated Science Assessment
9Q
for Sulfur Oxides (SO2 ISA). Following an extensive evaluation of health evidence from
epidemiologic and laboratory studies, the U.S. EPA has concluded that there is a causal
relationship between respiratory health effects and short-term exposure to SO2. The immediate
6-5
-------
effect of SC>2 on the respiratory system in humans is bronchoconstriction. Asthmatics are more
sensitive to the effects of SO2 likely resulting from preexisting inflammation associated with this
disease. In addition to those with asthma (both children and adults), potentially sensitive groups
include all children and the elderly. In laboratory studies involving controlled human exposures
to SC>2, respiratory effects have consistently been observed following 5-10 min exposures at SC>2
concentrations > 0.4 ppm in asthmatics engaged in moderate to heavy levels of exercise, with
more limited evidence of respiratory effects among exercising asthmatics exposed to
concentrations as low as 0.2-0.3 ppm. A clear concentration-response relationship has been
demonstrated in these studies following exposures to SC>2 at concentrations between 0.2 and 1.0
ppm, both in terms of increasing severity of respiratory symptoms and decrements in lung
function, as well as the percentage of asthmatics adversely affected.
In epidemiologic studies, respiratory effects have been observed in areas where the mean
24-hour SC>2 levels range from 1 to 30 ppb, with maximum 1 to 24-hour average SC>2 values
ranging from 12 to 75 ppb. Important new multicity studies and several other studies have found
an association between 24-hour average ambient 862 concentrations and respiratory symptoms
in children, particularly those with asthma. Generally consistent associations also have been
observed between ambient SC>2 concentrations and emergency department visits and
hospitalizations for all respiratory causes, particularly among children and older adults (> 65
years), and for asthma. A limited subset of epidemiologic studies has examined potential
confounding by copollutants using multipollutant regression models. These analyses indicate
that although copollutant adjustment has varying degrees of influence on the SC>2 effect
estimates, the effect of SO2 on respiratory health outcomes appears to be generally robust and
independent of the effects of gaseous and particulate copollutants, suggesting that the observed
effects of 862 on respiratory endpoints occur independent of the effects of other ambient air
pollutants.
Consistent associations between short-term exposure to SO2 and mortality have been
observed in epidemiologic studies, with larger effect estimates reported for respiratory mortality
than for cardiovascular mortality. While this finding is consistent with the demonstrated effects
of SC>2 on respiratory morbidity, uncertainty remains with respect to the interpretation of these
associations due to potential confounding by various copollutants. The U.S. EPA has therefore
concluded that the overall evidence is suggestive of a causal relationship between short-term
exposure to 862 and mortality. Significant associations between short-term exposure to 862 and
emergency department visits and hospital admissions for cardiovascular diseases have also been
reported. However, these findings have been inconsistent across studies and do not provide
adequate evidence to infer a causal relationship between 862 exposure and cardiovascular
morbidity.
6.1.3.3 Health Effects of Nitrogen Oxides
The most recent review of the health effects of oxides of nitrogen completed by the EPA
can be found in the 2008 Integrated Science Assessment for Nitrogen Oxides (NOx ISA).30 The
EPA concluded that the findings of epidemiologic, controlled human exposure, and animal
toxicological studies provide evidence that is sufficient to infer a likely causal relationship
between respiratory effects and short-term NO2 exposure. The 2008 NOx ISA concluded that the
strongest evidence for such a relationship comes from epidemiologic studies of respiratory
6-6
-------
effects including increased respiratory symptoms, emergency department visits, and hospital
admissions. Based on both short- and long-term exposure studies, the 2008 NOx ISA concluded
that individuals with preexisting pulmonary conditions (e.g., asthma or COPD), children, and
older adults are potentially at greater risk of MVrelated respiratory effects. Based on findings
from controlled human exposure studies, the 2008 NOx ISA also drew two broad conclusions
regarding airway responsiveness following NC>2 exposure. First, the NOx ISA concluded that
NO2 exposure may enhance the sensitivity to allergen-induced decrements in lung function and
increase the allergen-induced airway inflammatory response following 30-minute exposures of
asthmatic adults to NO2 concentrations as low as 260 ppb. Second, exposure to NO2 has been
found to enhance the inherent responsiveness of the airway to subsequent nonspecific challenges
in controlled human exposure studies of healthy and asthmatic adults. Small but statistically
significant increases in nonspecific airway hyperresponsiveness were reported for asthmatic
adults following 30-minute exposures to 200-300 ppb NO2 and following 1-hour exposures of
asthmatics to 100 ppb NO2. Enhanced airway responsiveness could have important clinical
implications for asthmatics since transient increases in airway responsiveness following NO2
exposure have the potential to increase symptoms and worsen asthma control. Together, the
epidemiologic and experimental data sets form a plausible, consistent, and coherent description
of a relationship between NO2 exposures and an array of adverse health effects that range from
the onset of respiratory symptoms to hospital admission.
In evaluating a broader range of health effects, the 2008 NOx ISA concluded evidence
was "suggestive but not sufficient to infer a causal relationship" between short-term NO2
exposure and premature mortality and between long-term NO2 exposure and respiratory effects.
The latter was based largely on associations observed between long-term NO2 exposure and
decreases in lung function growth in children. Furthermore, the 2008 NOx ISA concluded that
evidence was "inadequate to infer the presence or absence of a causal relationship" between
short-term NO2 exposure and cardiovascular effects as well as between long-term NO2 exposure
and cardiovascular effects, reproductive and developmental effects, premature mortality, and
cancer.31 The conclusions for these health effect categories were informed by uncertainties in the
evidence base such as the independent effects of NO2 exposure within the broader mixture of
traffic-related pollutants, limited evidence from experimental studies, and/or an overall limited
literature base.
6.1.4 Health Effects of Carbon Monoxide
Information on the health effects of carbon monoxide (CO) can be found in the January
2010 Integrated Science Assessment for Carbon Monoxide (CO ISA).32 The CO ISA concludes
that ambient concentrations of CO are associated with a number of adverse health effects.33 This
section provides a summary of the health effects associated with exposure to ambient
concentrations of CO.34
Controlled human exposure studies of subjects with coronary artery disease show a
decrease in the time to onset of exercise-induced angina (chest pain) and electrocardiogram
changes following CO exposure. In addition, epidemiologic studies show associations between
short-term CO exposure and cardiovascular morbidity, particularly increased emergency room
visits and hospital admissions for coronary heart disease (including ischemic heart disease,
myocardial infarction, and angina). Some epidemiologic evidence is also available for increased
6-7
-------
hospital admissions and emergency room visits for congestive heart failure and cardiovascular
disease as a whole. The CO ISA concludes that a causal relationship is likely to exist between
short-term exposures to CO and cardiovascular morbidity. It also concludes that available data
are inadequate to conclude that a causal relationship exists between long-term exposures to CO
and cardiovascular morbidity.
Animal studies show various neurological effects with in-utero CO exposure. Controlled
human exposure studies report central nervous system and behavioral effects following low-level
CO exposures, although the findings have not been consistent across all studies. The CO ISA
concludes the evidence is suggestive of a causal relationship with both short- and long-term
exposure to CO and central nervous system effects.
A number of studies cited in the CO ISA have evaluated the role of CO exposure in birth
outcomes such as preterm birth or cardiac birth defects. The epidemiologic studies provide
limited evidence of a CO-induced effect on preterm births and birth defects, with weak evidence
for a decrease in birth weight. Animal toxicological studies have found perinatal CO exposure to
affect birth weight, as well as other developmental outcomes. The CO ISA concludes the
evidence is suggestive of a causal relationship between long-term exposures to CO and
developmental effects and birth outcomes.
Epidemiologic studies provide evidence of associations between ambient CO
concentrations and respiratory morbidity such as changes in pulmonary function, respiratory
symptoms, and hospital admissions. A limited number of epidemiologic studies considered
copollutants such as ozone, SO2, and PM in two-pollutant models and found that CO risk
estimates were generally robust, although this limited evidence makes it difficult to disentangle
effects attributed to CO itself from those of the larger complex air pollution mixture. Controlled
human exposure studies have not extensively evaluated the effect of CO on respiratory
morbidity. Animal studies at levels of 50-100 ppm CO show preliminary evidence of altered
pulmonary vascular remodeling and oxidative injury. The CO ISA concludes that the evidence
is suggestive of a causal relationship between short-term CO exposure and respiratory morbidity,
and inadequate to conclude that a causal relationship exists between long-term exposure and
respiratory morbidity.
Finally, the CO ISA concludes that the epidemiologic evidence is suggestive of a causal
relationship between short-term concentrations of CO and mortality. Epidemiologic studies
provide evidence of an association between short-term exposure to CO and mortality, but limited
evidence is available to evaluate cause-specific mortality outcomes associated with CO exposure.
In addition, the attenuation of CO risk estimates which was often observed in copollutant models
contributes to the uncertainty as to whether CO is acting alone or as an indicator for other
combustion-related pollutants. The CO ISA also concludes that there is not likely to be a causal
relationship between relevant long-term exposures to CO and mortality.
6-8
-------
6.1.5 Health Effects of Air Toxics
6.1.5.1 Benzene
The EPA's IRIS database lists benzene as a known human carcinogen (causing leukemia)
by all routes of exposure, and concludes that exposure is associated with additional health
effects, including genetic changes in both humans and animals and increased proliferation of
bone marrow cells in mice.35'36'37 EPA states in its IRIS database that data indicate a causal
relationship between benzene exposure and acute lymphocyte leukemia and suggest a
relationship between benzene exposure and chronic non-lymphocytic leukemia and chronic
lymphocyte leukemia. EPA's IRIS documentation for benzene also lists a range of 2.2 x 10~6 to
7.8 x 10~6 as the unit risk estimate (URE) for benzene.38'39 The International Agency for
Research on Carcinogens (IARC) has determined that benzene is a human carcinogen and the
U.S. Department of Health and Human Services (DHHS) has characterized benzene as a known
u-4041
human carcinogen.
A number of adverse noncancer health effects including blood disorders, such as
preleukemia and aplastic anemia, have also been associated with long-term exposure to
benzene.42'43 The most sensitive noncancer effect observed in humans, based on current data, is
the depression of the absolute lymphocyte count in blood.44'45 EPA's inhalation reference
concentration (RfC) for benzene is 30 |ig/m3. The RfC is based on suppressed absolute
lymphocyte counts seen in humans under occupational exposure conditions. In addition, recent
work, including studies sponsored by the Health Effects Institute (HEI), provides evidence that
biochemical responses are occurring at lower levels of benzene exposure than previously
known.46'47'48'49 EPA's IRIS program has not yet evaluated these new data. EPA does not
currently have an acute reference concentration for benzene. The Agency for Toxic Substances
and Disease Registry (ATSDR) Minimal Risk Level (MRL) for acute exposure to benzene is 29
|ig/m3 for 1-14 days exposure. °'51
6.1.5.2 1,3-Butadiene
EPA has characterized 1,3-butadiene as carcinogenic to humans by inhalation.52'53 The
IARC has determined that 1,3-butadiene is a human carcinogen and the U.S. DHHS has
characterized 1,3-butadiene as a known human carcinogen.5 '55'56 There are numerous studies
consistently demonstrating that 1,3-butadiene is metabolized into genotoxic metabolites by
experimental animals and humans. The specific mechanisms of 1,3-butadiene-induced
carcinogenesis are unknown; however, the scientific evidence strongly suggests that the
carcinogenic effects are mediated by genotoxic metabolites. Animal data suggest that females
may be more sensitive than males for cancer effects associated with 1,3-butadiene exposure;
there are insufficient data in humans from which to draw conclusions about sensitive
subpopulations. The URE for 1,3-butadiene is 3 x io~5 per |ig/m3.57 1,3-butadiene also causes a
variety of reproductive and developmental effects in mice; no human data on these effects are
available. The most sensitive effect was ovarian atrophy observed in a lifetime bioassay of
female mice.58 Based on this critical effect and the benchmark concentration methodology, an
RfC for chronic health effects was calculated at 0.9 ppb (approximately 2 |ig/m3).
6-9
-------
6.1.5.3 Ethanol
EPA is planning to develop an assessment of the health effects of exposure to ethanol, a
compound which is not currently listed on EPA's IRIS database. Extensive health effects data
are available for ingestion of ethanol, while data on inhalation exposure effects are sparse. In
developing the assessment, EPA is evaluating pharmacokinetic models as a means of
extrapolating across species (animal to human) and across exposure routes (oral to inhalation) to
better characterize the health hazards and dose-response relationships for low levels of ethanol
exposure in the environment.
6.1.5.4 Formaldehyde
In 1991, EPA concluded that formaldehyde is a carcinogen based on nasal tumors in
animal bioassays.59 An Inhalation Unit Risk for cancer and a Reference Dose for oral noncancer
effects were developed by the Agency and posted on the Integrated Risk Information System
(IRIS) database. Since that time, the National Toxicology Program (NTP) and International
Agency for Research on Cancer (IARC) have concluded that formaldehyde is a known human
carcinogen.60'61'62
The conclusions by IARC and NTP reflect the results of epidemiologic research
published since 1991 in combination with previous animal, human and mechanistic evidence.
Research conducted by the National Cancer Institute reported an increased risk of
nasopharyngeal cancer and specific lymphohematopoietic malignancies among workers exposed
to formaldehyde.63'64'65 A National Institute of Occupational Safety and Health study of garment
workers also reported increased risk of death due to leukemia among workers exposed to
formaldehyde.66 Extended follow-up of a cohort of British chemical workers did not report
evidence of an increase in nasopharyngeal or lymphohematopoietic cancers, but a continuing
statistically significant excess in lung cancers was reported.67 Finally, a study of embalmers
reported formaldehyde exposures to be associated with an increased risk of myeloid leukemia
/-Q
but not brain cancer.
Health effects of formaldehyde in addition to cancer were reviewed by the Agency for
Toxics Substances and Disease Registry in 199969 and supplemented in 2010,70 and by the World
Health Organization.71 These organizations reviewed the literature concerning effects on the
eyes and respiratory system, the primary point of contact for inhaled formaldehyde, including
sensory irritation of eyes and respiratory tract, pulmonary function, nasal histopathology, and
immune system effects. In addition, research on reproductive and developmental effects and
neurological effects were discussed.
EPA released a draft Toxicological Review of Formaldehyde - Inhalation Assessment
through the IRIS program for peer review by the National Research Council (NRC) and public
comment in June 2010.72 The draft assessment reviewed more recent research from animal and
human studies on cancer and other health effects. The NRC released their review report in April
201173 (http://www.nap.edu/catalog.php?record_id=13142). The EPA is currently revising the
draft assessment in response to this review.
6-10
-------
6.1.5.5 Acetaldehyde
Acetaldehyde is classified in EPA's IRIS database as a probable human carcinogen,
based on nasal tumors in rats, and is considered toxic by the inhalation, oral, and intravenous
routes.74 The URE in IRIS for acetaldehyde is 2.2 x 10'6 per |ig/m3.75 Acetaldehyde is
reasonably anticipated to be a human carcinogen by the U.S. DHHS in the 12* Report on
Carcinogens and is classified as possibly carcinogenic to humans (Group 2B) by the IARC.76'77
EPA is currently conducting a reassessment of cancer risk from inhalation exposure to
acetaldehyde.
The primary noncancer effects of exposure to acetaldehyde vapors include irritation of
7R
the eyes, skin, and respiratory tract. In short-term (4 week) rat studies, degeneration of
olfactory epithelium was observed at various concentration levels of acetaldehyde exposure.79'80
Data from these studies were used by EPA to develop an inhalation reference concentration of 9
|ig/m3. Some asthmatics have been shown to be a sensitive subpopulation to decrements in
functional expiratory volume (FEV1 test) and bronchoconstriction upon acetaldehyde
inhalation.81 The agency is currently conducting a reassessment of the health hazards from
inhalation exposure to acetaldehyde.
6.1.5.6 Acrolein
EPA most recently evaluated the toxicological and health effects literature related to
acrolein in 2003 and concluded that the human carcinogenic potential of acrolein could not be
determined because the available data were inadequate. No information was available on the
carcinogenic effects of acrolein in humans and the animal data provided inadequate evidence of
oa
carcinogenicity. The IARC determined in 1995 that acrolein was not classifiable as to its
carcinogenicity in humans.83
Lesions to the lungs and upper respiratory tract of rats, rabbits, and hamsters have been
observed after subchronic exposure to acrolein.84 The Agency has developed an RfC for acrolein
o o c
of 0.02 |ig/m and an RfD of 0.5 |ig/kg-day. EPA is considering updating the acrolein
assessment with data that have become available since the 2003 assessment was completed.
Acrolein is extremely acrid and irritating to humans when inhaled, with acute exposure
resulting in upper respiratory tract irritation, mucus hypersecretion and congestion. The intense
irritancy of this carbonyl has been demonstrated during controlled tests in human subjects, who
suffer intolerable eye and nasal mucosal sensory reactions within minutes of exposure.86 These
data and additional studies regarding acute effects of human exposure to acrolein are
summarized in EPA's 2003 IRIS Human Health Assessment for acrolein.87 Studies in humans
indicate that levels as low as 0.09 ppm (0.21 mg/m3) for five minutes may elicit subjective
complaints of eye irritation with increasing concentrations leading to more extensive eye, nose
and respiratory symptoms. Acute exposures in animal studies report bronchial hyper-
responsiveness. Based on animal data (more pronounced respiratory irritancy in mice with
oo
allergic airway disease in comparison to non-diseased mice ) and demonstration of similar
effects in humans (e.g., reduction in respiratory rate), individuals with compromised respiratory
function (e.g., emphysema, asthma) are expected to be at increased risk of developing adverse
responses to strong respiratory irritants such as acrolein. EPA does not currently have an acute
6-11
-------
reference concentration for acrolein. The available health effect reference values for acrolein
have been summarized by EPA and include an ATSDR MRL for acute exposure to acrolein of
7 |ig/m3 for 1-14 days exposure; and Reference Exposure Level (REL) values from the
California Office of Environmental Health Hazard Assessment (OEHHA) for one-hour and 8-
hour exposures of 2.5 |ig/m3 and 0.7 |ig/m3, respectively.89
6.1.5.7 PAN
PAN (peroxy acetyl nitrate) has not been evaluated by EPA's IRIS program. Information
regarding the potential carcinogenicity of PAN is limited. As noted in the EPA air quality
criteria document for ozone and related photochemical oxidants, cytogenetic studies indicate that
PAN is not a potent mutagen, clastogen (a compound that can cause breaks in chromosomes), or
DNA-damaging agent in mammalian cells either in vivo or in vitro. Some studies suggest that
PAN may be a weak bacterial mutagen at concentrations much higher than exist in present urban
atmospheres.90
Effects of ground-level smog causing intense eye irritation have been attributed to
photochemical oxidants, including PAN.91 Animal toxicological information on the inhalation
effects of the non-ozone oxidants has been limited to a few studies on PAN. Acute exposure to
levels of PAN can cause changes in lung morphology, behavioral modifications, weight loss, and
susceptibility to pulmonary infections. Human exposure studies indicate minor pulmonary
function effects at high PAN concentrations, but large inter-individual variability precludes
definitive conclusions.92
6.1.5.8 Polycyclic Organic Matter
The term polycyclic organic matter (POM) defines a broad class of compounds that
includes the polycyclic aromatic hydrocarbon compounds (PAHs). One of these compounds,
naphthalene, is discussed separately below. POM compounds are formed primarily from
combustion and are present in the atmosphere in gas and particulate form. Cancer is the major
concern from exposure to POM. Epidemiologic studies have reported an increase in lung cancer
in humans exposed to diesel exhaust, coke oven emissions, roofing tar emissions, and cigarette
smoke; all of these mixtures contain POM compounds.9394 Animal studies have reported
respiratory tract tumors from inhalation exposure to benzo[a]pyrene and alimentary tract and
liver tumors from oral exposure to benzo[a]pyrene.95 In 1997 EPA classified seven PAHs
(benzo[a]pyrene, benz[a]anthracene, chrysene, benzo[b]fluoranthene, benzo[k]fluoranthene,
dibenz[a,h]anthracene, and indeno[l,2,3-cd]pyrene) as Group B2, probable human carcinogens.96
Since that time, studies have found that maternal exposures to PAHs in a population of pregnant
women were associated with several adverse birth outcomes, including low birth weight and
reduced length at birth, as well as impaired cognitive development in preschool children (3 years
of age).97'98 These and similar studies are being evaluated as a part of the ongoing IRIS
assessment of health effects associated with exposure to benzo[a]pyrene.
6.1.5.9 Naphthalene
Naphthalene is found in small quantities in gasoline and diesel fuels. Naphthalene
emissions have been measured in larger quantities in both gasoline and diesel exhaust compared
6-12
-------
with evaporative emissions from mobile sources, indicating it is primarily a product of
combustion. Acute (short-term) exposure of humans to naphthalene by inhalation, ingestion, or
dermal contact is associated with hemolytic anemia and damage to the liver and the nervous
system." Chronic (long term) exposure of workers and rodents to naphthalene has been reported
to cause cataracts and retinal damage.100 EPA released an external review draft of a
reassessment of the inhalation carcinogenicity of naphthalene based on a number of recent
animal carcinogenicity studies.101 The draft reassessment completed external peer review.102
Based on external peer review comments received, a revised draft assessment that considers all
routes of exposure, as well as cancer and noncancer effects, is under development. The external
review draft does not represent official agency opinion and was released solely for the purposes
of external peer review and public comment. The National Toxicology Program listed
naphthalene as "reasonably anticipated to be a human carcinogen" in 2004 on the basis of
bioassays reporting clear evidence of carcinogenicity in rats and some evidence of
carcinogenicity in mice.103 California EPA has released a new risk assessment for naphthalene,
and the IARC has reevaluated naphthalene and re-classified it as Group 2B: possibly
carcinogenic to humans.104
Naphthalene also causes a number of chronic non-cancer effects in animals, including
abnormal cell changes and growth in respiratory and nasal tissues.105 The current EPA IRIS
assessment includes noncancer data on hyperplasia and metaplasia in nasal tissue that form the
basis of the inhalation RfC of 3 |ig/m3.106 The ATSDR MRL for acute exposure to naphthalene
is 0.6 mg/kg/day.
6.1.5.10 Other Air Toxics
In addition to the compounds described above, other compounds in gaseous hydrocarbon
and PM emissions from vehicles will be affected by this proposal. Mobile source air toxic
compounds that would potentially be impacted include ethylbenzene, propionaldehyde, toluene,
and xylene. Information regarding the health effects of these compounds can be found in EPA's
IRIS database.107
6.1.6 Traffic-associated health effects
In addition to health concerns resulting from specific air pollutants, a large number of
studies have examined the health status of populations near major roadways. These studies
frequently have employed exposure metrics that are not specific to individual pollutants, but
rather reflect the large number of different pollutants found in elevation near major roads.
In this section of the RIA, information on health effects associated with air quality near
major roads or traffic in general is summarized. Generally, the section makes use of publications
that systematically review literature on a given health topic. In particular, this section makes
frequent reference of a report of by the Health Effects Institute (HEI) Panel on the Health Effects
of Traffic-Related Air Pollution, published in 2010 as a review of relevant studies.108'109 Other
systematic reviews of relevant literature are cited were appropriate.
6-13
-------
6.1.6.1 Populations near major roads
Numerous studies have estimated the size and demographics of populations that live near
major roads. Other studies have estimated the number of schools near major roads, and the
populations of students in such schools.
Every two years, the U.S. Census Bureau's American Housing Survey (AHS) has
reported whether housing units are within 300 feet of an "airport, railroad, or highway with four
or more lanes." The 2009 survey reports that over 22 million homes, or 17 percent of all housing
units in the U.S., were located in such areas. Assuming that populations and housing units are in
the same locations, this corresponds to a population of more than 50 million U.S. residents in
close proximity to high-traffic roadways or other transportation sources. According to the
Central Intelligence Agency's World Factbook, in 2010, the United States had 6,506,204 km or
roadways, 224,792 km of railways, and 15,079 airports. As such, highways represent the
overwhelming majority of transportation facilities described by this factor in the AHS.
The AHS reports are published every two years, and until 2011 recorded whether homes
were located near highways with four or more lanes, railroads, or airports. As such, trends in the
AHS can be reported to describe whether a greater or lesser proportion of homes are located near
major roads over time. Figure 6-1 depicts trends in the number and proportion of homes located
near major transportation sources, which generally indicate large roadways. As the figure
indicates, since 2005, there has been a substantial increase in the number and percentage of
homes located near major transportation sources. As such, the population in close proximity to
these sources, which may be affected by near-road air quality and health concerns, appears to
have increased over time.
6-14
-------
1 18 +
o
I
I 16 +
14 --
12 --
10
18.0%
16.0%
14.0%
Ł
^
12.0% "§
-i— i
w
10.0% H
8.0% o>
6.0%
4.0%
2.0%
0.0%
i iMiiiinn Housing Units
•Near4+ Lane Highway,
Railroad, or Airport
1997 1999 2001 2003 2005 2007 2009
Yearof American Housing Survey
Figure 6-1 Trends in Populations Near Large Highways, Railroads, and Airports
Furthermore, according to data from the 2008 American Time Use Survey (ATUS),
conducted by the Bureau of Labor Statistics (BTS), Americans spend more than an hour
traveling each day, on average.110 Although the ATUS does not indicate their mode of travel, the
majority of trips undertaken nationally is by motor vehicle.111 As such, daily travel activity
brings nearly all residents into a high-exposure microenvironment for part of the day.
6.1.6.2 Premature mortality
The HEI panel report concluded that evidence linking traffic-associated air pollution with
premature mortality from all causes was "suggestive but not sufficient" to infer a causal
relationship. This conclusion was based largely on several long-term studies that "qualitatively"
examined whether or not someone was exposed to traffic-associated air pollution. In addition,
based on several short-term studies of exposure, the panel concluded that there was evidence that
there was "suggestive but not sufficient" evidence to infer a causal relation between traffic-
related exposure and cardiovascular mortality.
6.1.6.3 Cardiovascular effects
6.1.6.3.1 Cardiac physiology
Exposure to traffic-associated pollutants has been associated with changes in cardiac
physiology, including cardiac function. One common measure of cardiac function is heart rate
6-15
-------
variability (HRV), an indicator of the heart's ability to respond to variations in stress, reflecting
119
the nervous system's ability to regulate the heart. Reduced HRV is associated with adverse
cardiovascular events, such as myocardial infarction, in heart disease patients. The HEI panel
concluded that available evidence provides evidence for a causal association between exposure
to traffic-related pollutants and reduced control of HRV by the nervous system. Overall, the
panel concluded that the evidence was "suggestive but not sufficient" to infer a causal relation
between traffic-related pollutants and cardiac function. Studies suggest that the HRV changes
from traffic-related air pollution result in changes to heart rhythms, which can lead to
arrhythmia.113'114
6.1.6.3.2 Heart attack and atherosclerosis
The HEI panel concluded that epidemiologic evidence of the association between traffic-
related pollutants and heart attacks and atherosclerosis was "suggestive but not sufficient" to
infer a causal association. In addition, the panel concluded that the toxicology studies they
reviewed provided "suggestive evidence that exposure to traffic emissions, including ambient
and laboratory-generated [PM] and diesel- and gasoline-engine exhaust, alters cardiovascular
function." The panel noted there are few studies of human volunteers exposed to real-world
traffic mixture, which were not entirely consistent. The panel notes that the studies provide
consistent evidence for exposure to PM and impaired cardiovascular responses. In addition to
the HEI study, several other reviews of available evidence conclude that there is evidence
supporting a causal association between traffic-related air pollution and cardiovascular
disease.11
A number of mechanisms for cardiovascular disease are highlighted in the HEI and AHA
report, including modified blood vessel endothelial function (e.g., the ability to dilate),
atherosclerosis, and oxidative stress. The HEI review cites "two well executed studies" in which
hospitalization for acute myocardial infarction (i.e., heart attack) were associated with traffic
exposures and a prospective study finding higher rates of arterial hardening and coronary heart
disease near traffic.
6.1.6.4 Respiratory effects
6.1.6.4.1 Asthma
Pediatric asthma and asthma symptoms are the effects that have been evaluated by the
largest number of studies in the epidemiologic literature on the topic. In general, studies
consistently show effects of residential or school exposure to traffic and asthma symptoms, and
the effects are frequently statistically significant. Studies have employed both short-term and
long-term exposure metrics, and a range of different respiratory measures. HEI Special Report
17 (HEI Panel on the Health Effects of Traffic-Related Air Pollution, 2010) concluded that there
is sufficient evidence for a causal association between exposure to traffic-related air pollution
and exacerbation of asthma symptoms in children.
While there is general consistency in studies examining asthma incidence in children, the
available studies employ different definitions of asthma (e.g., self-reported vs. hospital records),
methods of exposure assessment, and population age ranges. As such, the overall evidence,
6-16
-------
while supportive of an association between traffic exposure and new onset asthma, are less
consistent than for asthma symptoms. The HEI report determined that there is "sufficient" or
"suggestive" evidence of a causal relationship between exposure to traffic-related air pollution
and incident (new onset) asthma in children (HEI Panel on the Health Effects of Traffic-Related
Air Pollution, 2010). A recent meta-analysis of studies on incident asthma and air pollution in
general, based on studies dominated by traffic-linked exposure metrics, also concluded that
available evidence that exposures is consistent with a effect of exposure on asthma incidence
(Anderson et al., 2011). The study reported excess main risk estimates for different pollutants
ranging from 7-16 percent per 10 |J.g/m3 of long-term exposure (random effects models). Other
qualitative reviews (Salam et al., 2008; Braback and Forsberg, 2009) conclude that available
evidence is consistent with the hypothesis that traffic-associated air pollutants are associated with
incident asthma.
6.1.6.4.2 Chronic obstructive pulmonary disease (COPD)
The HEI panel reviewed available studies examining COPD in the context of traffic-
associated air pollution. Because of how the panel selected studies for inclusion in review, there
were only two studies that they used to review the available evidence. Both studies reported
some positive associations, but not for all traffic metrics. The small number of studies and lack
of consistency across traffic metrics led the panel to conclude that there is insufficient evidence
for traffic-associated air pollution causing COPD.
6.1.6.4.3 Allergy
There are numerous human and animal experimental studies that provides strongly
suggestive evidence that traffic-related air pollutants can enhance allergic responses to common
allergens.116'117'118 However, in its review of 16 epidemiologic studies that address traffic-related
air pollution's effect on allergies, the HEI expert panel (HEI, 2010) reported that only two such
studies showed consistently positive associations. As a result, despite the strong experimental
evidence, the panel concluded that there is "inadequate/insufficient" evidence of an association
between allergy and traffic-associated air pollution. As noted above, the HEI panel considered
toxicological evidence only based on whether or not they provide mechanistic support for
observations and inferences derived from epidemiology.
6.1.6.4.4 Lung function
There are numerous measurements of breathing (spirometry) that indicate the presence or
degree of airway disease, such as asthma and chronic obstructive pulmonary disease (COPD).
Forced vital capacity (FVC) is measured when a patient maximally fills their lungs and then
blows their hardest in completely exhaling. The peak expiratory flow (PEF) is the maximum air
flow achievable during exhalation. The forced expiratory volume in the first second of
exhalation is referred to as FEVi. FEVi and PEF reflect the function of the large airways. FVC
and FEVi, along with their ratio (FVC/FEVi) are used to classify airway obstruction in asthma
and COPD. Measurements of air flow at various times during forced exhalation, such as 25
percent, 50 percent, and 75 percent, are also used. The flow at 75 percent of forced exhalation
) reflects the status of small airways, which asthma and COPD affect.
6-17
-------
The HEI panel concluded that the available literature suggests that long-term exposure to
traffic-related air pollution is associated with reduced lung function in adolescents and young
adults and that lung function is lower in populations in areas with high traffic-related air
pollutant levels. However, the panel noted the difficulty of disentangling traffic-specific
exposures from urban air pollution in general. The studies reviewed that were more specifically
oriented toward traffic were not consistent in their findings. As a result, the panel found that the
evidence linking lung function and traffic exposure is "inadequate and insufficient" to infer a
causal relationship.
6.1.6.5 Reproductive and developmental effects
Several studies have reported associations between traffic-related air pollution and
adverse birth outcomes, such as preterm birth and low birth weight. At the time of the HEI
review, the panel concluded that evidence for adverse birth outcomes being causally associated
with traffic-related exposures was "inadequate and insufficient." Only four studies met the
panel's inclusion criteria, and had limited geographic coverage. One study provided evidence of
small but consistently increased risks using multiple exposure metrics. No studies were at the
time available that examined traffic-specific exposures and congenital abnormalities. Since then,
several studies investigating birth outcomes have been published, but no new systematic reviews.
One new meta-analysis of air pollution and congenital abnormalities has been published, though
none of the reviewed studies includes traffic-specific exposure information.
The HEI panel also reviewed toxicological studies of traffic-related air pollutants and
fertility. While numerous studies examining animal or human exposure and sperm count have
been published, the panel concluded that the generally high exposure concentrations employed in
the studies limited the applicability to typical ambient concentrations. Because there was no
overlap in the effects studied by epidemiology and toxicology studies, no synthesis review of the
combined literature was undertaken.
Since the HEI panel's publication, a systematic review and meta-analysis of air pollution
and congenital abnormalities was published.119 In that review, only one study directly included
nearby traffic in its exposure analysis. As such, there are so systematic reviews that specifically
address traffic's impact on congenital abnormalities.
6.1.6.6 Cancer
6.1.6.6.1 Childhood cancer
A number of studies examining various types of childhood cancer have been published
with mixed results. The HEI panel concluded that the available epidemiologic evidence was
"inadequate and insufficient" to infer a causal relationship between traffic-related air pollution
and childhood cancer. An earlier review article on the topic noted that studies reporting positive
effects tended to be small, while those with null effects tended to be larger, suggesting the
potential for publication bias in the available literature. 12°
6-18
-------
6.1.6.6.2 Adult cancer
Several studies have examined the risk of adult lung cancers in relation to exposure to
traffic-related air pollutants. The HEI panel evaluated four such studies, and rated the available
evidence as "inadequate and insufficient" to infer a causal relation for non-occupational lung
cancer.
6.1.6.7 Neurological effects
The HEI panel found that current toxicologic and epidemiologic literature on the
neurotoxicity of traffic-related air pollution was inadequate for their evaluation. The panel noted
that there were a number of toxicologic studies of traffic-associated pollutants, but found them to
have diverse exposure protocols, animal models, and endpoints, making them unsuitable for
systematic evaluation.
6.2 Environmental Effects of Criteria and Toxic Pollutants
6.2.1 Visibility Degradation
Visibility can be defined as the degree to which the atmosphere is transparent to visible
light.121 Visibility impairment is caused by light scattering and absorption by suspended
particles and gases. Visibility is important because it has direct significance to people's
enjoyment of daily activities in all parts of the country. Individuals value good visibility for the
well-being it provides them directly, where they live and work, and in places where they enjoy
recreational opportunities. Visibility is also highly valued in significant natural areas, such as
national parks and wilderness areas, and special emphasis is given to protecting visibility in these
areas. For more information on visibility see the final 2009 PM ISA. 22
EPA is working to address visibility impairment. In 1999, EPA finalized the regional
haze program (64 FR 35714) to protect the visibility in Mandatory Class I Federal areas. There
are 156 national parks, forests and wilderness areas categorized as Mandatory Class I Federal
areas (62 FR 38680-38681, July 18, 1997). These areas are defined in CAA section 162 as those
national parks exceeding 6,000 acres, wilderness areas and memorial parks exceeding 5,000
acres, and all international parks which were in existence on August 7, 1977. Figure 6-2 shows
the location of the 156 Mandatory Class I Federal areas.
6-19
-------
Produced by NPS Air Resources Division
* Rainbow Lake.WI and Brad well Bay, FL are Class 1 Areas
where visibility is not an important air quality related value
Figure 6-2 Mandatory Class I Federal Areas in the U.S.
EPA has also concluded that PIVb.s causes adverse effects on visibility in other areas that
are not protected by the Regional Haze Rule, depending on PM2.5 concentrations and other
factors that control their visibility impact effectiveness such as dry chemical composition and
relative humidity (i.e., an indicator of the water composition of the particles). EPA revised the
PM2.5 standards in December 2012 and established a target level of protection that is expected to
be met through attainment of the existing secondary standards for PM2.5.
6.2.1.1 Vi sibility Monitoring
In conjunction with the U.S. National Park Service, the U.S. Forest Service, other Federal
land managers, and State organizations in the U.S., the U.S. EPA has supported visibility
monitoring in national parks and wilderness areas since 1988. The monitoring network was
originally established at 20 sites, but it has now been expanded to 1 10 sites that represent all but
one of the 156 Mandatory Federal Class I areas across the country (see Figure 6-2). This long-
term visibility monitoring network is known as IMPROVE (Interagency Monitoring of Protected
Visual Environments).
IMPROVE provides direct measurement of fine particles that contribute to visibility
impairment. The IMPROVE network employs aerosol measurements at all sites, and optical and
scene measurements at some of the sites. Aerosol measurements are taken for PMio and PM2.5
mass, and for key constituents of PM2.5, such as sulfate, nitrate, organic and elemental carbon
OC and EC), soil dust, and several other elements. Measurements for specific aerosol
constituents are used to calculate "reconstructed" aerosol light extinction by multiplying the mass
for each constituent by its empirically-derived scattering and/or absorption efficiency, with
adjustment for the relative humidity. Knowledge of the main constituents of a site's light
extinction "budget" is critical for source apportionment and control strategy development. In
6-20
-------
addition to this indirect method of assessing light extinction, there are optical measurements
which directly measure light extinction or its components. Such measurements are made
principally with either a nephelometer to measure light scattering, some sites also include an
aethalometer for light absorption, or at a few sites using a transmissometer, which measures total
light extinction. Scene characteristics are typically recorded using digital or video photography
and are used to determine the quality of visibility conditions (such as effects on color and
contrast) associated with specific levels of light extinction as measured under both direct and
aerosol-related methods. Directly measured light extinction is used under the IMPROVE
protocol to cross check that the aerosol-derived light extinction levels are reasonable in
establishing current visibility conditions. Aerosol-derived light extinction is used to document
spatial and temporal trends and to determine how changes in atmospheric constituents would
affect future visibility conditions.
Annual average visibility conditions (reflecting light extinction due to both anthropogenic
and non-anthropogenic sources) vary regionally across the U.S. Visibility is typically worse in
the summer months and the rural East generally has higher levels of impairment than remote
sites in the West. Figures 9-9 through 9-11 in the PM ISA detail the percent contributions to
particulate light extinction for ammonium nitrate and sulfate, EC and OC, and coarse mass and
1 9^
fine soil, by season.
6.2.2 Particulate Matter Deposition
Particulate matter contributes to adverse effects on vegetation and ecosystems, and to
soiling and materials damage. These welfare effects result predominately from exposure to
excess amounts of specific chemical species, regardless of their source or predominant form
(particle, gas or liquid). The following characterizations of the nature of these environmental
effects are based on information contained in the 2009 PM ISA and the 2005 PM Staff Paper as
well as the 2008 Integrated Science Assessment for Oxides of Nitrogen and Sulfur- Ecological
Criteria (secondary NOX/SOX ISA).124'125'126
6.2.2.1 Deposition of Nitrogen and Sulfur
Nitrogen and sulfur interactions in the environment are highly complex. Both nitrogen
and sulfur are essential, and sometimes limiting, nutrients needed for growth and productivity of
ecosystem components (e.g. algae, plants). In terrestrial and aquatic ecosystems excesses of
nitrogen or sulfur can lead to acidification and nutrient enrichment.127
The process of acidification affects both freshwater aquatic and terrestrial ecosystems.
Acid deposition causes acidification of sensitive surface waters. The effects of acid deposition
on aquatic systems depend largely upon the ability of the ecosystem to neutralize the additional
acid. As acidity increases, aluminum leached from soils, flows into lakes and streams and can be
toxic to both terrestrial and aquatic biota. The lower pH concentrations and higher aluminum
levels resulting from acidification make it difficult for some fish and other aquatic organisms to
survive, grow, and reproduce. Research on effects of acid deposition on forest ecosystems has
come to focus increasingly on the biogeochemical processes that affect uptake, retention, and
cycling of nutrients within these ecosystems. Decreases in available base cations from soils are
at least partly attributable to acid deposition. Base cation depletion is a cause for concern
6-21
-------
because of the role these ions play in acid neutralization, and because calcium, magnesium and
potassium are essential nutrients for plant growth and physiology. Changes in the relative
proportions of these nutrients, especially in comparison with aluminum concentrations, have
been associated with declining forest health.
At current ambient levels, risks to vegetation from short-term exposures to dry deposited
particulate nitrate or sulfate are low. However, when found in acid or acidifying deposition, such
particles do have the potential to cause direct leaf injury. Specifically, the responses of forest
trees to acid precipitation (rain, snow) include accelerated weathering of leaf cuticular surfaces,
increased permeability of leaf surfaces to toxic materials, water, and disease agents; increased
leaching of nutrients from foliage; and altered reproductive processes—all which serve to
weaken trees so that they are more susceptible to other stresses (e.g., extreme weather, pests,
pathogens). Acid deposition with levels of acidity associated with the leaf effects described
above are currently found in some locations in the eastern U.S.128 Even higher concentrations of
acidity can be present in occult depositions (e.g., fog, mist or clouds) which more frequently
impacts higher elevations. Thus, the risk of leaf injury occurring from acid deposition in some
areas of the eastern U.S. is high. Nitrogen deposition has also been shown to impact ecosystems
in the western U.S. For example, a study conducted in the Columbia River Gorge National
Scenic Area (CRGNSA), located along a portion of the Oregon/Washington border, indicates
that lichen communities in the CRGNSA have shifted to a higher proportion of nitrophilous
species and the nitrogen content of lichen tissue is elevated.129 Lichens are sensitive indicators
of nitrogen deposition effects to terrestrial ecosystems and the lichen studies in the Columbia
River Gorge clearly show that ecological effects from air pollution are occurring.
Some of the most significant detrimental effects associated with excess nitrogen
deposition are those associated with a condition known as nitrogen saturation. Nitrogen
saturation is the condition in which nitrogen inputs from atmospheric deposition and other
sources exceed the biological requirements of the ecosystem. The effects associated with
nitrogen saturation include: (1) decreased productivity, increased mortality, and/or shifts in plant
community composition, often leading to decreased biodiversity in many natural habitats
wherever atmospheric reactive nitrogen deposition increases significantly above background and
critical thresholds are exceeded; (2) leaching of excess nitrate and associated base cations from
soils into streams, lakes, and rivers, and mobilization of soil aluminum; and (3) fluctuation of
ecosystem processes such as nutrient and energy cycles through changes in the functioning and
species composition of beneficial soil organisms.1 °
In the U.S. numerous forests now show severe symptoms of nitrogen saturation. These
forests include: the northern hardwoods and mixed conifer forests in the Adirondack and
Catskill Mountains of New York; the red spruce forests at Whitetop Mountain, Virginia, and
Great Smoky Mountains National Park, North Carolina; mixed hardwood watersheds at Fernow
Experimental Forest in West Virginia; American beech forests in Great Smoky Mountains
National Park, Tennessee; mixed conifer forests and chaparral watersheds in southern California
and the southwestern Sierra Nevada in Central California; the alpine tundra/subalpine conifer
forests of the Colorado Front Range; and red alder forests in the Cascade Mountains in
Washington.
6-22
-------
Excess nutrient inputs into aquatic ecosystems (i.e. streams, rivers, lakes, estuaries or
oceans) either from direct atmospheric deposition, surface runoff, or leaching from nitrogen
saturated soils into ground or surface waters can contribute to conditions of severe water oxygen
depletion; eutrophication and algae blooms; altered fish distributions, catches, and physiological
states; loss of biodiversity; habitat degradation; and increases in the incidence of disease.
Atmospheric deposition of nitrogen is a significant source of total nitrogen to many
estuaries in the United States. The amount of nitrogen entering estuaries that is ultimately
attributable to atmospheric deposition is not well-defined. On an annual basis, atmospheric
nitrogen deposition may contribute significantly to the total nitrogen load, depending on the size
and location of the watershed. In addition, episodic nitrogen inputs, which may be ecologically
important, may play a more important role than indicated by the annual average concentrations.
Estuaries in the U.S. that suffer from nitrogen enrichment often experience a condition known as
eutrophi cation. Symptoms of eutrophication include changes in the dominant species of
phytoplankton, low levels of oxygen in the water column, fish and shellfish kills, outbreaks of
toxic alga, and other population changes which can cascade throughout the food web. In
addition, increased phytoplankton growth in the water column and on surfaces can attenuate light
causing declines in submerged aquatic vegetation, which serves as an important habitat for many
estuarine fish and shellfish species.
Severe and persistent eutrophication often directly impacts human activities. For
example, losses in the nation's fishery resources may be directly caused by fish kills associated
with low dissolved oxygen and toxic blooms. Declines in tourism occur when low dissolved
oxygen causes noxious smells and floating mats of algal blooms create unfavorable aesthetic
conditions. Risks to human health increase when the toxins from algal blooms accumulate in
edible fish and shellfish, and when toxins become airborne, causing respiratory problems due to
inhalation. According to a National Oceanic and Atmospheric Administration report, more than
half of the nation's estuaries have moderate to high expressions of at least one of these symptoms
- an indication that eutrophication is well developed in more than half of U.S. estuaries. 31
6.2.2.2 Deposition of Heavy Metals
Heavy metals, including cadmium, copper, lead, chromium, mercury, nickel and zinc,
have the greatest potential for impacting forest growth.132 Investigation of trace metals near
roadways and industrial facilities indicate that a substantial load of heavy metals can accumulate
on vegetative surfaces. Copper, zinc, and nickel have been documented to cause direct toxicity
to vegetation under field conditions. Trace metals are associated with ambient PM, however
little research has been conducted on the effects associated with mixtures of contaminants.
While metals typically exhibit low solubility, limiting their bioavailability and direct toxicity,
chemical transformations of metal compounds occur in the environment, particularly in the
presence of acidic or other oxidizing species. These chemical changes influence the mobility
and toxicity of metals in the environment. Once taken up into plant tissue, a metal compound
can undergo chemical changes, exert toxic effects on the plant itself, accumulate and be passed
along to herbivores or can re-enter the soil and further cycle in the environment. Although there
has been no direct evidence of a physiological association between tree injury and heavy metal
exposures, heavy metals have been implicated because of similarities between metal deposition
patterns and forest decline. This hypothesized relationship/correlation was further explored in
6-23
-------
high elevation forests in the northeastern U.S. These studies measured levels of a group of
intracellular compounds found in plants that bind with metals and are produced by plants as a
response to sublethal concentrations of heavy metals. These studies indicated a systematic and
significant increase in concentrations of these compounds associated with the extent of tree
injury. These data strongly imply that metal stress causes tree injury and contributes to forest
decline in the northeastern United States.133 Contamination of plant leaves by heavy metals can
lead to elevated soil levels. Trace metals absorbed into the plant frequently bind to the leaf
tissue, and then are lost when the leaf drops. As the fallen leaves decompose, the heavy metals
are transferred into the soil.134'135 Upon entering the soil environment, PM pollutants can alter
ecological processes of energy flow and nutrient cycling, inhibit nutrient uptake, change
ecosystem structure, and affect ecosystem biodiversity. Many of the most important effects
occur in the soil. The soil environment is one of the most dynamic sites of biological interaction
in nature. It is inhabited by microbial communities of bacteria, fungi, and actinomycetes. These
organisms are essential participants in the nutrient cycles that make elements available for plant
uptake. Changes in the soil environment that influence the role of the bacteria and fungi in
nutrient cycling determine plant and ultimately ecosystem response.136
The environmental sources and cycling of mercury are currently of particular concern due
to the bioaccumulation and biomagnification of this metal in aquatic ecosystems and the potent
toxic nature of mercury in the forms in which is it ingested by people and other animals.
Mercury is unusual compared with other metals in that it largely partitions into the gas phase (in
elemental form), and therefore has a longer residence time in the atmosphere than a metal found
predominantly in the particle phase. This property enables mercury to travel far from the
primary source before being deposited and accumulating in the aquatic ecosystem. The major
source of mercury in the Great Lakes is from atmospheric deposition, accounting for
approximately eighty percent of the mercury in Lake Michigan.137'138 Over fifty percent of the
mercury in the Chesapeake Bay has been attributed to atmospheric deposition.139 Overall, the
National Science and Technology Council identifies atmospheric deposition as the primary
source of mercury to aquatic systems.140 Forty-four states have issued health advisories for the
consumption offish contaminated by mercury; however, most of these advisories are issued in
areas without a mercury point source.
Elevated levels of zinc and lead have been identified in streambed sediments, and these
elevated levels have been correlated with population density and motor vehicle use.141'142'143
Zinc and nickel have also been identified in urban water and soils. In addition, platinum,
palladium, and rhodium, metals found in the catalysts of modern motor vehicles, have been
measured at elevated levels along roadsides.144 Plant uptake of platinum has been observed at
these locations.
6.2.2.3 Deposition of Polycyclic Organic Matter
Polycyclic organic matter (POM) is a byproduct of incomplete combustion and consists
of organic compounds with more than one benzene ring and a boiling point greater than or equal
to 100 degrees centigrade.145 Polycyclic aromatic hydrocarbons (PAHs) are a class of POM that
contains compounds which are known or suspected carcinogens.
6-24
-------
Major sources of PAHs include mobile sources. PAHs in the environment may be
present as a gas or adsorbed onto airborne parti culate matter. Since the majority of PAHs are
adsorbed onto particles less than 1.0 jim in diameter, long range transport is possible. However,
studies have shown that PAH compounds adsorbed onto diesel exhaust particulate and exposed
to ozone have half lives of 0.5 to 1.0 hours.146
Since PAHs are insoluble, the compounds generally are particle reactive and accumulate
in sediments. Atmospheric deposition of particles is believed to be the major source of PAHs to
the sediments of Lake Michigan.147'148 Analyses of PAH deposition in Chesapeake and
Galveston Bay indicate that dry deposition and gas exchange from the atmosphere to the surface
water predominate.149'150 Sediment concentrations of PAHs are high enough in some segments
of Tampa Bay to pose an environmental health threat. EPA funded a study to better characterize
the sources and loading rates for PAHs into Tampa Bay.151 PAHs that enter a water body
through gas exchange likely partition into organic rich particles and can be biologically recycled,
while dry deposition of aerosols containing PAHs tend to be more resistant to biological
recycling.152 Thus, dry deposition is likely the main pathway for PAH concentrations in
sediments while gas/water exchange at the surface may lead to PAH distribution into the food
web, leading to increased health risk concerns.
Trends in PAH deposition levels are difficult to discern because of highly variable
ambient air concentrations, lack of consistency in monitoring methods, and the significant
influence of local sources on deposition levels.153 Van Metre et al. noted PAH concentrations in
urban reservoir sediments have increased by 200-300 percent over the last forty years and
correlate with increases in automobile use.154
Cousins et al. estimate that more than ninety percent of semi-volatile organic compound
(SVOC) emissions in the United Kingdom deposit on soil.155 An analysis of PAH concentrations
near a Czechoslovakian roadway indicated that concentrations were thirty times greater than
background.156
6.2.2.4 Materials Damage and Soiling
The effects of the deposition of atmospheric pollution, including ambient PM, on
materials are related to both physical damage and impaired aesthetic qualities. The deposition of
PM (especially sulfates and nitrates) can physically affect materials, adding to the effects of
natural weathering processes, by potentially promoting or accelerating the corrosion of metals,
by degrading paints, and by deteriorating building materials such as concrete and limestone.
Only chemically active fine particles or hygroscopic coarse particles contribute to these physical
effects. In addition, the deposition of ambient PM can reduce the aesthetic appeal of buildings
and culturally important articles through soiling. Particles consisting primarily of carbonaceous
compounds cause soiling of commonly used building materials and culturally important items
such as statues and works of art.
6.2.3 Plant and Ecosystem Effects of Ozone
The welfare effects of ozone can be observed across a variety of scales, i.e., subcellular,
cellular, leaf, whole plant, population and ecosystem. Ozone effects that begin at small spatial
6-25
-------
scales, such as the leaf of an individual plant, when they occur at sufficient magnitudes (or to a
sufficient degree) can result in effects being propagated along a continuum to larger and larger
spatial scales. For example, effects at the individual plant level, such as altered rates of leaf gas
exchange, growth and reproduction can, when widespread, result in broad changes in
ecosystems, such as productivity, carbon storage, water cycling, nutrient cycling, and community
composition.
Ozone can produce both acute and chronic injury in sensitive species depending on the
concentration level and the duration of the exposure.157 In those sensitive species158, effects from
repeated exposure to ozone throughout the growing season of the plant tend to accumulate, so
that even low concentrations experienced for a longer duration have the potential to create
chronic stress on vegetation.159 Ozone damage to sensitive species includes impaired
photosynthesis and visible injury to leaves. The impairment of photosynthesis, the process by
which the plant makes carbohydrates (its source of energy and food), can lead to reduced crop
yields, timber production, and plant productivity and growth. Impaired photosynthesis can also
lead to a reduction in root growth and carbohydrate storage below ground, resulting in other,
more subtle plant and ecosystems impacts.160 These latter impacts include increased
susceptibility of plants to insect attack, disease, harsh weather, interspecies competition and
overall decreased plant vigor. The adverse effects of ozone on areas with sensitive species could
potentially lead to species shifts and loss from the affected ecosystems161, resulting in a loss or
reduction in associated ecosystem goods and services. Additionally, visible ozone injury to
leaves can result in a loss of aesthetic value in areas of special scenic significance like national
1 f\'J
parks and wilderness areas and reduced use of sensitive ornamentals in landscaping.
The Integrated Science Assessment (ISA) for Ozone presents more detailed information
on how ozone effects vegetation and ecosystems.163 The ISA concludes that ambient
concentrations of ozone are associated with a number of adverse welfare effects and
characterizes the weight of evidence for different effects associated with ozone.164 The ISA
concludes that visible foliar injury effects on vegetation, reduced vegetation growth, reduced
productivity in terrestrial ecosystems, reduced yield and quality of agricultural crops, and
alteration of below-ground biogeochemical cycles are causally associated with exposure to
ozone. It also concludes that reduced carbon sequestration in terrestrial ecosystems, alteration of
terrestrial ecosystem water cycling, and alteration of terrestrial community composition are
likely to be causally associated with exposure to ozone.
6.2.4 Environmental Effects of Air Toxics
Emissions from producing, transporting and combusting fuel contribute to ambient levels
of pollutants that contribute to adverse effects on vegetation. Volatile organic compounds
(VOCs), some of which are considered air toxics, have long been suspected to play a role in
vegetation damage.165 In laboratory experiments, a wide range of tolerance to VOCs has been
observed.166 Decreases in harvested seed pod weight have been reported for the more sensitive
plants, and some studies have reported effects on seed germination, flowering and fruit ripening.
Effects of individual VOCs or their role in conjunction with other stressors (e.g., acidification,
drought, temperature extremes) have not been well studied. In a recent study of a mixture of
VOCs including ethanol and toluene on herbaceous plants, significant effects on seed production,
leaf water content and photosynthetic efficiency were reported for some plant species. 6?
6-26
-------
Research suggests an adverse impact of vehicle exhaust on plants, which has in some
cases been attributed to aromatic compounds and in other cases to nitrogen oxides.168'169'170 The
impacts of VOCs on plant reproduction may have long-term implications for biodiversity and
survival of native species near major roadways. Most of the studies of the impacts of VOCs on
vegetation have focused on short-term exposure and few studies have focused on long-term
effects of VOCs on vegetation and the potential for metabolites of these compounds to affect
herbivores or insects.
6-27
-------
References
1 U.S. Environmental Protection Agency. (2005). Supplemental guidance for assessing susceptibility
from early-life exposure to carcinogens. Washington, DC: Risk Assessment Forum. EPA/630/
R-03/003F. http://www.epa.gov/raf/publications/pdfs/childrens_supplement_final.pdf.
2 Regulatory definitions of PM size fractions, and information on reference and equivalent methods for measuring
PM in ambient air, are provided in 40 CFR Parts 50, 53, and 58. With regard to national ambient air quality
standards (NAAQS) which provide protection against health and welfare effects, the 24-hour PM10 standard
provides protection against effects associated with short-term exposure to thoracic coarse particles (i.e., PM10.2 5).
3 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009.
4 U.S. EPA. (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F.
5 The causal framework draws upon the assessment and integration of evidence from across epidemiological,
controlled human exposure, and lexicological studies, and the related uncertainties that ultimately influence our
understanding of the evidence. This framework employs a five-level hierarchy that classifies the overall weight of
evidence and causality using the following categorizations: causal relationship, likely to be causal relationship,
suggestive of a causal relationship, inadequate to infer a causal relationship, and not likely to be a causal relationship
(U.S. EPA. (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, Table 1-3).
6 78 FR 3086 (January 15, 2013) pages 3103-3104.
7 77 FR 38890 (June 29, 2012) pages 38906-38911.
8 These causal inferences are based not only on the more expansive epidemiological evidence available in this
review but also reflect consideration of important progress that has been made to advance our understanding of a
number of potential biologic modes of action or pathways for PM-related cardiovascular and respiratory effects
(U.S. EPA. (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, chapter 5).
9 78 FR 3086 (January 15, 2013) pages 3103-3104.
10 U.S. EPA. (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, chapter 6 (section 6.5) and chapter 7 (Section 7.6).
11 78 FR 3103 (January 15, 2013).
12 78 FR 3103 (January 15, 2013).
13 78 FR 3103-3104 (January 15, 2013).
14 78 FR 3104 (January 15, 2013).
15 U.S. EPA. (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F. pg 2-13.
16 U.S. EPA. (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F. Chapter 8 and Chapter 2.
17 77 FR 38890,(June 29, 2012) page 38910.
18 78 FR 3086 (January 15, 2013) page 3104.
19 U.S. EPA. (2011). Policy Assessment for the Review of the PM NAAQS. U.S. Environmental Protection
Agency, Washington, DC, EPA/452/R-11-003. section 2.2.1.
20 U.S. EPA. (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F. Chapter 8 and Chapter 2 (Section 2.4.1).
6-28
-------
21 U.S. EPA. (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F. Section 2.3.4 and Table 2-6.
22 78 FR 3167-8 (January 15, 2013).
23 77 FR 38947-51 (June 29, 2012).
24
U.S. EPA. (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F. Section 2.3.5 and Table 2-6.
25 78 FR 3121 (January 15, 2013).
26 Human exposure to ozone varies over time due to changes in ambient ozone concentration and because people
move between locations which have notable different ozone concentrations. Also, the amount of ozone delivered to
the lung is not only influenced by the ambient concentrations but also by the individuals breathing route and rate.
27 U.S. EPA. Integrated Science Assessment of Ozone and Related Photochemical Oxidants (Final Report). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-10/076F, 2013. The ISA is available at
http://cfpub.epa. gov/ncea/isa/recordisplav.cfm?deid=247492#Download.
28 The ISA evaluates evidence and draws conclusions on the causal relationship between relevant pollutant
exposures and health effects, assigning one of five "weight of evidence" determinations: causal relationship, likely
to be a causal relationship, suggestive of a causal relationship, inadequate to infer a causal relationship, and not
likely to be a causal relationship. For more information on these levels of evidence, please refer to Table II in the
Preamble of the IS A.
29 U.S. EPA (2008). Integrated Science Assessment (ISA) for Sulfur Oxides - Health Criteria (Final Report).
EPA/600/R-08/047F. Washington, DC,: U.S.EPA. Retrieved on March 19, 2009 from
http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=198843.
30 U.S. EPA (2008). Integrated Science Assessment for Oxides of Nitrogen -Health Criteria (Final Report).
EPA/600/R-08/071. Washington, DC,: U.S.EPA. Retrieved on March 19, 2009 from
http ://cfpub .epa. gov/ncea/cfm/recordisplav. cfm?deid= 194645.
31 U.S. EPA (2008). Integrated Science Assessment for Oxides of Nitrogen -Health Criteria (Final Report).
EPA/600/R-08/071. Washington, DC: U.S.EPA.
32 U.S. EPA, (2010). Integrated Science Assessment for Carbon Monoxide (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-09/019F, 2010.
http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=218686.
33 The ISA evaluates the health evidence associated with different health effects, assigning one of five "weight of
evidence" determinations: causal relationship, likely to be a causal relationship, suggestive of a causal relationship,
inadequate to infer a causal relationship, and not likely to be a causal relationship. For definitions of these levels of
evidence, please refer to Section 1.6 of the ISA.
34 Personal exposure includes contributions from many sources, and in many different environments. Total
personal exposure to CO includes both ambient and nonambient components; and both components may contribute
to adverse health effects.
35 U.S. EPA. (2000). Integrated Risk Information System File for Benzene. This material is available electronically
at: http://www.epa.gov/iris/subst/0276.htm.
36 International Agency for Research on Cancer. (1982). IARC monographs on the evaluation of carcinogenic risk of
chemicals to humans, Volume 29, Some industrial chemicals and dyestuffs, International Agency for Research on
Cancer, World Health Organization, Lyon, France 1982.
37 Irons, R.D.; Stillman, W.S.; Colagiovanni, D.B.; Henry, V.A. (1992). Synergistic action of the benzene metabolite
hydroquinone on myelopoietic stimulating activity of granulocyte/macrophage colony-stimulating factor in vitro,
Proc. Natl. Acad. Sci. 89:3691-3695.
38 A unit risk estimate is defined as the increase in the lifetime risk of an individual who is exposed for a lifetime to
1 ug/m3 benzene in air.
6-29
-------
39 U.S. EPA. (2000). Integrated Risk Information System File for Benzene. This material is available electronically
at: http://www.epa.gov/iris/subst/0276.htm.
40 International Agency for Research on Cancer (IARC). 1987. Monographs on the evaluation of carcinogenic risk
of chemicals to humans, Volume 29, Supplement 7, Some industrial chemicals and dyestuffs, World Health
Organization, Lyon, France.
41 U.S. Department of Health and Human Services National Toxicology Program. (2011). 12th Report on
Carcinogens available at: http://ntp.niehs.nih.gov/?objectid=03C9AF75-ElBF-FF40-DBA9EC0928DF8B15.
42Aksoy, M. (1989). Hematotoxicity and carcinogenicity of benzene. Environ. Health Perspect. 82:193-197.
EP A-HQ-O AR-2011-0135.
43 Goldstein, B.D. (1988). Benzene toxicity. Occupational medicine. State of the Art Reviews. 3:541-554.
44Rothman, N., G.L. Li, M. Dosemeci, W.E. Bechtold, G.E. Marti, Y.Z. Wang, M. Linet, L.Q. Xi, W. Lu, M.T.
Smith, N. Titenko-Holland, L.P. Zhang, W. Blot, S.N. Yin, and R.B. Hayes. (1996). Hematotoxicity among Chinese
workers heavily exposed to benzene. Am. J. Ind. Med. 29: 236-246.
45U.S. EPA (2002). Toxicological Review of Benzene (Noncancer Effects). Environmental Protection Agency,
Integrated Risk Information System (IRIS), Research and Development, National Center for Environmental
Assessment, Washington DC. This material is available electronically at http://www.epa.gov/iris/subst/0276.htm.
46Qu, O.; Shore, R.; Li, G.; Jin, X.; Chen, C.L.; Cohen, B.; Melikian, A.; Eastmond, D.; Rappaport, S.; Li, H.; Rupa,
D.; Suramaya, R.; Songnian, W.; Huifant, Y.; Meng, M.; Winnik, M.; Kwok, E.; Li, Y.; Mu, R.; Xu, B.; Zhang,
X.; Li, K. (2003). HEI Report 115, Validation & Evaluation of Biomarkers in Workers Exposed to Benzene in
China.
47 Qu, Q., R. Shore, G. Li, X. Jin, L.C. Chen, B. Cohen, et al. (2002). Hematological changes among Chinese
workers with a broad range of benzene exposures. Am. J. Industr. Med. 42: 275-285.
48Lan, Qing, Zhang, L., Li, G., Vermeulen, R., et al. (2004). Hematotoxically in Workers Exposed to Low Levels of
Benzene. Science 306: 1774-1776.
49 Turtletaub, K.W. and Mani, C. (2003). Benzene metabolism in rodents at doses relevant to human exposure from
Urban Air. Research Reports Health Effect Inst. Report No.113.
50 U.S. Agency for Toxic Substances and Disease Registry (ATSDR). (2007). Toxicological profile for benzene.
Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service.
http://www.atsdr.cdc.gov/ToxProfiles/tp3.pdf.
51 A minimal risk level (MRL) is defined as an estimate of the daily human exposure to a hazardous substance that is
likely to be without appreciable risk of adverse noncancer health effects over a specified duration of exposure.
52U.S. EPA. (2002). Health Assessment of 1,3-Butadiene. Office of Research and Development, National Center
for Environmental Assessment, Washington Office, Washington, DC. Report No. EPA600-P-98-001F. This
document is available electronically at http://www.epa.gov/iris/supdocs/buta-sup.pdf.
53U.S. EPA. (2002) "Full IRIS Summary for 1,3-butadiene (CASRN 106-99-0)" Environmental Protection Agency,
Integrated Risk Information System (IRIS), Research and Development, National Center for Environmental
Assessment, Washington, DC http://www.epa.gov/iris/subst/0139.htm.
54 International Agency for Research on Cancer (IARC). (1999). Monographs on the evaluation of carcinogenic risk
of chemicals to humans, Volume 71, Re-evaluation of some organic chemicals, hydrazine and hydrogen peroxide
and Volume 97 (in preparation), World Health Organization, Lyon, France.
55 International Agency for Research on Cancer (IARC). (2008). Monographs on the evaluation of carcinogenic risk
of chemicals to humans, 1,3-Butadiene, Ethylene Oxide and Vinyl Halides (Vinyl Fluoride, Vinyl Chloride and
Vinyl Bromide) Volume 97, World Health Organization, Lyon, France.
56 NTP. (2011). Report on Carcinogens, Twelfth Edition. Research Triangle Park, NC: U. S. Department of Health
and Human Services, Public Health Service, National Toxicology Program. 499 pp.
6-30
-------
57 U.S. EPA. (2002). "Full IRIS Summary for 1,3-butadiene (CASRN 106-99-0)" Environmental Protection Agency,
Integrated Risk Information System (IRIS), Research and Development, National Center for Environmental
Assessment, Washington, DC http://www.epa.gov/iris/subst/0139.htm.
58Bevan, C.; Stadler, J.C.; Elliot, G.S.; et al. (1996). Subchronic toxicity of 4-vinylcyclohexene in rats and mice by
inhalation. Fundam. Appl. Toxicol. 32:1-10.
59 EPA. Integrated Risk Information System. Formaldehyde (CASRN 50-00-0)
http://www.epa.gov/iris/subst/0419/htm.
60 National Toxicology Program, U.S. Department of Health and Human Services (HHS), 12th Report on
Carcinogens, June 10, 2011.
61IARC Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 88 (2006): Formaldehyde, 2-
Butoxyethanol and l-tert-Butoxypropan-2-ol.
62 IARC Mongraphs on the Evaluation of Carcinogenic Risks to Humans Volume 100F (2012): Formaldehyde.
63 Hauptmann, M..; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Blair, A. 2003. Mortality from lymphohematopoetic
malignancies among workers in formaldehyde industries. Journal of the National Cancer Institute 95: 1615-1623.
64Hauptmann, M..; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Blair, A. 2004. Mortality from solid cancers among
workers in formaldehyde industries. American Journal of Epidemiology 159: 1117-1130.
65Beane Freeman, L. E.; Blair, A.; Lubin, J. H.; Stewart, P. A.; Hayes, R. B.; Hoover, R. N.; Hauptmann, M. 2009.
Mortality from lymphohematopoietic malignancies among workers in formaldehyde industries: The National Cancer
Institute cohort. J. National Cancer Inst. 101: 751-761.
66 Pinkerton, L. E. 2004. Mortality among a cohort of garment workers exposed to formaldehyde: an update.
Occup. Environ. Med. 61: 193-200.
67 Coggon, D, EC Harris, J Poole, KT Palmer. 2003. Extended follow-up of a cohort of British chemical workers
exposed to formaldehyde. J National Cancer Inst. 95:1608-1615.
68 Hauptmann, M,; Stewart P. A.; Lubin J. H.; Beane Freeman, L. E.; Hornung, R. W.; Herrick, R. F.; Hoover, R. N;
Fraumeni, J. F.; Hayes, R. B. 2009. Mortality from lymphohematopoietic malignancies and brain cancer among
embalmers exposed to formaldehyde. Journal of the National Cancer Institute 101:1696-1708.
69 ATSDR. 1999. Toxicological Profile for Formaldehyde, U.S. Department of Health and Human Services (HHS),
July 1999.
70 ATSDR. 2010. Addendum to theToxicological Profile for Formaldehyde. U.S. Department of Health and Human
Services (HHS), October 2010.
71IPCS. 2002. Concise International Chemical Assessment Document 40. Formaldehyde. World Health
Organization.
72 EPA (U.S. Environmental Protection Agency). 2010. Toxicological Review of Formaldehyde (CAS No. 50-00-0)
- Inhalation Assessment: In Support of Summary Information on the Integrated Risk Information System (IRIS).
External Review Draft. EPA/635/R-10/002A. U.S. Environmental Protection Agency, Washington DC [online].
Available: http://cfpub.epa.gov/ncea/irs_drats/recordisplay.cfm?deid=223614.
73 NRC (National Research Council). 2011. Review of the Environmental Protection Agency's Draft IRIS
Assessment of Formaldehyde. Washington DC: National Academies Press.
http://books.nap.edu/openbook.php?record_id=13142.
74U.S. EPA (1991). Integrated Risk Information System File of Acetaldehyde. Research and Development,
National Center for Environmental Assessment, Washington, DC. This material is available electronically at
http://www.epa.gov/iris/subst/0290.htm.
75U.S. EPA (1991). Integrated Risk Information System File of Acetaldehyde. This material is available
electronically at http://www.epa.gov/iris/subst/0290.htm.
6-31
-------
76 NTP. (2011). Report on Carcinogens, Twelfth Edition. Research Triangle Park, NC: U.S. Department of Health
and Human Services, Public Health Service, National Toxicology Program. 499 pp.
77 International Agency for Research on Cancer (IARC). (1999). Re-evaluation of some organic chemicals,
hydrazine, and hydrogen peroxide. IARC Monographs on the Evaluation of Carcinogenic Risk of Chemical to
Humans, Vol 71. Lyon, France.
78 U.S. EPA (1991). Integrated Risk Information System File of Acetaldehyde. This material is available
electronically at http://www.epa.gov/iris/subst/0290.htm.
79 U.S. EPA. (2003). Integrated Risk Information System File of Acrolein. Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
http://www.epa.gov/iris/subst/0364.htm.
80 Appleman, L.M., R.A. Woutersen, and V.J. Feron. (1982). Inhalation toxicity of acetaldehyde in rats. I. Acute and
subacute studies. Toxicology. 23: 293-297.
81 Myou, S.; Fujimura, M; Nishi K.; Ohka, T.; and Matsuda, T. (1993). Aerosolized acetaldehyde induces
histamine-mediated bronchoconstriction in asthmatics. Am. Rev. Respir.Dis. 148(4 Pt 1): 940-943.
82 U.S. EPA. (2003). Integrated Risk Information System File of Acrolein. Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available at
http://www.epa.gov/iris/subst/0364.htm.
83 International Agency for Research on Cancer (IARC). (1995). Monographs on the evaluation of carcinogenic risk
of chemicals to humans, Volume 63. Dry cleaning, some chlorinated solvents and other industrial chemicals, World
Health Organization, Lyon, France.
84 U.S. EPA. (2003). Integrated Risk Information System File of Acrolein. Office of Research and Development,
National Center for Environmental Assessment, Washington, DC. This material is available at
http://www.epa.gov/iris/subst/0364.htm.
85 U.S. EPA. (2003). Integrated Risk Information System File of Acrolein. Office of Research and Development,
National Center for Environmental Assessment, Washington, DC. This material is available at
http://www.epa.gov/iris/subst/0364.htm.
86 U.S. EPA. (2003). Toxicological review of acrolein in support of summary information on Integrated Risk
Information System (IRIS) National Center for Environmental Assessment, Washington, DC. EPA/635/R-03/003. p.
10. Available online at: http://www.epa.gov/ncea/iris/toxreviews/0364tr.pdf.
87 U.S. EPA. (2003). Toxicological review of acrolein in support of summary information on Integrated Risk
Information System (IRIS) National Center for Environmental Assessment, Washington, DC. EPA/635/R-03/003.
Available online at: http://www.epa.gov/ncea/iris/toxreviews/0364tr.pdf.
88 Morris JB, Symanowicz PT, Olsen JE, et al. (2003). Immediate sensory nerve-mediated respiratory responses to
irritants in healthy and allergic airway-diseased mice. J Appl Physiol 94(4): 1563-1571.
89 U.S. EPA. (2009). Graphical Arrays of Chemical-Specific Health Effect Reference Values for Inhalation
Exposures (Final Report). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-09/061, 2009.
http://cfpub.epa. gov/ncea/cfm/recordisplay.cfm?deid=211003.
90U.S. EPA. (2006). Air quality criteria for ozone and related photochemical oxidants (Ozone CD). Research
Triangle Park, NC: National Center for Environmental Assessment; report no. EPA/600/R-05/004aF-cF.3v. page 5-
78 Available at http://cfpub.epa.gov/ncea/.
91 U.S. EPA Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. Environmental
Protection Agency, Washington, D.C., EPA 600/R-05/004aF-cF, 2006. page 5-63. This document is available in.
This document may be accessed electronically at: http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_cd.html.
92U.S. EPA Air Quality Criteria for Ozone and Related Photochemical Oxidants (Final). U.S. Environmental
Protection Agency, Washington, D.C., EPA 600/R-05/004aF-cF, 2006. page 5-78. This document is available in.
This document may be accessed electronically at: http://www.epa.gov/ttn/naaqs/standards/ozone/s_o3_cr_cd.html.
6-32
-------
93 Agency for Toxic Substances and Disease Registry (ATSDR). (1995). lexicological profile for Polycyclic
Aromatic Hydrocarbons (PAHs). Atlanta, GA: U.S. Department of Health and Human Services, Public Health
Service. Available electronically athttp://www.atsdr.cdc.gov/ToxProfiles/TP.asp?id=122&tid=25.
94 U.S. EPA (2002). Health Assessment Document for Diesel Engine Exhaust. EPA/600/8-90/057F Office of
Research and Development, Washington DC. http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=29060.
95 International Agency for Research on Cancer (IARC). (2012). Monographs on the Evaluation of the
Carcinogenic Risk of Chemicals for Humans, Chemical Agents and Related Occupations. Vol. 100F. Lyon, France.
96U.S. EPA (1997). Integrated Risk Information System File of indeno(l,2,3-cd)pyrene. Research and Development,
National Center for Environmental Assessment, Washington, DC. This material is available electronically at
http://www.epa.gov/ncea/iris/subst/0457.htm.
97 Perera, P.P.; Rauh, V.; Tsai, W-Y.; et al. (2002). Effect of transplacental exposure to environmental pollutants on
birth outcomes in a multiethnic population. Environ Health Perspect. Ill: 201 -205.
98 Perera, P.P.; Rauh, V.; Whyatt, R.M.; Tsai, W.Y.; Tang, D.; Diaz, D.; Hoepner, L.; Barr, D.; Tu, Y.H.; Camann,
D.; Kinney, P. (2006). Effect of prenatal exposure to airborne polycyclic aromatic hydrocarbons on
neurodevelopment in the first 3 years of life among inner-city children. Environ Health Perspect 114: 1287-1292.
99 U. S. EPA. 1998. Toxicological Review of Naphthalene (Reassessment of the Inhalation Cancer Risk),
Environmental Protection Agency, Integrated Risk Information System, Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
http://www.epa.gov/iris/subst/0436.htm.
100 U. S. EPA. 1998. Toxicological Review of Naphthalene (Reassessment of the Inhalation Cancer Risk),
Environmental Protection Agency, Integrated Risk Information System, Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
http://www.epa.gov/iris/subst/0436.htm.
101 U. S. EPA. (1998). Toxicological Review of Naphthalene (Reassessment of the Inhalation Cancer Risk),
Environmental Protection Agency, Integrated Risk Information System, Research and Development, National
Center for Environmental Assessment, Washington, DC. This material is available electronically at
http://www.epa.gov/iris/subst/0436.htm.
102 Oak Ridge Institute for Science and Education. (2004). External Peer Review for the IRIS Reassessment of the
Inhalation Carcinogenicity of Naphthalene. August 2004.
http://cfpub.epa. gov/ncea/cfm/recordisplav.cfm?deid=84403.
103 N-pp. (2011). Report on Carcinogens, Twelfth Edition. Research Triangle Park, NC: U.S. Department of Health
and Human Services, Public Health Service, National Toxicology Program. 499 pp.
104 International Agency for Research on Cancer (IARC). (2002). Monographs on the Evaluation of the
Carcinogenic Risk of Chemicals for Humans. Vol. 82. Lyon, France.
105 U. S. EPA. (1998). Toxicological Review of Naphthalene, Environmental Protection Agency, Integrated Risk
Information System, Research and Development, National Center for Environmental Assessment, Washington, DC.
This material is available electronically at http://www.epa.gov/iris/subst/0436.htm.
106 U.S. EPA. (1998). Toxicological Review of Naphthalene. Environmental Protection Agency, Integrated Risk
Information System (IRIS), Research and Development, National Center for Environmental Assessment,
Washington, DC http://www.epa.gov/iris/subst/0436.htm.
107 U.S. EPA Integrated Risk Information System (IRIS) database is available at: www.epa.gov/iris.
108 It should be noted that there are no peer reviewed EPA-authored reviews of traffic-related health studies. The
HEI panel primarily used epidemiology studies for inferring whether there was sufficient evidence of a causal
association exists between a particular health effect and traffic-related air pollution, In its weight-of-evidence
determinations, the panel also placed "considerable weight" on controlled human exposure studies. However, it
restricted consideration of other lexicological studies to whether or not the studies provided "general mechanistic
support" for the inferences of causality made on the basis of epidemiology.
6-33
-------
109 Health Effects Institute Panel on the Health Effects of Traffic-Related Air Pollution. (2010). Traffic-related air
pollution: a critical review of the literature on emissions, exposure, and health effects. HEI Special Report 17.
[Online at http://www.healtheffects.org].
110 Bureau of Labor Statistics. (2009). American Time Use Survey. [Online at http://www.bls.gov/tus].
111 Bureau of Transportation Statistics. (2003). Highlights of the 2001 National Household Travel Survey. Report
BTS03-05. [Online at http://www.bts.gov].
112 The autonomic nervous system (ANS) consists of sympathetic and parasympathetic components. The
sympathetic ANS signals body systems to "fight or flight." The parasympathetic ANS signals the body to "rest and
digest." In general, HRV is indicative of parasympathetic control of the heart.
113 Baja, E.S.; Schwartz, J.D.; Wellenius, G.A.; Coull, B.A.; Zanobetti, A.; Vokonas, P.S.; Suh, H.H. (2010).
Traffic-related air pollution and QT interval: modification by diabetes, obesity, and oxidate stress gene
polymorphisms in the Normative Aging Study. Environ Health Perspect 118: 840-846. doi:10.1289/ehp.0901396.
114 Zanobettia, A.; Stone, P.H.; Speizer, F.E.; Schwarz, J.D.; Coull, B.A.; Suh, H.H.; Nearing, B.D.; Mittleman,
M.A.; Verrier, R.L.; Gold, D.R. (2009). T-wave alterans, air pollution and traffic in high-risk subjects. Am J
Cardiol 104: 665-670. doi:10.1016/j.amjcard.2009.04.046.
115 Brook, R.D.; Rajagopalan, S.; Pope, C.A.; Brook, J.R.; Bhatnagar, A.; Diez-Rouz, A.V.; Holguin, F.; Hong, Y.;
Luepker, R.V.; et la. (2010). Paniculate matter air pollution and cardiovascular disease: an update to the scientific
statement from the American Heart Association. Circulation 121: 2331 -2378.
doi:10.1161/CIR.Ob013e3181dbecel.
116 Bastain, T.M.; Gilliland, F.D.; Li, Y.; Saxon, A.; Diaz-Sanchez, D. (2003) Intraindividual reproducibility of nazal
allergic responses to diesel exhaust particles indicates a susceptible phenotype. Clinical Immunol 109: 130-136.
117 Gilliland, F.D.; Li, Y.; Diaz-Sanchez, D. (2004) Effect of glutathione-S-transferase Ml and PI genotypes on
xenobiotic enhancement of allergic responses: randomized, placebo-controlled crossover study. Lancet 363:119-
125.
118 Svartengren, M, Strand, V.; Bylin, G. Ja'rup, L.; Pershagen, G. (2000) Short-term exposure to air pollution in a
road tunnel enhances the asthmatic response to allergen. Eur Respir J 15: 716-724.
119 Vrijheid, M.; Martinez, D.; Manzanares, S.; Dadvand, P.; Schembari, A.; Rankin, F.; Nieuwenhuijsen, M. (2011).
Ambient air pollution and risk of congenital anomalies: a systematic review and meta-analysis. Environ Health
Perspect 119: 598-606. doi: 10.1289/ehp. 1002946.
120 Raaschou-Nielson, O.; Reynolds, P. (2006). Air pollution and childhood cancer: a review of the epidemiological
literature. IntJ Cancer 118: 2920-2929. doi: 10.1002/ijc.21787 [Online at http://dx.doi.org].
121 National Research Council, (1993). Protecting Visibility in National Parks and Wilderness Areas. National
Academy of Sciences Committee on Haze in National Parks and Wilderness Areas. National Academy Press,
Washington, DC. This book can be viewed on the National Academy Press Website at
http://www.nap.edu/books/0309048443/html/.
122 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009.
123 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009. pg 9-19 through 9-23.
124 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009.
125 U.S. EPA. (2005). Review of the National Ambient Air Quality Standard for Paniculate Matter: Policy
Assessment of Scientific and Technical Information, OAQPS Staff Paper. EPA-452/R-05-005.
126 U.S. EPA. (2008). Integrated Science Assessment for Oxides of Nitrogen and Sulfur- Ecological Criteria (Final).
U.S. EPA, WashingtonD.C., EPA/600/R-08/082F.
6-34
-------
127 U.S. EPA. (2008). Integrated Science Assessment for Oxides of Nitrogen and Sulfur- Ecological Criteria (Final).
U.S. EPA, WashingtonD.C., EPA/600/R-08/082F.
128 Environmental Protection Agency (2003). Response Of Surface Water Chemistry to the Clean Air Act
Amendments of 1990. National Health and Environmental Effects Research Laboratory, Office of Research and
Development, U.S. Environmental Protection Agency. Research Triangle Park, NC. EPA 620/R-03/001.
129 Fenn, M.E. and Blubaugh, TJ. (2005). Winter Deposition of Nitrogen and Sulfur in the Eastern Columbia River
Gorge National Scenic Area, USDA Forest Service.
130 Galloway, J. N.; Cowling, E. B. (2002). Reactive nitrogen and the world: 200 years of change. Ambio 31: 64-71.
131 Bricker, Suzanne B., et al., (1999). National Estuarine Eutrophication Assessment, Effects of Nutrient
Enrichment in the Nation's Estuaries, National Ocean Service, National Oceanic and Atmospheric Administration,
September, 1999.
132 Smith, W.H. (1991). "Air pollution and Forest Damage." Chemical Engineering News, 69(45): 30-43.
133 Gawel, J.E.; Ahner, B.A.; Friedland, A.J.; and Morel, F.M.M. (1996). "Role for heavy metals in forest decline
indicated by phytochelatin measurements." Nature, 381: 64-65.
134 Cotrufo, M.F.; DeSanto, A.V.; Alfani, A.; et al. (1995). "Effects of urban heavy metal pollution on organic matter
decomposition in Quercus ilix L. woods." Environmental Pollution, 89: 81-87.
135Niklinska, M; Laskowski, R.; Maryanski, M. (1998). "Effect of heavy metals and storage time on two types of
forest litter: basal respiration rate and exchangeable metals." Ecotoxicological Environmental Safety, 41: 8-18.
136 U.S. EPA (2009). Integrated Science Assessment for Paniculate Matter (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-08/139F, 2009. Section 9.4.5.2.
137 Mason, R.P. and Sullivan, K. A. (1997). "Mercury in Lake Michigan." Environmental Science & Technology, 31:
942-947. (from Delta Report "Atmospheric deposition of toxics to the Great Lakes").
138 Landis, M.S. and Keeler, G. J. (2002). "Atmospheric mercury deposition to Lake Michigan during the Lake
Michigan Mass Balance Study." Environmental Science & Technology, 21: 4518-24.
139 U.S. EPA. (2000). EPA453/R-00-005, "Deposition of Air Pollutants to the Great Waters: Third Report to
Congress," Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
140 National Science and Technology Council (NSTC) (1999). "The Role of Monitoring Networks in the
Management of the Nation's Air Quality."
141 Callender, E. and Rice, K.C. (2000). "The Urban Environmental Gradient: Anthropogenic Influences on the
Spatial and Temporal Distributions of Lead and Zinc in Sediments." Environmental Science & Technology, 34: 232-
238.
142 Rice, K.C. (1999). "Trace Element Concentrations in Streambed Sediment Across the Conterminous United
States." Environmental Science & Technology, 33: 2499-2504.
143 U.S. EPA. (2013) Integrated Science Assessment for Lead (Final Report). U.S. Environmental Protection
Agency, Washington, DC, EPA/600/R-10/075F, 2013.
144 Ely, JC; Neal, CR; Kulpa, CF; et al. (2001). "Implications of Platinum-Group Element Accumulation along U.S.
Roads from Catalytic-Converter Attrition." Environ. Sci. Technol. 35: 3816-3822.
145 U.S. EPA. (1998). EPA454/R-98-014, "Locating and Estimating Air Emissions from Sources of Fob/cyclic
Organic Matter," Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
146U.S. EPA. (1998). EPA454/R-98-014, "Locating and Estimating Air Emissions from Sources of Fob/cyclic
Organic Matter," Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
147 Simcik, M.F.; Eisenreich, S.J.; Golden, K.A.; et al. (1996). "Atmospheric Loading of Fob/cyclic Aromatic
Hydrocarbons to Lake Michigan as Recorded in the Sediments." Environmental Science and Technology, 30: 3039-
3046.
6-35
-------
148 Simcik, M.F.; Eisenreich, S.J.; and Lioy, P.J. (1999). "Source apportionment and source/sink relationship of
PAHs in the coastal atmosphere of Chicago and Lake Michigan." Atmospheric Environment, 33: 5071-5079.
149 Arzayus, K.M.; Dickhut, R.M.; and Canuel, E.A. (2001). "Fate of Atmospherically Deposited Polycyclic
Aromatic Hydrocarbons (PAHs) in Chesapeake Bay." Environmental Science & Technology, 35, 2178-2183.
150Park, J.S.; Wade, T.L.; and Sweet, S. (2001). "Atmospheric distribution of polycyclic aromatic hydrocarbons and
deposition to Galveston Bay, Texas, USA." Atmospheric Environment, 35: 3241-3249.
151 Poor, N.; Tremblay, R.; Kay, H.; et al. (2002). "Atmospheric concentrations and dry deposition rates of
polycyclic aromatic hydrocarbons (PAHs) for Tampa Bay, Florida, USA." Atmospheric Environment 38: 6005-
6015.
152 Arzayus, K.M.; Dickhut, R.M.; and Canuel, E.A. (2001). "Fate of Atmospherically Deposited Polycyclic
Aromatic Hydrocarbons (PAHs) in Chesapeake Bay." Environmental Science & Technology, 35, 2178-2183.
153 U.S. EPA. (2000). EPA453/R-00-005, "Deposition of Air Pollutants to the Great Waters: Third Report to
Congress," Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
154 Van Metre, P.C.; Mahler, B.I; and Furlong, E.T. (2000). "Urban Sprawl Leaves its PAH Signature."
Environmental Science & Technology, 34: 4064-4070.
155 Cousins, I.T.; Beck, A.J.; and Jones, K.C. (1999). "A review of the processes involved in the exchange of semi-
volatile organic compounds across the air-soil interface." The Science of the Total Environment, 228: 5-24.
156 Tuhackova, J. et al. (2001). Hydrocarbon deposition and soil microflora as affected by highway traffic.
Environmental Pollution, 113: 255-262.
157
is:
73 FR 16486 (March 27, 2008).
73FR16491 (March 27, 2008). Only a small percentage of all the plant species growing within the U.S. (over
43,000 species have been catalogued in the USD A PLANTS database) have been studied with respect to ozone
sensitivity.
159 The concentration at which ozone levels overwhelm a plant's ability to detoxify or compensate for oxidant
exposure varies. Thus, whether a plant is classified as sensitive or tolerant depends in part on the exposure levels
being considered. Chapter 9, section 9.3.4 of U.S. EPA, 2013 Integrated Science Assessment for Ozone and
Related Photochemical Oxidants. Office of Research and Development/National Center for Environmental
Assessment. U.S. Environmental Protection Agency. EPA600/R-10/076F.
160 73 FR 16492 (March 27, 2008).
161 73 FR 16493-16494 (March 27, 2008). Per footnote 2 above, ozone impacts could be occurring in areas where
plant species sensitive to ozone have not yet been studied or identified..
162 73 FR 16490/ 16497 (March 27, 2008).
163 U.S. EPA. Integrated Science Assessment of Ozone and Related Photochemical Oxidants (Final Report). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-10/076F, 2013. The ISA is available at
http://cfpub.epa. gov/ncea/isa/recordisplay.cfm?deid=247492#Download.
164 The Ozone ISA evaluates the evidence associated with different ozone related health and welfare effects,
assigning one of five "weight of evidence" determinations: causal relationship, likely to be a causal relationship,
suggestive of a causal relationship, inadequate to infer a causal relationship, and not likely to be a causal
relationship. For more information on these levels of evidence, please refer to Table II of the ISA.
165U.S. EPA. (1991). Effects of organic chemicals in the atmosphere on terrestrial plants. EPA/600/3-91/001.
166 Cape JN, ID Leith, J Binnie, J Content, M Donkin, M Skewes, DN Price AR Brown, AD Sharpe. (2003). Effects
of VOCs on herbaceous plants in an open-top chamber experiment. Environ. Pollut. 124:341-343.
167 Cape JN, ID Leith, J Binnie, J Content, M Donkin, M Skewes, DN Price AR Brown, AD Sharpe. (2003). Effects
of VOCs on herbaceous plants in an open-top chamber experiment. Environ. Pollut. 124:341-343.
6-36
-------
168 Viskari E-L. (2000). Epicuticular wax of Norway spruce needles as indicator of traffic pollutant deposition.
Water, Air, and Soil Pollut. 121:327-337.
169 Ugrekhelidze D, F Korte, G Kvesitadze. (1997). Uptake and transformation of benzene and toluene by plant
leaves. Ecotox. Environ. Safety 37:24-29.
170KammerbauerH, H Selinger, RRommelt, A Ziegler-Jons, D Knoppik, B Hock. (1987). Toxic components of
motorvehicle emissions forthe spruce Picea abies. Environ. Pollut. 48:235-243.
6-37
-------
Chapter 7 Impacts of the Rule on Emissions and Air Quality
This chapter presents the overall emissions and air quality impacts of the Tier 3
standards. Section 7.1 describes the national impacts on criteria and toxic emissions resulting
from the Tier 3 program. Section 7.2 describes the air quality effects of the 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 program on greenhouse gas emissions.
7.1 Criteria and Toxic Pollutant Emission Impacts
7.1.1 Overview
This section presents the projected national emission impacts of Tier 3 standards on
criteria and toxic air pollutants for selected calendar years, and the methodology used to estimate
these reductions. The Tier 3 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 (802). The
implementation of lower sulfur gasoline will reduce criteria and air toxic emissions from the
existing gasoline-powered vehicle fleet, and cause some reductions in SO2 emissions 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 present the emission reductions in calendar year
2018, near the beginning of the fuel program.
The vehicle standards will reduce emissions as the cleaner cars and trucks begin to enter
the fleet in model year 2017 and model year 2018, respectively. 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 in calendar year 2030, when 70 percent of the
miles travelled are from vehicles that meet the fully phased-in Tier 3 standards. Furthermore,
2030 is a standard out-year for evaluation; it is used for air quality modeling in this rule as well
as recent EPA rules. However, the full impact of the vehicle program will 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 vehicles meeting the fully phased-in Tier 3 standards. As
explained in Section 7.2, air quality modeling was done only for 2018 and 2030.
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 Motor
Vehicle Emission Simulator (MOVES) model, as described in detail in Section 7.1.3; and the
NONROAD model for offroad sources. Reductions were estimated compared to a reference case
that assumed an average gasoline sulfur level of 30 ppm (10 ppm in California) and continuation
7-1
-------
of the Tier 2 vehicle program indefinitely, with the exception of California and Section 177
states that have adopted the LEV III program. For those states, adoption of LEV III was
included in the reference case.A
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 7.2.1.1.
7.1.2 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). Table
7-1 below presents an overview of the reference and control scenarios for calendar years 2018
and 2030. Both scenarios reflect the renewable fuel volumes and market fractions projected by
the Annual Energy Outlook 2013 Report.1 We thus refer to this renewable fuel level as "AEO
2013".
A These states include Connecticut, Delaware, Maryland, Maine, Massachusetts, New Jersey, New York, Oregon,
Pennsylvania, Rhode Island, Washington, and Vermont.
7-2
-------
Table 7-1 Overview of Reference and Control Scenarios
Reference Scenario
Control Scenario
2018
Renewable Fuels: AEO 2013a
17.5 B gallons renewable fuels
(18.3 B ethanol-equivalent gallons):
16.0 B gallons ethanol: E10b, E15C
E85d
Fuel Sulfur Level:
30 ppm (10 ppm California)
Fleet:e
96 percent Tier 2 and older vehicles
4 percent LEV III vehicles
Renewable Fuels: AEO 2013a
17.5 B gallons renewable fuels
(18.3 B ethanol-equivalent gallons):
16.0 B gallons ethanol: E10b, E15C
E85d
Fuel Sulfur Level:
10 ppm
Fleet:e
86 percent Tier 2 and older vehicles
14 percent Tier 3/LEV III vehicles
2030
Renewable Fuels: AEO 2013a
17.6 B gallons renewable fuels
(18.6 B ethanol-equivalent gallons):
15.3 B gallons ethanol: E10b, E15C
E85d
Fuel Sulfur Level:
30 ppm (10 ppm California)
Fleet:e
76 percent Tier 2 and older vehicles
24 percent LEV III vehicles
Renewable Fuels: AEO 2013a
17.6 B gallons renewable fuels
(18.6 B ethanol-equivalent gallons):
15.3 B gallons ethanol: E10b, E15C
E85d
Fuel Sulfur Level:
10 ppm
Fleet:e
21 percent Tier 2 and older vehicles
79 percent Tier 3/LEV III vehicles
aU.S. Energy Information Administration, Annual Energy Outlook 2013 (April 15, 2013)
b Gasoline containing 10 percent ethanol by volume
0 Gasoline containing 15 percent ethanol by volume
d Gasoline containing up to 85 percent ethanol by volume (74 percent nominal used in this analysis)
e 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 Tier 3 program, federal gasoline will 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 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, with the exception of California and
Section 177 states that have adopted the LEV III program. Our Tier 3 control scenario modeled
the suite of exhaust and evaporative emission standards 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:
7-3
-------
• 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 for
the light-duty fleet under 6,000 Ibs. GVWR and phasing in from MYs 2018 to
2025 for the light-duty fleet over 6,000 Ibs. GVWR, and MDPVs
• 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 for the light-duty fleet under 6,000 Ibs. GVWR and phasing in
from MYs 2018 to 2025 for the light-duty fleet over 6,000 Ibs. GVWR, and
MDPVs
• Per-vehicle FTP PM standard of 3 mg/mi for LDVs, LDTs and MDPVs,
phasing in from MYs 2017 to 2022 for the light-duty fleet under 6,000 Ibs.
GVWR and phasing in from MYs 2018 to 2022 for the light-duty fleet over
6,000 Ibs. GVWR, and MDPVs
• Per-vehicle US06-only PM standard of 10 mg/mi for LDVs, LDTs and
MDPVs through MY2021 and of 6 mg/mi for MY2022 and later model years
• New standards for Class 2b and 3 trucks phasing in by MY 2022 including
NMOG+NOx declining fleet average, and more stringent PM standards
• More stringent evaporative emission standards for diurnal plus hot soak
emissions, a new canister bleed test and emission standard, and new
requirements addressing evaporative leaks on in-use vehicles.
• New refueling emission control requirements for all complete HDGVs equal
to or less than 14,000 Ibs GVWR (i.e., Class 2b/3 HDGVs), starting in the
2018 model year, and for all larger HDGVs by the 2022 model year
The Tier 3 standards will 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 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.
Accordingly, the analyses for the Tier 3 rule include the final LD GHG standards in both the
reference and control scenarios, and thus account for their impacts on the future vehicle fleet and
future fuel consumption.
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)
7-4
-------
• 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)
• Greenhouse Gas Emissions Standards and Fuel Efficiency Standards for
Medium- and Heavy-Duty Engines and Vehicles (76 FR 57106, September 15,
2011)
• 2017 and Later Model Year Light-Duty Vehicle Greenhouse Gas Emissions and
Corporate Average Fuel Economy Standards (77 FR 62623, October 15, 2012)
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.
Furthermore, the Tier 3 emissions modeling for both the national inventory and air quality
analysis includes California's LEV III program and its associated emission reductions from both
California and the states that adopted the LEV III program, in the baseline scenario. See the Tier
3 emissions modeling TSD for more detail.
7.1.3 Onroad Emissions
7.1.3.1 Methodol ogy Overvi ew
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.3
A version of MOVES2010b, 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 7.1.2 above, for calendar
years 2018, 2030, and 2050B
T3
The MOVES updates are reflected in a version of the MOVES model code (October 2, 2012 Version F) and
concurrently updated versions of the MOVES default database; October 2, 2012 Version K_truncatedGFRE for the
inventory runs for the air quality modeling and October 2, 2012 Version L_truncatedGFREIM for the national
inventory runs (see Section 7.2.1.1 for details). Both the code and the databases are available in the Tier 3 docket.
As these updates are still draft, these code and/or databases are not approved for official use in SIP and conformity
analyses.
7-5
-------
The MOVES model updates affecting reference and control case emissions were
extensive, and are documented in a separate memorandum to the Tier 3 docket.4 Updates made
to MOVES2010b for this analysis primarily incorporated major new research in four 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 fourth involves estimating the exhaust emissions from vehicles using E85 fuel (gasoline
containing up to 85 percent ethanol by volume). Other than sulfur, these changes had more
bearing on updating the reference case emissions than on the projected reductions from the Tier
3 standards.
The effects of changes in fuel properties on exhaust emissions of Tier 2 vehicles, which
comprise the majority of the fleet by 2018, were assessed through the results of the EPAct Phase-
3 Program.5 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.6 Improvements in estimating and projecting evaporative emissions are described in this
document (see Section 7.1.3.3.7). Finally, because the updates to fuel sulfur effects are critical
for estimating the reductions from the Tier 3 program, they are also presented in detail in Section
7.1.3.4.
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 MOVES2010b
*~1 Q
emission rates for NHs, NO and NO2. '
The MOVES version used for this analysis also includes an added capability to model
many hazardous air pollutants. And, additional changes were made to the MOVES2010b 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 for the
proposed rule addressing MOVES updates9.
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 the air quality modeling. The abbreviated approach makes
c 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.
7-6
-------
the analysis easier for stakeholders and other commenters to replicate. National emission
inventories were developed using the pre-aggregation feature of MOVES at the state level. For
all pollutants, the default pre-aggregation level of 'hour' was selected, which condenses the
county-level temperatures into a single state 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 the difference in temperature resolution (state average vs. gridded
hourly temperature), the national emission inventory analysis also used information contained in
the MOVES2010b default database for fleet age distributions, activity inputs (e.g., speeds),
temperatures, emission standards, and inspection/maintenance programs. In contrast, as
discussed in Section 7.2, the air quality modeling inventory methodology used the data supplied
by state and regional modelers for many of these inputs, and employed hourly meteorological
data. Both the national emission inventories and the air quality modeling inventory accounted
for the state adoption of California LEV III standards, as well as previous LEV programs, in the
modeling baseline. The future year projections of vehicle population and vehicle miles travelled
were updated to reflect the latest estimates from the Department of Energy's Annual Energy
Outlook (AEO2013).
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 4.3 was applied to January and 7.5 to July; these factors were determined based
on analysis of annual PM emissions during modeling for the RFS2 rule.10 The updated MOVES
version, and all inputs and outputs that produced the results presented in Section 7.1.5 of this
Chapter, are contained in the Tier 3 rulemaking docket.
7.1.3.2 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 state-level 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.
In order to further simplify the final analysis, provide more consistent results for future
efforts, and create additional data for renewable volume sensitivity, we elected to use projections
found in the Annual Energy Outlook 2013 repor^for our primary case. This report provided the
basis for our analysis with conventional fuel volumes, as well as renewable fuel volumes and
flex fuel vehicle usage year by year. In some years, the volume and usage differ significantly
from the analysis found in the proposal (based on the RFS2 "mid-ethanol" full compliance case).
7-7
-------
Comparisons between our final analysis and the analysis completed in the proposal can be found
in the following subsection below.
7.1.3.2.1AEO Fuel Volume Analysis
The Annual Energy Outlook 2013 report provided the basis for the fuel volumes used in
this analysis. The AEO2013 report contains updated effects on fuel volumes from programs not
included in the analysis for the proposal, most notably the latest light duty greenhouse gas
standards2. The inclusion of that rulemaking significantly reduces the overall volume of fuel used
as compared to the volume found in the proposal. The AEO2013 projections also show a lower
volume of renewable fuel used, and coupled with the reduced fuel volumes, show an increasing
fraction at which renewable fuel would have to be blended into the supply as compared to
volumes found in the RFS2 rulemaking as well as the NPRM modeling. A comparison between
the volumes used for the final analysis and the proposal can be found in Table 7-2, below:
Table 7-2 Comparison of NPRM Fuel Volume with FRM Fuel Volume
(Million gallons)
Total Fuel Volume
Ethanol Volume
E10 Volume
E15 Volume
E85 (as E74) Volume
NPRM 2017
144.9
21.6
84.5
60.0
0.4a
FRM 2018
129.6
15.2
100.7
27.5
1.4
NPRM 2030
148.3
30.5
1.7
146.6
0.0
FRM 2030
113.5
14.6
68.4
43.4
1.7
a This small volume of E85 fuel was not modeled in the NPRM 2017 version
The AEO2013 report shows that due to the increase in renewable fuel blending levels,
E85 must be included in the fuel supply for the final analysis as compared to the NPRM. E85
fueling rates were derived from the AEO2013 report on flex fuel vehicle sales and E85 fuel
volume, and can be found in Table 7-3, below. This table describes the fraction of the gasoline
vehicle population that is flex-fuel capable, the fraction of E85 in the overall gasoline fuel pool,
and finally the fraction of refueling events where flex-fuel capable vehicles fill with E85 fuel.
Table 7-3 FFV Population, E85 Fuel Volume and Overall FFV E85 Usage Rates
FFV Pop. % (of gasoline vehicles)
E85 Vol % (of gasoline fuel volume)
Overall E85 Usage by FFVs %
2018
6.9%
1.0%
15.3%
2030
8.1%
1.5%
18.9%
Updates in modeling technique have allowed us to vary ethanol blending penetration by
region, with data provided by the AEO2013 report. These regional rates vary between 63% -
83% for E10 penetration and 17% - 37% for E15 penetration in 2018 and between 45% - 78%
for E10 penetration and 22% - 55% for E15 penetration in 2030. The remainder of fuel was E85,
with no EO included in the on-road fuel supply. The non-road/off-road fuel supply was assumed
to be 100 percent E10. For more information regarding non-road modeling, please see Section
7-8
-------
7.1.4. A breakdown of ethanol penetration by region can be found in subsection 7.1.3.2.7, Table
7-7.
7.1.3.2.2 Baseline and Reference Case
Since the platform for air quality modeling is based on 2007, a baseline fuel supply was
needed for that year. The regional fuel supply approach was applied using Alliance fuel survey
data as well as the EPA refinery compliance data for the year 2007, as described in the
subsection 7.1.3.2.4.
Reference cases in the years 2018 and 2030 were required for the final analysis. These
reference cases were also created following the regional fuel supply approach using Alliance fuel
survey data as well as the EPA refinery compliance data from the years 2011 and 2012 as the
most up-to-date surrogate for fuel properties in those years. Additionally, these fuel supplies
were further corrected to provide consistent results region to region when compared to the
control cases. These corrections are as follows below:
Benzene and sulfur levels for all counties were 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 Petroleum Administration Defense
Districts (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 Table 7-4 below:
Table 7-4 MSAT2 Downstream Fuel Benzene Levels
PADD
1
2
3
4
5
CA
CG
0.61
0.63
0.63
0.86
0.65
N/A
RFC
0.54
0.60
0.54
N/A
0.61
0.62
Sulfur corrections were made to all counties based on the default sulfur level found in the
2007 baseline 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 an 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:
7-9
-------
1) A "high sulfur" aromatics level was determined using the following equation:
high sulfur aromatics level — - * 0. 479 + 24. 60
\sulfur level
Equation 7-1 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 2007 fuel supply
Diesel fuel sulfur levels were also adjusted to 15ppm to reflect low sulfur diesel levels.
There were no other changes to the 2018 and 2030 reference cases. For detailed fuel property
information by region for the reference case, please refer to the subsection 7.1.3.2.7, Table 7-9.
7.1.3.2.3 Control Case
The Tier 3 control fuel scenarios for the years 2018 and 2030 used the fuel supplies
constructed for the 2018 and 2030 reference cases described in the previous section as a
foundation. To develop the control scenario fuel supplies, we modified the reference case fuel
supplies to reflect the sulfur program in the Tier 3 control case by reducing sulfur from 30 ppm
to 10 ppm for all gasoline. Changes in other fuel properties resulting from sulfur control were
determined by refinery modeling and reflected in the control case, such as an increase in
aromatics and decrease in olefms and distillation properties, as shown in Table 7-13. These
changes were made to every county with fuel exceeding a sulfur level of lOppm. Please note, the
impacts to other fuel properties due to sulfur reduction are significantly lower in the final
analysis than what was used for the proposal. Updated refinery modeling results have reduced
the impact on these other fuel properties, which is properly reflected in the final analysis fuel
supplies. For more information regarding refinery modeling adjustments, please refer to
subsection 7.1.3.2.9.
The result of this effort was two additional alternate fuel supply databases tables for use
in MOVES, reflecting the control case fuel supplies in 2018 and 2030; these tables were used for
the development of the final national emissions inventory as well as the inventory used for air
quality modeling.
7.1.3.2.4 Regional Fuel Supply
In addition to simplifying the fuel volumes, we also made an effort to simplify the
county-level fuel properties within the MOVES fuel supply for the final analysis. Many counties
in the previous analysis contained unique fuel properties not associated with neighboring
7-10
-------
counties or pipeline locations based on fuel survey data performed in those individual counties.
We did not feel that it was appropriate to continue this approach for the final analysis, and
developed new fuel properties based on averages of these survey data as well as data provided to
us at the refinery gate as part of EPA fuel compliance. We believe this average data by region
provides a more consistent and maintainable basis for this final analysis as well as future
analyses. Details regarding this regional fuels approach can be found in the following
subsections, 7.1.3.2.5 - 7.1.3.2.6.
7.1.3.2.5 Rationale for Updating MO VES Fuel Supply County Level Properties
The fuel supplies used in the proposal analysis as well as the default version of MOVES
are based on single-point county data on a small number of cities. Fuel survey data has tended to
include large deviation depending on when fuel was sampled, which batch of fuel the sample
was pulled from, which blend of fuel was being sampled, and/or when the station was refueled.
Depending on which batch of fuel was at a station during sampling and testing, properties could
dramatically impact the result in one city versus another. Updating this data would also result in
potentially wide swings in assumed fuel properties for that area, when in reality the average fuel
for that location had not changed. As a result, the single point county data are not very useful for
creating national fuel supplies. Refer to Figure 7-1 below, for an example of deviation in fuel
batch parameters for E200 through a one year period. Note while there is an overall trend in fuel
property variation throughout the year, which is expected, there is also a deviation of up to 50%
in some cases.
Figure 7-1 Fuel Property Batch Variation Throughout One Year
. -?..». «•
«t» <..••• • '•".•"• •*--'c:h.-»"?v.
SSSfSkVv- ;/••.-• i-'^-^^-
? '• ** ',••••.-"-%"."»v"v^nt»".•;"''.•*->-akp?iŁ1»^}
^^SJ^^Si^^^^P
m*
250 300
Time
400 45O
+7.333e5
As a result, the use of this survey data city-by-city has lead to a fuel supply contained in
the default version of MOVES, as well as the proposal analysis, that includes many variations of
fuel properties in a non-contiguous fashion, often with little application to fuel distribution or
overall properties. Without aggregation of the data, the fuel supply became very large, and less
7-11
-------
representative of our actual knowledge of fuel property variation overall. For example, the fuel
supply used in the proposal analysis contained approximately 425 various fuel formulations,
spread over the county in a way not representative of distribution networks, natural borders, and
state/regional programs. Please see Figure 7-2, below, for a detailed layout of the proposal fuel
supply:
Figure 7-2 National Map of Unique Fuel Properties Used in Proposal
With a new method aggregating fuel survey data, as well as incorporating refinery batch-
by-batch fuel compliance data, a more representative fuel supply was created for the final
analysis. This new version of the fuel supply better accounts for fuel production and distribution
networks, natural borders, and regional/state/local variations in fuel policy. Reducing the number
of fuel formulations in the supply also increased our confidence that the fuels in a particular
region represent fuel being used in that region; rather than being based on samples taken in a
particular city or small set of cities. The new regional fuel supply method created approximately
45 fuel formulations. Please see Figure 7-3, below, for a detailed layout of the regional fuel
supply used in the final analysis:
Figure 7-3 National Map of Unique Fuel Properties Using Regional Fuels Approach
7-12
-------
7.1.3.2.6 Fuel Regions
There are eleven general fuel regions used in the new regional fuel supply approach.
Table 7-5 identifies and briefly describes each region as used in the MOVES fuel supply
database. Please see Figure 7-3, above, for an illustration of these fuel regions on a national map.
Table 7-5 Description of Fuel Regions
Region ID#
1
2
3
4
5
6
11
12
13
14
15
Region Name
East Coast
Midwest
South
North
Rocky Mts
CA/NV/AR
East Coast RFC
MD/VA
Texas RFC
Midwest RFC
California
Description
East coast states up to Appalachians, Florida, and gulf
coast region
Midwest states up to Appalachians (not including
Wisconsin), Tennessee, Kentucky
Southern states not including gulf coast, Nebraska,
Iowa
North and South Dakota, Minnesota, Wisconsin
Pacific northwest, Rocky mountain states, Utah
California, Nevada, Arizona NOT using RFG
East coast states and regions using RFG fuel or under a
controlled fuel program
Maryland and Virginia regions using RFG fuel or
under a controlled fuel program
Texas regions using RFG fuel or under a controlled
fuel program
Midwest regions using RFG fuel or under a controlled
fuel program
California using California fuel, Nevada and Arizona
regions using California Fuel
7.1.3.2.7 Fuel Properties by Region
The following subsection contains tables detailing fuel properties by region for the final
analysis 2007 baseline, 2018 and 2030 reference cases, and the 2018 and 2030 control cases.
7-13
-------
Fuel properties in Table 7-8, Table 7-9 and Table 7-10 below are for conventional gasoline only.
To derive E10 and E15 properties, please refer to Table 7-6, below, for the adjustment factors
used.
Table 7-6 Ethanol Blending Adjustments for Conventional Gasoline Properties Within
Fuel Regions
ETHANOL ADJUSTMENT FACTORS (from EO to EXX)
FUEL
BIOS
E10W
E15S
E15 W
DESCRIPTION
E10 Summer Fuel
E10 Winter Fuel
El 5 Summer Fuel
El 5 Winter Fuel
RVP
1.00
1.00
SULF
AROM
-2.02
-3.65
-3.36
-5.69
OLEF
-0.46
-2.07
-1.64
-3.27
BENZ
E200
3.11
4.88
9.24
11.11
E300
0.39
0.54
0.91
1.01
T50
-6.34
-9.96
-18.86
-22.67
T90
-1.77
-2.45
-4.14
-4.59
Table 7-7 Non - Flex Fuel Ethanol Fuel Blending Levels by Region1
REGION
1
2
3
4
5
6
11
12
13
14
15
East Coast
Midwest
South
North
Rocky Mts
CA/NV/AR
East Coast REG
MD/VARFG
Texas REG
Midwest REG
California
2018
E10
0.834
0.722
0.746
0.632
0.817
0.822
0.834
0.834
0.746
0.722
0.000
E15
0.166
0.278
0.254
0.368
0.183
0.178
0.166
0.166
0.254
0.278
1.000
2030
E10
0.628
0.592
0.447
0.544
0.686
0.780
0.628
0.628
0.447
0.592
0.000
E15
0.372
0.408
0.553
0.456
0.314
0.220
0.372
0.372
0.553
0.408
1.000
1 This table does not include the contribution of fuel blending due to flex fuel (E85) usage. For flex fuel blending
levels, please refer to Table 7-3
7-14
-------
Table 7-8 2007 Baseline Case Fuel Properties by Region
REGION
1
2
3
4
5
6
11
12
13
14
15
East Coast
Midwest
South
North
Rocky Mts
CA/NV/AR
East Coast REG
MD/VARFG
Texas REG
Midwest REG
California
SUMMER
RVP
8.70
8.70
8.70
8.70
8.70
8.70
6.90
6.91
6.92
7.06
7.06
SULF
34.66
42.00
78.88
37.92
65.06
78.88
33.12
35.10
30.12
32.12
9.00
AROM
27.60
28.52
28.55
24.53
28.17
28.55
21.69
20.11
16.65
17.13
21.98
OLEF
12.78
9.95
9.02
8.65
9.67
9.02
11.61
11.76
11.12
7.85
4.44
BENZ
1.004
1.540
1.689
1.211
1.765
1.689
0.641
0.626
0.533
0.774
0.530
E200
44.06
47.87
45.56
49.17
46.03
45.56
50.26
50.35
50.58
50.98
44.52
E300
80.22
81.24
84.71
82.23
85.26
84.71
84.97
83.84
85.05
85.24
88.81
T50
214.9
211.5
211.3
201.2
207.9
211.3
196.2
195.4
196.1
193.2
211.0
T90
337.6
339.6
318.2
333.7
318.3
318.2
322.7
331.1
329.0
326.7
303.0
WINTER
RVP
11.31
11.84
10.85
12.55
11.75
10.85
11.31
11.31
10.85
11.84
11.84
SULF
34.27
35.77
80.44
32.90
57.50
80.44
33.12
35.10
30.12
32.12
9.00
AROM
24.68
25.51
26.47
21.67
25.76
26.47
21.69
20.11
16.65
17.13
21.98
OLEF
12.71
9.36
8.55
8.67
9.50
8.55
11.61
11.76
11.12
7.85
4.44
BENZ
0.915
1.495
1.661
1.081
1.737
1.661
0.641
0.626
0.533
0.774
0.530
E200
48.52
52.03
48.78
52.49
50.41
48.78
55.82
55.89
55.53
57.00
50.54
E300
82.52
84.02
86.80
85.06
86.84
86.80
87.18
86.04
87.02
87.63
91.20
T50
202.4
198.7
205.3
195.2
199.5
205.3
184.9
184.1
186.0
180.9
198.5
T90
330.8
239.2
312.5
322.3
311.4
312.5
312.7
321.1
320.1
315.8
292.0
Table 7-9 Reference Case Fuel Properties by Region
REGION
1
2
3
4
5
6
11
12
13
14
15
East Coast
Midwest
South
North
Rocky Mts
CA/NV/AR
East Coast REG
MD/VARFG
Texas REG
Midwest REG
California
SUMMER
RVP
8.70
8.70
8.70
8.70
8.70
8.70
6.90
6.90
6.90
6.90
6.90
SULF
30.00
30.00
30.00
30.00
30.00
30.00
30.00
30.00
30.00
30.00
9.00
AROM
22.89
25.40
26.82
24.28
27.41
26.82
21.69
20.11
16.65
17.13
21.98
OLEF
13.19
7.13
10.75
8.18
8.13
10.75
11.61
11.76
11.12
7.85
4.44
BENZ
0.610
0.630
0.630
0.860
0.650
0.650
0.540
0.540
0.540
0.600
0.530
E200
45.78
47.65
45.34
47.20
45.24
45.34
50.26
50.35
50.58
50.98
44.52
E300
80.81
82.32
84.65
81.13
84.64
84.65
84.97
83.84
85.05
85.24
88.81
T50
208.4
204.6
209.3
205.5
209.5
209.3
199.3
199.1
198.6
197.8
211.0
T90
339.4
332.5
321.9
337.9
322.0
321.9
320.5
325.6
320.1
319.2
303.0
WINTER
RVP
10.82
11.96
10.36
12.92
11.61
10.36
11.31
11.31
10.85
11.84
11.84
SULF
30.00
30.00
30.00
30.00
30.00
30.00
30.00
30.00
30.00
30.00
9.00
AROM
21.19
22.40
24.92
20.79
25.12
24.92
21.69
20.11
16.65
17.13
21.98
OLEF
12.36
7.12
9.69
8.42
8.00
9.69
11.61
11.76
11.12
7.85
4.44
BENZ
0.610
0.630
0.630
0.860
0.650
0.650
0.540
0.540
0.540
0.600
0.530
E200
50.32
52.93
48.93
51.33
49.73
48.93
55.82
55.89
55.53
57.00
50.54
E300
83.58
85.12
86.92
83.74
86.46
86.92
87.18
86.04
87.02
87.63
91.20
T50
199.2
193.8
202.0
197.1
200.4
202.0
187.9
187.8
188.5
185.5
198.5
T90
326.8
319.8
311.6
326.0
313.7
311.6
310.4
315.6
311.2
308.4
292.0
7-15
-------
Table 7-10 Control Case Fuel Properties by Region
REGION
1
2
3
4
5
6
11
12
13
14
15
East Coast
Midwest
South
North
Rocky Mts
CA/NV/AR
East Coast REG
MD / VA REG
Texas REG
Midwest REG
California
SUMMER
RVP
8.70
8.70
8.70
8.70
8.70
8.70
6.90
6.90
6.90
6.90
6.90
SULF
10.00
10.00
10.00
10.00
10.00
10.00
10.00
10.00
10.00
10.00
9.00
AROM
22.91
25.42
26.84
24.30
27.43
26.84
21.71
20.13
16.67
17.15
21.98
OLEF
12.39
7.13
10.75
8.18
8.13
10.75
11.61
11.76
11.12
7.85
4.44
BENZ
0.610
0.630
0.630
0.860
0.650
0.650
0.540
0.540
0.540
0.600
0.530
E200
45.78
47.65
45.34
47.20
45.24
45.34
50.26
50.35
50.58
50.98
44.52
E300
80.78
82.29
84.62
81.10
84.61
84.62
84.94
83.81
85.02
85.21
88.81
T50
208.4
204.6
209.3
205.5
209.5
209.3
199.3
199.1
198.6
197.8
211.0
T90
339.5
332.6
322.0
338.0
322.1
322.0
320.6
325.7
320.2
319.4
303.0
WINTER
RVP
10.82
11.96
10.36
12.92
11.61
10.36
11.31
11.31
10.85
11.84
11.84
SULF
10.00
10.00
10.00
10.00
10.00
10.00
10.00
10.00
10.00
10.00
9.00
AROM
21.28
22.49
25.01
20.88
25.21
25.01
21.78
20.20
16.74
17.22
21.98
OLEF
11.40
6.16
8.73
7.46
7.04
8.73
10.65
10.80
10.16
6.89
4.44
BENZ
0.610
0.630
0.630
0.860
0.650
0.650
0.540
0.540
0.540
0.600
0.530
E200
50.23
52.84
48.84
51.24
49.64
48.84
55.73
55.80
55.44
56.91
50.54
E300
83.50
85.04
86.84
83.66
86.38
86.84
87.10
85.96
86.94
87.55
91.20
T50
199.3
194.0
202.2
197.3
200.6
202.2
188.1
188.0
188.7
185.7
198.5
T90
327.1
320.1
312.0
326.4
314.0
312.0
310.8
316.0
311.5
308.7
292.0
7.1.3.2.8 Boutique Fuels
The new regional fuel supply used for the final analysis also contains fuel formulations
for counties or regions using local fuel controls, also referred to as "boutique fuels". These fuel
controls are usually expressed in additional control on RVP and/or ethanol blending levels. Table
7-11 presents the fuel property adjustments due to RVP control used in the final analysis.
Table 7-11 RVP Adjustment Factors for Boutique Fuel Areas Within Fuel Regions
RVP ADJUSTMENT FACTORS (per PSI)
FUEL
per PSI
DESCRIPTION
boutique fuel adj
RVP
-1.00
SULF
AROM
OLEF
BENZ
E200
-1.26
E300
-0.50
T50
2.57
T90
2.27
In addition to the properties listed above, a small change in aromatics and olefins was
found to be associated with changes to RVP in boutique fuel areas. After additional refinery
modeling, effects to these properties were not found to be significant and were excluded from the
analysis.
7.1.3.2.9 Application of Refinery Modeling
It was necessary to estimate the gasoline qualities and changes in gasoline quality to
estimate the emissions impact of the Tier 3 program. This was conducted in two separate steps -
the first step estimated the impact of blending in more ethanol, the second step for estimating
desulfurization on gasoline qualities.
For estimating ethanol's impact on gasoline qualities, it was necessary to establish a base
case for refinery modeling, for which we chose the year 2005. However, we revised the base
case from what actually occurred in 2005, by replacing the content of MTBE blended into
7-16
-------
gasoline in 2005 with 5.55 billion gallons of ethanol. Next, we established a reference case and
the sole differences that we modeled between the 2005 revised base case and the 2030 reference
case were the gasoline and other refinery product volumes (which change with changing
demand), and product pricing (which changes based on forecasted changes in crude oil prices).
Finally, we then modeled two control cases which reflect different ethanol volumes, one with
100 percent E10 and the other with 100 percent E15.D This allowed us to estimate the impacts of
the two different amounts of ethanol on gasoline qualities. These two ethanol control cases were
also modeled in 2030. The changes in gasoline quality are summarized in Table 7-12. For
estimating how ethanol affects gasoline quality, we solely used the national average change in
gasoline qualities and applied those changes for all El0 or El5 gasoline.
D Because E85 is blended up with the same gasoline blendstock which comprises E10 or E15, there was no need to
estimate separate gasoline qualities for the gasoline blendstock which is used for E85.
7-17
-------
Table 7-12 Difference in Gasoline Qualities between the 2030 E10 and E15 Control Cases
with the 2030 Reference Case
2030 (E10)
PADD1
PADD2
PADD3
PADD
4/5
USavg
minus CA
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
Refinery
Modeling
Summer
6.57
1.54
0.67
-3.81
-3.55
-0.50
-0.49
0.01
0.47
0.01
0.06
0.37
-0.03
-1.03
5.23
5.40
-2.10
0.89
-3.48
-1.54
3.11
0.39
0.35
-2.02
-0.46
Winter
4.21
0.03
0.70
-2.70
-3.24
5.41
1.22
-0.24
-4.56
-0.46
4.48
0.56
-0.23
-3.92
-2.09
7.39
0.78
-0.30
-4.71
-1.17
4.88
0.54
0.15
-3.65
-2.07
2030 (E15)
PADD1
PADD 2
PADD 3
PADD
4/5
USavg
minus CA
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
Refinery
Modeling
Summer
14.07
3.06
0.06
-6.06
-5.09
4.65
-0.43
-0.72
-0.45
-0.43
6.13
0.86
-0.72
-1.13
3.47
7.55
-5.18
0.00
-4.62
-1.49
9.24
0.91
-0.33
-3.36
-1.64
Winter
10.94
1.38
0.37
-6.55
-5.49
11.65
0.91
-0.65
-4.54
-0.35
9.89
0.41
-0.57
-4.73
-3.20
13.38
1.07
-0.58
-7.58
-1.62
11.11
1.01
-0.20
-5.69
-3.27
The second step for estimating gasoline qualities was to model the impact of
desulfurization on gasoline qualities. The total impact of desulfurization on gasoline qualities is
comprised of the reduction in gasoline sulfur, the associated reduction in olefins and the impacts
of recovering the lost octane. The sulfur reduction is fixed by the standard and the olefins
reduction is a function of the selectivity of the desulfurization technologies. We reviewed the
information that we had obtained for the gasoline desulfurization technologies and estimated that
desulfurizing gasoline from 30 ppm to 10 ppm would result in a 1 percent reduction in olefm
level. Since we estimated the cost of making up lost octane using the LP refinery model, we
used that case for estimating the impact of octane recovery on gasoline qualities. The gasoline
7-18
-------
qualities for the reference case and the control case that we developed reflect a 1 octane number
loss in the FCC naphtha pool. However, we estimate that the gasoline desulfurization equipment
that refiners will use to comply with Tier 3 will only cause about a one half octane number
reduction in the gasoline pool, so we reduce the gasoline quality changes by half. The difference
in gasoline qualities between the reference and control cases reflecting a one half reduction in
octane number is summarized in Table 7-13. Because of the tendency for the LP refinery model
to shift gasoline blendstocks around resulting in odd gasoline quality changes in individual
PADDs, we solely used the national average change in gasoline qualities and applied those
changes for all gasoline for the emissions analysis.
Table 7-13 Differences in Gasoline Qualities between the Control and Reference Cases
Change in Gasoline Quality forl/2 Octane
Number Decrease
PADD1
PADD2
PADD3
PADD 4/5
USavg
minus CA
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
E200
E300
RVP
Aromatics
Olefins
Summer
-0.11
-0.12
0.00
0.02
0.20
-0.10
-0.10
0.00
0.01
0.17
0.38
0.29
0.00
-0.02
0.36
0.00
-0.02
0.00
0.11
0.01
0.00
-0.03
0.00
0.02
0.20
Winter
-0.04
-0.09
0.00
0.18
0.11
-0.10
-0.03
0.01
-0.03
-0.01
-0.22
-0.16
0.01
0.11
-0.03
-0.02
-0.06
0.00
0.10
0.00
-0.09
-0.08
0.01
0.09
0.04
7-19
-------
7.1.3.3 Vehi cl e Program Inputs
Modeling the controls introduced by the Tier 3 vehicle program required the development
of a set of alternate MOVES database tables to reflect each aspect of the Tier 3 program. These
database tables included:
• Gaseous exhaust emission rates (HC/CO/NOx) for light-duty cars, trucks, and
light-heavy trucks (gas and diesel) to reflect the 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-duty trucks (gas and diesel) to reflect the Tier 3
FTP and US06 PM standards and phase-in.
• Evaporative hydrocarbon emission rates for the permeation process to reflect the
diurnal test standard, certification fuel, and phase-in.
• Leak prevalence rates for tank vapor and liquid leaks to reflect requirements for
evaporative leak detection.
The development of these alternative inputs is discussed below by pollutant, fuel and
vehicle regulatory class.
7.1.3.3.1 Gasoline Light-Duty HC/CO/NOX Exhaust
Emission rates for gaseous pollutants (HC/CO/NOx) in MOVES are contained in a
database table (EmissionRateByAge). Rates are expressed in terms of mass per time (g/hr),
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, engine soak time for start). Developing these rates for Tier 3 vehicles required
accounting for expected changes in each of these dimensions.
The development of emission rates representing implementation of the Tier 3 standards
followed the same procedures used to develop rates for the National LEV (NLEV, covering
model years 2001-2003) and Tier 2 standards (covering model years 2004 and later) in the
default MOVES database, as described in the documentation for light-duty exhaust emission
rates for MOVES2010 (known as the "MOVES Light-Duty report").11 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
7-20
-------
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 standards. For Tier 3, we
developed estimates of FTP results for Tier 3 vehicles based on IUVP data from vehicles
certified to Tier 2 Bin 2 and 3 standards, including cycle composites, "cold-start" emissions"
(Bagl minus Bag3) and "hot-running" emissions (FTP Bag 2 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. Note that for purposes of inventory modeling for the FRM, the onset of the
Tier 3 phase-in was delayed from 2017 to 2018 for truck classes with gross vehicle weight
ratings > 6,000 Ib (LDT3 and LDT4). For LDV, LDT1 and LDT2, the phase-in begins in 2017,
as in the inventory modeling for the NPRM.
3. Merge FTP results and Phase-in assumptions. For running emissions, calculate
weighted ratios of FTP and US06 emissions in each model year relative to those for cars (LDV)
in MY2000, which represent Tier 1 LDV.
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 120,000 miles, which was assumed for Tier 2 and NLEV
vehicles. This is the outcome of applying In-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/Mreference 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 Irates.
Each of these six steps is described in more detail below. Addition information is
17
available in a separate memo available in the docket.
7.1.3.3.1.1 Average FTP Results (Step 1) (Standard)
Our projected emissions for Tier 3 vehicles are driven by the NMOG+NOx standard, set
at 30 mg/mi. However, because MOVES models 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
7-21
-------
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 "Bagl 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 Table 7-14, 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 Equation 7-2.
T3 = 0.775 -B2 + 0.225 -B3
Equation 7-2
Table 7-14 Hydrocarbons (HC): Useful-Life FTP Standards and Associated Cold-Start
and Hot-Running Results on the FTP and US06 Cycles.
Bin
8
5
4
3
2
Tier 3C
Useful-life Standard
(mg/mi)
125
90
70
55
10
20
FTP Composite51
(mg/mi)
41.3
35.5
24.8
21.5
5.6
9.2
FTP Cold Start3
(mg)
591
534
383
329
87
142
FTP hot Running3
(Bag 2)
(mg/mi)
3.56
2.63
2.28
1.74
0.42
0.7
US06b
(mg/mi)
35.8
35.8
35.8
35.8
2.6
10.0
a Values represent "non-methane organic gases" (NMOG).
b Values represent "non-methane hydrocarbons" (NMHC).
0 Values for Tier 3 calculated using 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 Table 7-15.
7-22
-------
Table 7-15 CO: Useful-Life FTP Standards and Associated Cold-Start and Hot-Running
Results on the FTP and US06 Cycles.
Bin
8
5
4
3
2
Tier 3a
Useful-life Standard
(mg/mi)
4,200
4,200
4,200
2,100
2,100
2,100
FTP Composite
(mg/mi)
861
606
537
463
235
286
Cold Start
(mg)
6,680
5,510
5,500
3,470
1,620
2,040
FTP hot Running
(Bag 2)
(mg/mi)
451
238
201
119
70
81
US06
(mg/mi)
2,895
2,895
2,895
2,895
948
1,390
a Values for Tier 3 calculated using Equation 7-2.
Corresponding results for NOx are presented in Table 7-16. 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 7-16 NOX: Useful-Life FTP Standards and Associated Cold-Start and Hot-Running
Results on the FTP and US06 Cycles.
Bin
8
5
4
3
2
Tier3
Useful-life Standard
(mg/mi)
200
70
40
30
20
10
FTP Composite
(mg/mi)
64.2
21.2
8.7
5.7
5.5
5.5
Cold Start
(mg)
418
165
90
71
67
67
FTP hot Running
(Bag 2)
(mg/mi)
35.1
8.2
4.7
3.8
0.4
0.4
US06
(mg/mi)
61.3
45.9
30.6
30.6
18.4
18.4
7.1.3.3.1.2 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 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.
7-23
-------
LDV/LDT1
LDT234
Model Year
Figure 7-4 NMOG+NOX FTP Fleet Average Requirements during Phase-In of the Tier 3
Exhaust Emissions Standards for Light-Duty Vehicles.
7.1.3.3.1.3 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.
7.1.3.3.1.4 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 cycle (< -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 Figure 7-5).
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 2017 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 Table 7-14 through Table 7-16. For HC and
CO, we used Equation 7-2, as before. For NOx, we applied the Bin-2 values. Figure 7-6 and
7-24
-------
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 7-6 displays the
information on linear scale, highlighting the differences at the higher operating modes, while
Figure 7-7 shows the same information on a logarithmic scale, illustrating the patterns for the
lower operating modes.
30 +
IF 27-30
c
Ł 24-27
1 27-24
| 78-27
Ł 15-18
u
t 72-75
u
$9-72
73 6'9
-S 3-6
^ 0-3
<0
Speed Class
7-25 25-50
16 30
29
28
^^^^^^^^^^M
1 27
15 25
14 24
13 23
12 22
11 21
(mph)
50 +
40
39.x
^x*^
38
=^Ł^^=— — —
^*^^^^~^^^^^S
37
•
35
«
33
^.
The "US06 Region"
High Power
The "hot-running FTP Region"
Low to Moderate
Power
(also includes braking (0)
and idle (1)
Figure 7-5 Operating modes for running Exhaust Emissions, divided broadly into "hot-
running FTP" and "US06" regions.
7-25
-------
s
I
g
'
(A
E
LU
-5 0 5 10 15 20 25 30 35 40
Vehicle Specific Power (kW/Mg)
5 10 15 20 25
Vehicle Specific Power (kW/Mg)
30
35
40
ra
Oi
E
LU
-5
5 10 15 20 25 30
Vehicle Specific Power (kW/Mg)
Figure 7-6 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).
7-26
-------
10,000
Ł 1,000
I
(A
(A
E
LU
100 !-
0 10 20 30
Vehicle Specific Power (kW/Mg)
40
O)
I
oi
c
,0
'35
M
I
LU
100.00
10.00
1.00
0.10
0.01
0 10 20
Vehicle Specific Power (kW/Mg)
30
40
0 5 10 15 20 25
Vehicle Specific Power (kW/Mg)
30
35
40
Figure 7-7 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).
7-27
-------
7.1.3.3.1.5 Apply Deterioration (Step 5)
Based on our extensive emissions analysis during MOVES2010 development, we
assumed that deterioration for different technologies was best represented by a multiplicative
model, in which different technologies, represented by successive model-year groups, showed
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 vehicles manufactured after the onset of the Tier 3 phase-in, the deterioration
assumptions were modified to represent an extension of the full useful life (FUL) from 120,000
mi to 150,000 mi. Thus, the inventory modeling assumes a standard of 30 mg/mi NMOG+NOx
and a 150,000 mi useful life. However, under the final rule, manufacturers will retain the option
of certifying some engine test groups to a somewhat lower standard with a 120,000 mi useful
life. Over the useful life of vehicles, we assume that the two options (higher standard with longer
useful life, lower standard with shorter useful life) will yield approximately equivalent
deterioration trends. Therefore, no attempt was made to represent both options in the inventory
modeling.
Note that we did not extrapolate the logarithmic deterioration trend beyond the 8-9 year
age group, as we know that emissions tend to stabilize beyond this age, while the In-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).
7.1.3.3.1.6 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).
7.1.3.3.1.7 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.
7-28
-------
7.1.3.3.1.8 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 Tier 3 standards, emissions can
be constructed for FTP composite and US06 from MOVES emission rates for the Tier 2 (labeled
as MY2010) and Tier 3 (labeled as MY 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.
0.12
0.00
•2010
•2025
10 15
Age (Years)
20
25
Figure 7-8 FTP Composite NOX emissions for reference (2010) and Tier 3 (2025)
constructed from MOVES rates
7-29
-------
0.16
•2010
•2025
0.00
0
10 15
Age (Years)
20
25
Figure 7-6 FTP Composite THC emissions constructed from MOVES rates
0.18
•2010
•2025
10 15
Age (Years)
20
25
Figure 7-7 US06 NOx emissions constructed from MOVES rates
7-30
-------
0.06
0.00
0
'2010
•2025
10 15
Age (Years)
20
25
Figure 7-8 US06 THC emissions constructed from MOVES rates
7.1.3.3.2 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 Section 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.
7.1.3.3.3 Gasoline Medium-duty HC/CO/NOX
The Tier 3 program will affect not only light-duty vehicles (below 8,500 pounds
GVWR), but also chassis-certified vehicles between 8,500 and 14,000 pounds GVWR. This
class of vehicles is referred to "light-heavy-duty" or "medium-duty" vehicles. In MOVES, these
vehicles are designated as regulatory class 41. This regulatory class comprises several types of
vehicles, including engine-certified trucks, Class 2b and Class 3 heavy trucks, and medium-duty
passenger vehicles (MDPV), which are not regulated under the medium-duty standards
considered here. As described in the reference-case medium-duty updates to MOVES,13 we
assumed that during this timeframe, engine-certified vehicles and medium-duty passenger
vehicles (MDPV) comprise five percent and fifteen percent of the regulatory class, respectively,
with the remainder composed of Class 2b and 3 trucks.
7-31
-------
Table 7-17 Population Percentage
Category
Engine -certified
MDPV
Class 2b
Class 3
Fraction of RegClass 4 1
(%)
5.0
15
60
20
The Class 2b and Class 3 vehicle program was modeled to begin in model year 2018 and
fully phase in during the 2022 model year (Figure 7-9). This projection 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.
2B and 3 Phase-in
550
500
450
400
JU
•| 350
tS 300
E 250
200
150
100
Class 2B
Class 3
2017
201B
2019
2020
2021
2022
MY
Figure 7-9 Class 2B and 3 Standard Phase-in
To represent mean emissions on the FTP for this regulatory class, we combined the Tier 3
phase-in for Class 2b and 3 vehicles with the existing emission standards for 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
1.2 times their standard on the chassis FTP.14 Calculated using the fractions in Table 7-17, the
weighted average of the Class 2b, Class 3, MDPV, and engine standards is 0.181 for model year
2022.
To account for the real-world performance of these vehicles, we related this average to
that for Tier 2 light-duty vehicles, for which we have a substantial volume of data from EPA's
in-use verification program. Using the same approaches used for developing the Tier 2 and Tier
3 light duty emission rates (see Section 7.1.3.3.1), we modeled MOVES Regulatory Class 41 in
7-32
-------
2022 as 90 percent Bin 5 and 10 percent Bin 8 vehicles.E For the phase-in years 2017-2021, we
calculated interim emission rates for each year as a weighted average of the Tier 3 rates and the
existing MOVES rates for regulatory class 41 (in MY 2016), such that the appropriate weighted
composite was calculated each year as shown in Table 7-18.
Table 7-18 Phase-in of MD Tier 3 Rates
Model Year
2017
2018
2019
2020
2021
2022
Tier 3 Rate
34%
49%
62%
75%
87%
100%
Current Rate
(MY2016)
66%
51%
38%
25%
13%
0%
Composite Standard
(g/mile)
0.33
0.29
0.26
0.24
0.21
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 g/hp-hr multiplied by 1.2),15
and for the Tier 3 MD vehicles, the standard ranges from 4.2 to 7.3 g/mile. Using the same
phase-in fractions as for NMOG+NOx, we calculated an aggregate CO standard of 4.4
grams/mile in 2022, 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.
As for light-duty vehicles, deterioration was modeled to represent a 150,000 mile useful
life. The same methodology was used for light-duty and medium-duty vehicles.
7.1.3.3.4 Diesel Medium-Duty HC/CO/NOX
For medium-duty (MD) diesel vehicles, the emission rates currently in MOVES imply
levels on the FTP substantially below the 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
E 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.
7-33
-------
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 Table 7-19. 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 7-19 Phase-in of MD Diesel Tier 3 NOX Rates
Model Year
2017
2018
2019
2020
2021
2022
Tier 3 Phase In
20%
38%
54%
69%
85%
100%
Reduction in NOx
Running Emission
Rate
12%
23%
33%
42%
52%
61%
Reduction in NOx
Start Emission Rate
5%
9%
12%
16%
19%
23%
7.1.3.3.5 Gasoline Particulate Matter (PM2.5)
Tier 3 will introduce standards for emissions of primary particulate matter (PM) for light-
duty vehicles. For the Federal Test Procedure (FTP), a full-useful life (FUL) standard of 3.0
mg/mi will apply. Additionally, for the US06, a standard of 10.0 mg/mi during the PM phase-in
and a final standard of 6.0 mg/mileF will apply to vehicles with GVWR up to 8,500 Ibs, as well
as medium-duty passenger vehicles (MDPV). Additionally, the full useful life (FUL) for most
vehicles will be increased from 120,000 to 150,000 miles under the Tier 3 rule. These standards
are targeting several processes that contribute to particulate matter in light-duty gasoline
vehicles: cold starts, high-power and high load operation, and deterioration of engine and
emission control technology over the life of the vehicle.
As mentioned in the preamble, most current new light-duty vehicles are effectively in
compliance with the 3.0 mg/mile standard (Section 1.5.1.1). The MOVES model includes strong
deterioration of light-duty gasoline emissions based on data collected in the Kansas City Vehicle
Emissions Study16, which causes predictions of fleet-average Tier 2 light-duty gasoline vehicles
to exceed the proposed 3 mg/mile standard after only 3-4 years of operation. We anticipate that
the Tier 3 PM standards will force continual improvement on fleet-wide PM emissions in several
ways. First, the rule will require reductions in emissions from the light-duty vehicles that
currently exceed the Tier 3 standard. Second, we expect that vehicle manufacturers will decrease
PM emissions in order to increase their compliance margin, which will help vehicles meet the
increased durability requirements for those test groups certified to the extended useful life of
150,000 miles. Third, we expect that manufacturers will lower US06 PM emissions in a similar
fashion in order to meet the respective standards. And finally, the Tier 3 standards will prevent
F The US06 PM standard will be 10 mg/mi starting and MY 2017 through MY 2021. After MY 2021, the US06 PM
standard will drop to 6 mg/mi.
7-34
-------
the backsliding of PM emissions from new vehicles with new technologies. The technologies
and techniques that manufacturers require to meet these standards are described in Chapter 1 of
this RIA.
To reflect the above projections, the emission rates in MOVES were modified in the
following ways. The light-duty PM2.5emission rates were reduced such that future vehicles
(when new) will meet the Tier 3 standards with a compliance margin of approximately 50
percent. MOVES estimates that current vehicles, within the first 3 years of use, emit PM2.5 on
the FTP at 2.72 mg/mile (cars), and 3.08 mg/mile (light trucks). To achieve a 50 percent
compliance margin with the fully-phased in Tier 3 rule, we estimated a reduction of 50 percent to
the Tier 3 standards. The reductions are applied:
• Uniformly to cars and trucks,
• Uniformly across start and running processes,
• Uniformly across all operating modes (including the ones that cover US06 type
driving)
The reductions in the PM2.5 emissions over the model years reflect the phase-in of the Tier 3
PM2.5 standards. The phase-in begins in 2017 for cars, and in 2018 for light trucks. By 2021, the
emission rates (described above) are fully phased-in for light-duty vehicles. Table 7-20 includes
the estimated FTP and US06 PM2.s for new (undeteriorated) vehicles, by model year. As shown
in the table, the reductions in PM2.5 yield compliance margins of more than 45 percent for both
the FTP and US06 standards.
7-35
-------
Table 7-20 Trends in PM2.5 Emissions on the FTP and US06 Cycles, by Model Year,
Reflecting the Phase-in of the Tier 3 Standards
Model
Year
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
Tier 3 phase-in
Car
0%
0%
0%
0%
0%
0%
10%
20%
40%
70%
100%
100%
100%
100%
100%
Truck
0%
0%
0%
0%
0%
0%
0%
20%
40%
70%
100%
100%
100%
100%
100%
FTP (mg/mile)
Car
2.72
2.72
2.72
2.72
2.72
2.72
2.59
2.45
2.18
1.77
1.36
1.36
1.36
1.36
1.36
Truck
3.08
3.08
3.08
3.08
3.08
3.08
3.08
2.77
2.46
2.00
1.54
1.54
1.54
1.54
1.54
US06 (mg/mile)
Car
2.94
2.94
2.94
2.94
2.94
2.94
2.79
2.64
2.35
1.91
1.47
1.47
1.47
1.47
1.47
Truck
4.85
4.85
4.85
4.85
4.85
4.85
4.85
4.36
3.88
3.15
2.42
2.42
2.42
2.42
2.42
PM2.5 emissions of the vehicle fleet deteriorate with age, due to wear and failure of
engine and emission control systems on some vehicles. The deterioration of the fleet-average
PM2.5 emission rates increase logarithmically with age as documented in the MOVES report10.
The emissions trend reflecting the durability associated with a FUL of 120,000 mi is shown in
Figure 7-10. The deterioration is applied multiplicatively to rates for "young" vehicles in the 0-3
yr ageGroup.
For vehicles manufactured after the onset of the Tier 3, the deterioration assumptions
were modified to represent an extension of the full useful life (FUL) from 120,000 mi to 150,000
mi. To model this assumption, we adjusted the deterioration trend such that emissions levels
were shifted to a point in age (or miles) that is 25% later (the value of 1.25 was calculated as the
ratio of the extended FUL to the previous FUL, or 150/120). For example, under the 120K
assumption, the fleet-average PM2.5 emission rate reaches 10 mg/mile after 14 years. Under the
150K scenario, the fleet-average PM2.5 emission rate does not reach 10 mg/mile until after
14*1.25 = 17.5 years.
Thus, the inventory modeling assumes a standard of 3.0 mg/mi and a 150,000 mi useful
life. However, under the final rule, manufacturers will retain the option of certifying some engine
test groups to a somewhat lower standard with a 120,000 mi useful life. Over the useful life of
vehicles, we assume that the two options (higher standard with longer useful life, lower standard
with shorter useful life) will yield approximately equivalent emissions trends over the life of the
vehicles. Therefore, no attempt was made to represent both options in the inventory modeling.
7-36
-------
9.00
ISOKmiFUL
120KmiFUL —
0
10 15
Vehicle Age (years)
20
25
Figure 7-10 Deterioration Trends for PM2.5, for Full Useful Lives of 120,000 and 150,000
miles
The discussion to this point has concerned total particulate (PM2.5). It is important to note
that the rates stored in the emissionRateByAge table in MOVES do not represent PM2.5 but
rather organic and elemental carbon components, OC2.5 and Ł€2.5, respectively. In the table,
these components are identified as pollutantID = 111 and 112 respectively. Starting with OC and
EC rates in the default database for MY 2016, the reductions described in Table 7-20 are applied
equally to OC and EC rates to generate corresponding rates under the Tier 3 standards.
7.1.3.3.6 Diesel PM
The Tier 3 controls were modeled as having no impact on light-duty diesel PM
emissions. Although light-duty diesels are subject to the same Tier 3 PM standards as light-duty
gasoline vehicles, all light-duty diesels are equipped with diesel particulate filters (DPF). The
application of a DPF results in diesel PM emissions being very close to the cleanest gasoline
vehicles in operation today. As a result the application of the Tier 3 PM standards to diesel
vehicles will not result in a change in diesel PM emissions.
7.1.3.3.7 Evaporative Emissions
The Tier 3 evaporative program requires lower emissions on the hot soak plus diurnal test
procedure on 9 RVP E10 certification fuel and strengthens in-use performance through a new
leak standard and OBD requirements for detection of vapor leaks. The new standards are
projected to result in significant reductions in evaporative hydrocarbon emissions. For this
analysis, tighter evaporative emission standards in conjunction with 9 RVP E10 certification fuel
are expected to reduce evaporative permeation emissions and fuel system venting, since the Tier
3 evaporative emission standards are aimed at not allowing any vented vapor emissions during
7-37
-------
the test. Moreover, the new requirements addressing leaks are expected to reduce the prevalence
(frequency rate) of fuel system vapor and liquid leaks.*3 The discussion below focuses on
reducing evaporative emissions by improving the permeation resistance of fuel tanks and vapor
lines and improved fuel system designs to reduce leaks. However, as mentioned above, and
discussed in Chapter 1, there are a variety of technologies manufacturers could use to achieve the
required reductions (including those which address vapor venting, permeation and leaks) and
manufacturers are likely to use the technologies they expect to get the largest reductions for the
lowest cost whether they be from reductions in fuel venting emissions, permeation, or leaks.
Furthermore, some technologies will be effective in achieving emission reductions against more
than one standard. For example, the canister honeycomb will help to meet the canister bleed and
Tier 3 hot soak plus diurnal standards and any measure to reduce leaks will help to meet the leak
standard and the hot soak plus diurnal standard. The technology choice is up to the manufacturer,
but we expect that the technologies used will to varying degrees address all three basic emission
types.
7.1.3.3.7.1 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 of about 50 percent from Tier 2 levels.
The Tier 2 permeation rate in MOVES is 0.0102 g/hour on EO 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 E10 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 emission rate in MOVES is developed by estimating permeation emissions
over one day of diurnal activity (65F-105F) on an ethanol-containing fuel and applying
reductions in the base rate over time. 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).
' One of the updates to MOVES for this analysis was to enable direct input of the leak prevalence rates.
7-38
-------
Table 7-21 Tier 3 Permeation Rates
Model Year
Tier 2
2016 2017
2018 2019
2020 2021
2022
Tier3
Phase-in
0%
40%
60%
80%
100%
Permeation
(g/hr)
0.0102
0.0072
0.0056
0.0041
0.0026
7.1.3.3.7.2 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. 7>18'19 The evaporative leak provisions
grew from this work, informing 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.
The impact of the Tier 3 evaporative emissions standards are quantified from MOVES,
which has been updated based on the new data collected. We expect that emissions reductions
will come from the reduction in the incidence rate of vapor leaks.
The vapor leak frequency estimates are generated from the high evaporative emissions
field study in Colorado during the summer of 2009.1? In that study, it was found that, compared
to the 1981-1995 model years pre-enhanced evaporative emissions vehicles, the vapor leak
frequency dropped significantly with the onset of the enhanced evaporative emissions standards
in 1996. The new standards did not explicitly require the measurement of vapor leaks through
OBD leak detection, but the standards encouraged changes in the materials and connections in
the fuel systems resulting in approximately 70 percent less vapor leaks. The upgraded materials
and connections for the following Tier 2 standards are estimated to have reduced vapor leak rates
another 33 percent.
We expect that manufacturers will again respond to the Tier 3 standards by making
further improvements, such as changes in materials and connections in order to reduce the
amount of fuel permeation and vapor leaks and will address canister bleed emissions. These
changes plus design enhancements (see Chapter 1, Section 1.6.1 "Tier 3 Evaporative
Emissions/Leak Control Technology Approaches" for more details) can be expected to further
reduce the amount of vapor leaks by one third, similar to the effect of Tier 2.
Tier 3 also includes provisions for an in-use measurement program that will explicitly
detect the remaining vapor leaks. The in-use leak standard will be enforced in the In-Use Vehicle
Program (IUVP) at different points in the vehicle's useful life. Tier 3 also imposes more
stringent requirements for evaporative on-board diagnostics (OBD) including a change in the
vapor leak check from a 0.040" equivalent diameter orifice in the current federal leak detection
7-39
-------
requirement to 0.020", as well as other more stringent readiness requirements which are already
in place for implementation in the State of California.
The analysis from the high evaporative emission field studies in Colorado (2008-2010)
90
found that evaporative OBD was only 30-50 percent effective in detecting vapor leaks. The
improvements to evaporative OBD in the Tier 3 rule are estimated to increase OBD vapor leak
detection to 80 percent effectiveness.
Based on the analysis as described above, the leak standard is therefore estimated to
result in an additional one third less evaporative vapor leaks. The prediction of an overall two
thirds reduction in vapor leaks is not unreasonable, considering the observed impact of the 1996
enhanced evap rule. Modeling details can be found in a separate memo to the docket.21
The leak prevalence rates in MOVES utilize observations that the State of Colorado
found in a follow-up study in 2010. Leaking vehicles were recruited in the Denver area to the
Colorado Tech Center and Laboratory to determine pre and post repair emissions. In this study
they found that approximately 70 percent of the evaporative leaks detected were due to the
deterioration of the evaporative and/or fuel system e.g. problems like corroded fuel lines, filler
neck, cracked hoses, etc. that could be repaired or improved with design and materials. The other
30 percent were due to issues beyond the manufacturers' control, such as improper maintenance
or missing gas caps.
tc
LJ Leaks Prevented
• Leaks w/ MIL on
• Leaks w/MIL off
Reference ;V III Case With Tier 3
Figure 7-10 Vapor Leak Control Assumed for 100 Leaking Vehicles
The Reference case in Figure 7-10 assumes 40 percent OBD effectiveness with 95
percent OBD readiness, resulting in 38 out of 100 MILs due to leaks turned on (the red portion
of first bar in above figure). This 40 percent OBD effectiveness is based on the analysis
comparing the evaporative system leak OBD DTCs to the portable SHED hot soak emissions in
the high evaporative emissions field studies in Denver (2008-2010), 30-50 percent of the time.
For the LEV III case (the overall nationwide situation without Tier 3), it is assumed that 33
percent (the one third discussed as Tier 3 permeation benefit, similar to Tier 2 benefit) of 70
7-40
-------
percent (-23 percent) are prevented with the reduction in SHED test standard (white portion of
middle). It is assumed that 80 percent of the remaining leaks will have a light on (red portion in
middle bar). The blue portion of the middle bar is the leaks remaining undetected in the LEV III
case.
Based on Tier 3 control, we assume 66 percent or two thirds (one third for Tier 3
permeation benefit in previous paragraph plus one third discussed earlier as leak standard
benefit, equals two thirds emissions benefit for Tier 3) of the 70 percent durability related leaks
(-46 percent) are prevented (white portion in third bar).The modeling then assumes that 80
percent of the remaining leaks will have an evaporative system leak OBD light on as in the LEV
III case (the red portion of the third bar). The remaining without lights is 11 out of 100, in the
blue portion of the third bar in the figure above.
7.1.3.3.7.3 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 remaining 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 7-22 Reductions of Liquid Leaks in Tier 3
Age
0-9
10-14
15-19
20+
Operating
45%
30%
0%
0%
Hot Soak
45%
30%
0%
0%
Cold Soak
45%
30%
0%
0%
7.1.3.4 Updates to MOVES Sulfur Effects
In order to evaluate the emission impacts of the sulfur standards, the version of MOVES
used for this analysis made significant updates to the effect of fuel sulfur levels below 30 ppm on
exhaust emissions. In MOVES2010b, these effects were based on extrapolation of data on sulfur
79
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 final report on this
research is available in the docket.23
7.1.3.4.1 EPA Testing of Gasoline Sulfur Effects on Tier 2 Vehicles and the In-Use
Fleet
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.
7-41
-------
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.24
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 converter 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 (MS AT) rule, found significant
reductions in NOx, CO and total HC when the vehicles were tested on low sulfur fuel, relative to
32 ppm fuel.25 In particular, the study found a nearly 50 percent increase in NOx when sulfur
was increased from 6 ppm to 32 ppm. Given the preparatory procedures related to catalyst
clean-out and loading used by these studies, these results may represent a "best case" scenario
relative to what will be expected under more typical driving conditions. Nonetheless, these data
suggested the effect of in-use sulfur loading was largely reversible for Tier 2 vehicles, and that
there were likely to be significant emission reductions possible with further reductions in
gasoline sulfur level. 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.26 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 what the state of catalyst sulfur loading is for the typical vehicle in the field. In
response to these data needs, 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 93 cars and light trucks recruited
from owners in southeast Michigan, covering model years 2007-9 with approximately 20,000-
40,000 odometer miles.H 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 test fuel, one at a sulfur level of 5
H The NPRM modeling was based on analysis of 81 passenger cars and trucks. Since the NPRM, twelve additional
Tier 2 vehicles were tested and included in the statistical analysis described in the docketed final report, examining
the effect of sulfur on emissions from Tier 2 vehicles. The analysis based on the complete set of 93 Tier 2 vehicles
is reflected in the results presented in this section and the emissions modeling for FRM.
7-42
-------
ppm and the other at 28 ppm. 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 as well as 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, demonstrating that sulfur loadings have a large effect on exhaust catalyst
performance, and that Tier 2 vehicles can achieve significant reductions based on removing, at
least in part, the negative impact of the sulfur loading on catalyst efficiency (Table 7-23). For
example, Bag 2 NOx emissions dropped 31 percent between the pre- and post-cleanout tests on
28 ppm fuel.
Table 7-23 Percent Reduction in In-Use Emissions after the Clean-out
Using 28 ppm Test Fuel
Bagl
Bag 2
Bag3
FTP Composite
Bag 1 - Bag 3
NOX
(p-value)
—
31.4%
(0.0003)
35.4%
(O.OOOl)
11.4%
(0.0002)
—
THC
(p-value)
—
14.9%
(0.0118)
20.4%
(O.OOOl)
3.8%
(0.0249)
—
CO
(p-value)
6.0%
(0.0151)
—
21.5%
(0.0001)
6.8%
(0.0107)
7.2%
(0.0656)
NMHC
(p-value)
—
18.7%
(0.0131)
27.7%
(O.OOOl)
3.5%
(0.0498)
—
CH4
(p-value)
—
14.4%
(0.0019)
10.3%
(O.OOOl)
6.0%
(0.0011)
—
PM
(p-value)
15.4%
(< 0.0001)
—
24.5%
(O.OOOl)
13.7%
(O.OOOl)
—
The clean-out effect is not significant at a = 0.10, when no reduction estimate is provided.
7-43
-------
To assess the impact of lower sulfur fuel on in-use emissions, further testing was
conducted on a representative subset of vehicles on 28 ppm and 5 ppm fuel with accumulated
mileage. A first step in this portion of the study was to assess the differences in the effectiveness
of the clean-out procedure under different fuel sulfur levels. Table 7-24 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 34
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. This further
indicates that current sulfur levels in gasoline continue to have a long-term, adverse effect on
exhaust emissions control that is not fully removed by intermittent clean-out procedures that can
occur in day-to-day operation of a vehicle and demonstrates that lowering sulfur levels to 10
ppm on average will significantly reduce the effects of sulfur impairment on emissions control
technology.
Table 7-24 Percent Reduction in Exhaust Emissions When Going from 28 ppm to 5 ppm
Sulfur Gasoline for the First Three Repeat FTP Tests Immediately Following Clean-out
Bagl
Bag 2
Bag3
FTP
Composite
Bag 1 - Bag 3
NOX
(p- value)
5.3%
(0.0513)
34.4%
(0.0036)
42.5%
(O.OOOl)
15.0%
(0.0002)
_?
THC
(p-value)
6.8%
(0.0053)
33.9%
(O.OOOl)
36.9%
(O.OOOl)
13.3%
(O.OOOl)
_?
CO
(p-value)
6.2%
(0.0083)
_?
14.7%
(0.0041)
8.5%
(0.0050)
_?
NMHC
(p-value)
5.7%
(0.0276)
26.4%
(0.0420)
51.7%
(O.OOOl)
10.9%
(0.0012)
_?
CH4
(p-value)
14.0%
(O.OOOl)
49.4%
(O.OOOl)
28.5%
(O.OOOl)
23.6%
(O.OOOl)
_?
PM?
—
—
—
—
-
1 The effectiveness of clean-out cycle is not significant at a = 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
Table 7-25. Reductions for Bag 2 NOx were particularly high, estimated at 52 percent between
28 ppm and 5 ppm overall. For all pollutants, 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.
7-44
-------
Table 7-25 Percent Reduction in Emissions from 28 ppm to 5 ppm Fuel Sulfur
on In-use Tier 2 Vehicles
Bagl
Bag 2
Bag3
FTP
Composite
Bagl-
Bag3
NOX
(p-value)
7.1%
(0.0216)
51.9%
(< 0.0001)
47.8%
(< 0.0001)
14.1%
(0.0008)
_ t
THC
(p-value)
9.2%
(0.0002)
43.3%
(< 0.0001)
40.2%
(< 0.0001)
15.3%
(< 0.0001)
5.9%
(0.0074)
CO
(p-value)
6.7%
(0.0131)
_t
15.9%
(0.0003)
9.5%
(< 0.0001)
_t
NMHC
(p-value)
8.1%
(0.0017)
42.7%
(0.0003)
54.7%
(< 0.0001)
12.4%
(< 0.0001)
_ t
CH4
(p-value)
16.6%
(< 0.0001)
51.8%
(< 0.0001)
29.2%
(< 0.0001)
29.3%
(< 0.0001)
_ t
NMOG+NOX
(p-value)
N/A
N/A
N/A
14.4%
(< 0.0001)
N/A
PM
t
-
-
-
-
-
T Sulfur level not significant at a = 0.10.
* Inconclusive because the mixed model did not converge.
Major findings from this study include:
• Largely 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 current levels to levels in the range of the
gasoline sulfur standards will be expected to achieve significant reductions in
emissions of NOx, hydrocarbons, and other pollutants of interest in the current in-
use fleet.
• Assuming that the emissions impacts vs. gasoline sulfur content are
approximately linear, changing gasoline sulfur content from 30 ppm to 10 ppm
would result in NMOG+NOX emissions decreasing from 52 mg/mi to 45 mg/mi,
respectively (a 13% decrease), and NOX emissions decreasing from 19 mg/mi to
16 mg/mi, respectively (a 16% decrease), for the vehicles in the study.
To evaluate the robustness of the statistical analyses assessing the overall in-use
emissions reduction between operation on high and low sulfur fuel (Table 7-25), a series of
sensitivity analyses were performed to assess the impacts on study results of measurements from
low-emitting vehicles and influential vehicles, as documented in detail in the report.27 The
sensitivity analyses showed that the magnitude and the statistical significance of the results were
not impacted and thus demonstrated that the results are statistically robust. We also subjected the
design of the experiment and data analysis to a contractor-led independent peer-review process
in accordance with EPA's peer review guidance. The results of the peer review28'29 largely
supported the study design, statistical analyses, and the conclusions from the program and raised
7-45
-------
only minor concerns that have not changed the overall conclusions and have subsequently been
addressed in the final version of the report.30
Overall, the reductions found in this study are in agreement with other low sulfur studies
conducted on Tier 2 vehicles, namely MSAT and Umicore studies mentioned above, in terms of
the magnitude of NOx and HC reductions when switching from 28 ppm to 5 ppm fuel.31'32 For
additional details on the impact of gasoline sulfur control on exhaust emissions, see Section
IV. A. 6 of the preamble.
7.1.3.4.2 Implementation in MOVES
The fuel sulfur effect applies multiplicatively in conjunction with other gasoline fuel
effects in MOVES. The results shown in Table 7-25 were applied in MOVES for model year
2001 and later gasoline vehicles to estimate the sulfur effects at or below 30 ppm. For sulfur
levels above 30 ppm, and for all pre-2001 model year vehicles, the sulfur effect originally
implemented in MOVES2010b remains in place.
Equation 7-3 shows the generic form of the low sulfur model applied to model year 2001
and later gasoline vehicles.
sulfur adjustment^ j= 1.0-/?s fbase-xs^
Equation 7-3 Low Sulfur Model
The sulfur coefficients (/?s) were developed by linearly interpolating between emission
levels at 28 to 5 ppm, corresponding to the reductions in emissions shown in Table 7-25, and
standardizing to sulfur level of 30 ppm. The sulfur coefficient simply represents the slope of the
interpolated line. The emission reductions from FTP bag 2 and FTP bagl-bag3 were used to
calculate the sulfur coefficients for running exhaust and start exhaust, respectively. Table 7-26
shows the resulting sulfur coefficients applied in MOVES by pollutant, process, and vehicle
type.
Table 7-26 Low Sulfur Coefficients by Vehicle Type, Process and Pollutant
Vehicle Type
Motorcycle
Passenger Car,
Passenger Truck &
Light Commercial Truck
All other Vehicle Types
THC
Starts
0
0.00257
0
Running
0
0.01813
0.015488
CO
Starts
0
0
0
Running
0
0
0.009436
NOX
Starts
0
0
0
Running
0
0.02158
0.027266
PM
Starts
0
0
0
Running
0
0
0
The sulfur base (Sbase) in the low sulfur model varies as a function of the scenario being
modeled and the model year group (see Table 7-27). For most states, the sulfur base is 30 ppm
except for 2017 and later model years for the control scenario. For California, the sulfur base is
10 ppm for all model years and scenarios. For the Section 177 states that have adopted the LEV
7-46
-------
Ill program, the sulfur base of 30 ppm and 10 ppm is used for the model years 2001 to 2016 and
model years 2017 and later, respectively.
Table 7-27 Sulfur Base in Low Sulfur Model
Most States
California
Section- 177
States
Reference
MYG 200 1-20 16
30 ppm
10 ppm
30 ppm
MYG 20 17+
30 ppm
10 ppm
10 ppm
Control
MYG 200 1-20 16
30 ppm
10 ppm
30 ppm
MYG 20 17+
10 ppm
10 ppm
10 ppm
These equations were then used to populate the database table that stores fuel effect
equations in the MOVES database ("GeneralFuelRatioExpression"). This table allows the
MOVES model to compute fuel effects based on the properties of any fuel contained in the
"FuelSupply" and "FuelFormulation" database tables. Additional details are documented in the
docket memo addressing MOVES updates.
7.1.4 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 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 decrease in gasoline sulfur levels. As with onroad, reference and
control case emissions were generated using the fuel supply inputs reflecting the projected fuel
volumes from AEO2013.
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 (NMEVI).33 The fuels in the NMEVI database, NCD2010201a,
were developed from the reference and control fuels used for onroad vehicles, as described in
Section 7.1.3.2. Onroad and nonroad gasoline formulations are assumed to be identical for all
years. In 2018 and 2030, nonroad equipment is assumed to use E10 only. For all years, the
reference case included the higher sulfur reference gasoline and the control case met the sulfur
limits.
Since aircraft, locomotive and commercial marine emission sources do not burn gasoline,
their emission factors are unaffected by the sulfur 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 the ones estimated for the Heavy-
Duty Greenhouse Gas Rule (2011).34 Estimation of emissions from locomotives and C1/C2
commercial marine used the same procedures developed for the Locomotive Marine Rule
(2008), detailed in the RIA for that rule.35 The procedures used for calculating C3 commercial
marine emissions are those developed in the recent C3 Rule (2010).36
7-47
-------
7.1.5 Criteria and Toxic Emission Impact Results
The Tier 3 rule will reduce NOx (including NCh), VOC, and 862 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 and CO 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 862 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 Table 7-28 for calendar years 2018 and 2030. We project
significant reductions immediately upon implementation of the program, growing to a nearly 25
percent reduction in onroad emissions by 2030. We project a nearly 33 percent reduction in
onroad emissions in 2050, when the fleet will have fully turned over to vehicles meeting the fully
phased in Tier 3 standards.
Table 7-28 Tier 3 NOX Reductions by Calendar Year (Annual U.S. Short Tons)
Year
2018
2030
Onroad
mobile
reference
2,753,732
1,331,788
Onroad
mobile
with
control
2,489,364
1,003,279
Reduction
264,369
328.509
Percent
reduction
in onroad
9.6%
24.7%
Table 7-29 shows the reduction in NOx emissions, in annual short tons, projected in
calendar years 2018 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 2018 over 90 percent of the program reductions are coming from lower
sulfur gasoline on the fleet already on the road. By 2030, over 80 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.1
Table 7-29 Projected NOx Reductions from Tier 3 Program (Annual U.S. Short Tons)
Total reduction
Reduction from pre-Tier 3
fleet due to sulfur standard
Reduction from Tier 3 fleet
due to vehicle and sulfur
standards
2018
264,369
242,434
21,934
2030
328,509
56,324
272,185
1 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
7-48
-------
VOC reductions are shown in Table 7-30 for calendar years 2018 and 2030. We project
reductions of over 40,000 tons (3 percent of the onroad fleet emissions) immediately upon
implementation of the program, growing to a 16 percent reduction in onroad emissions by 2030.
We project a 28 percent reduction in onroad emissions in 2050, when the fleet will have fully
turned over to vehicles meeting the fully phased in Tier 3 standards.
Table 7-30 Tier 3 VOC Reductions by Calendar Year (Annual U.S. Short Tons)
CY
2018
2030
Onroad
mobile
reference
1,703,902
1,078,892
Onroad
mobile with
control
1,656,399
911,301
Reduction
47,504
167,591
Percent
reduction in
onroad
2.8%
15.5%
Table 7-31 shows the VOC reductions in 2018 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 exhaust and evaporative standards. In 2018, 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 close to 40 percent reduction in VOC in 2030.
Table 7-31 Projected VOC Reductions from Tier 3 Program (Annual U.S. Short Tons)
Total reduction
Reduction from pre-Tier 3
fleet due to sulfur standard
Reduction from Tier 3 fleet
due to vehicle and sulfur
standards
Exhaust
Evaporative
2018
47,504
38,786
8,718
43,009
4,495
2030
167,591
11,249
156,343
105,253
62,339
CO reductions are shown in Table 7-32 for calendar years 2018 and 2030. We project
significant reductions immediately upon implementation of the program, growing to a 24 percent
reduction in onroad emissions by 2030. We project a 38 percent reduction in onroad emissions
in 2050, when the fleet will have fully turned over to vehicles meeting the fully phased-in Tier 3
standards.
Table 7-32 Tier 3 CO Reductions by Calendar Year (Annual U.S. Short Tons)
CY
2018
2030
Onroad
mobile
reference
17,517,356
14,663,722
Onroad
mobile with
control
17,238,477
11,205,680
Reduction
278,879
3,458,041
Percent
reduction
in onroad
1.6%
23.6%
7-49
-------
Table 7-33 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 only about 40 percent of total reductions in 2018. By 2030, the Tier 3
fleet is accounting for 99 percent of program reductions. Of the Tier 3 vehicle standard
reductions in 2030, we estimate that about 5 percent are contributed by the heavy-duty tailpipe
standards.
Table 7-33 CO Reductions from Tier 3 Program (Annual U.S. Short Tons)
Total reduction
Reduction from pre-Tier 3
fleet due to sulfur standard
Reduction from Tier 3 fleet
due to vehicle and sulfur
standards
2018
278,879
122,171
156,708
2030
3,458,041
17,734
3,440,307
Direct PM2.5 impacts are shown in Table 7-34 for calendar years 2018 and 2030. For
direct PM, the impact shown is solely from the 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 7,900 tons annually, which represents approximately 10 percent
of the onroad direct PM2.5 inventory. The relative reduction in onroad emissions is projected to
grow to 28 percent in 2050, when the fleet will have fully turned over to vehicles meeting the
fully phased-in Tier 3 standards.
Table 7-34 Tier 3 PM2.5 Reductions by Calendar Year (Annual U.S. Short Tons)
CY
2018
2030
Onroad mobile
reference
115,560
78,320
Onroad mobile
with control
115,430
70,428
Reduction
130
7,892
Percent
reduction in
onroad
0.1%
10.1%
Emissions of air toxics also will be reduced by the 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 7-35 presents reductions for
certain gaseous air toxics and polycyclic aromatic hydrocarbons (PAHs)J, reflecting reductions
of a few percent in 2018, and 10 to 30 percent of onroad emissions, depending on the individual
pollutant, in 2030.
1 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(l,2,3,cd)pyrene, phenanthrene, and pyrene. These PAHs are
included inEPA's national emissions inventory (NEI).
7-50
-------
Table 7-35 Reductions for Certain Individual Compounds (Annual U.S. Short Tons)
Acetaldehyde
Formaldehyde
Acrolein
1,3 -Butadiene
Benzene
Naphthalene
Ethanol
2,2,4-
Trimethylpentane
Ethyl Benzene
Hexane
Propi onal dehy de
Styrene
Toluene
Xylene
PAHs
2018
Reduction
600
513
40
257
1,916
99
2,704
806
761
1,112
27
71
3,772
2,894
10
Percent reduction
in onroad
3.4%
2.3%
2.6%
5.2%
5.8%
3.0%
1.7%
1.9%
2.8%
3.2%
2.7%
5.1%
2.2%
2.9%
1.2%
2030 Reduction
2,067
1,277
127
677
4,762
269
19,950
3,827
2,451
4,132
63
242
15,261
9,396
58
Percent reduction
in onroad
20.6%
9.9%
15.0%
29.3%
26.4%
15.2%
15.8%
11.9%
14.3%
19.0%
16.9%
29.5%
12.9%
15.1%
18.1%
The totals shown in Table 7-36 represent the sum of all toxic species listed in Table 7A-1
of the Appendix, including the species in Table 7-35. As shown, in 2030, the overall onroad
inventory of total toxics will be reduced by about 15 percent, with nearly one half of the
reductions coming from the evaporative standards.
Table 7-36 Reductions in Total Mobile Source Air Toxics (Annual U.S. Short Tons)
Total reduction
Reduction from pre-Tier 3 fleet due to
sulfur standard
Reduction from Tier 3 fleet due to
vehicle and sulfur standards
Exhaust
Evaporative
2018
15,583
11,981
3,602
13,340
2,243
2030
64,558
3,517
61,041
34,595
29,963
SC>2 emissions from mobile sources are a direct function of sulfur in the fuel, and
reducing sulfur in gasoline would result in immediate reductions in 862 from the on and off-road
fleet. The reductions, shown in Table 7-37, represent over 50 percent reduction in onroad SC>2
emissions. The breakdown of the relative contribution of onroad 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.37
7-51
-------
Table 7-37 Projected SOi Reductions from Tier 3 Program (Annual U.S. Short Tons)
Total reduction
Reduction from onroad
vehicles due to sulfur
standard
Reduction from off-road
equipment due to sulfur
standard
Percent reduction in
onroad SOi emissions
2018
15,565
14,813
752
56%
2030
13,261
12,399
862
56%
7.2 Criteria and Toxic Pollutant Air Quality Impacts
7.2.1 Emission Inventories for Air Quality Modeling
To estimate the benefits of the Tier 3 rule, we performed air quality modeling for the
years 2018 and 2030. 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 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.38
7.2.1.1 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. This additional 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
7-52
-------
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.
Because of the differences in methodology between the national inventories and air
quality inventories, particularly the treatment of local variables such as vehicle speed
distributions and the handling of 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. This is pronounced, especially, 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.
In addition to the methodological differences, because this modeling methodology with
added precision is time consuming and resource intensive, the inventories for the air quality
modeling had to begin months before the national inventory modeling and used a slightly older
version of the MOVES model. The model and the input differences between the national
inventory and the inventories developed for air quality modeling are described below.
Phase-in Assumptions for LDT3 andLDT4
As described in Section 7.1.3.3.1, for the national inventory, the onset of the Tier 3
phase-in for gaseous pollutants (HC, CO, NOx) starts in 2018 for truck classes with gross vehicle
weight ratings greater than 6,000 Ibs. (LDT3 and LDT4), in contrast to LDV, LDT1, and LDT2,
which begin phasing-in in 2017. However, for the air quality modeling, the phase-in was
modeled uniformly in 2017 for all light-duty fleet, resulting in slight overestimation of the Tier 3
emission reductions in calendar year 2018.
E200/E300 to T50/T90 Conversion
While the primary impact of gasoline sulfur control is the changes in sulfur content of the
fuel, we do expect slight changes in other fuel properties, including fuel distillation, as discussed
in Section 7.1.3.2. Fuel distillation is one of a number of fuel properties that have been found to
impact vehicle emissions. Specific fuel properties in question are the T50 and T90 of the
gasoline (the temperature at which 50 percent and 90 percent of the fuel is evaporated,
respectively), which are often discussed instead in terms of E200 and E300 (the percent of the
fuel evaporated at 200°F and 300°F, respectively). Both T50 and T90 are among the fuel
properties modeled in MOVES that have impacts on emissions. Thus, estimating the overall
impact of Tier 3 sulfur control on emissions requires the characterization of the changes in fuel
distillation resulting from changes in gasoline sulfur.
Both T50/T90 and E200/E300 data is available from the fuel compliance database. For
the reference case, the values for T50 and T90 were derived directly from the fuel compliance
database. However, the changes for the control case were modeled using a refinery model which
only provides changes in the form of E200 and E300. Therefore, we applied changes in
E200/E300 from the refinery modeling to E200 and E300 values derived from the fuel
compliance database, and then converted into T50 and T90 values using characteristic
transformation equations.
7-53
-------
Typically, E200/E300 and T50/T90 values are well correlated using the transformation
equations. However, if the distillation properties vary in a non-typical way, these
transformations can provide inaccurate results compared to tested values of E200/E300 and
T50/T90. For the 2012 fuel compliance data, fuels in all regions except the West and Pacific
Northwest behave in a typical way regarding distillation properties and characteristic
transformations between those properties. However, in the West and Pacific Northwest, when
the transformation equation was used, E200/E300 did not correlate with T50/T90 as was
expected.
The choice to start with T50/T90 data from the fuels compliance database for the
reference case and E200/E300 data from the same database for the control case resulted in
unexpected differences in fuel properties between the two scenarios, and in turn, affected the
emissions inventory in the West and Pacific Northwest. These effects on the emissions
inventory are not real effects; they are simply an artifact of our methodology. In summary, for
air quality modeling inputs in the West and Pacific Northwest, there is up to a 5 percent increase
in T50 and T90 values in the control case compared to the reference case that is caused by
improper translation from E200/E300 and not caused by a real change in these data. This error
was corrected for the inputs used in the national inventory modeling.
Evaporative Emissions
Overall, the updates made to the evaporative emissions for the national inventory resulted
in reduced evaporative emissions inventory by 10 to 20 percent. Most of the differences between
the national inventory and the inventories for air quality modeling results from a change in the
vapor venting leak prevalence rates. The distribution of leak sizes shifted towards smaller leaks
causing a reduction in emissions from leaks. An error was also corrected in the temperature
adjustment coefficients for running loss emissions, resulting in lower running loss emissions.
There was another update to correct an error in the database for the LEV Ill/Section 177 States'
I/M versus non-I/M rates, which was very minor. Additional details are provided in the docket
memo documenting the MOVES updates.
Modeling of I/M Programs
For states that have I/M programs and submitted county databasesK, the inventories for
air quality modeling did not properly account for the I/M program, resulting in overestimation of
the inventory for both the reference and the control cases. However, the emission reductions
from Tier 3 are not affected.
Modeling of Section 177 States
The sulfur base (see Table 7-27) for the three Non-Section 177 states (VA, NH, and DC)
were inadvertently modeled as Section 177 states in the inventories for air quality modeling. For
these states, the emission reductions from Tier 3 were underestimated.
K A total of 411 counties were affected. For additional detail, please refer to the docket memo describing MOVES
updates.
7-54
-------
The updates and the fixes to the above described issues were included only in the national
inventories. However, the fundamental MOVES updates incorporating new research, the custom
inputs developed for the reference and control scenarios, and the modeling of LEV III programs,
described in detail in Section 7.1.3, were also included in the inventories for the air quality
modeling.
The two sets of results are compared in Table 7-38 below.
Table 7-38 Comparison of
Inventories
Calendar Year 2030 Onroad Emission National Inventories and
Used for Air Quality Modeling [U.S. Short tons]
Pollutant
NOX
VOC
CO
PM2.5
Benzene
Ethanol
Acrolein
1,3-Butadiene
Formaldehyde
Acetaldehyde
SO2
Reference
National
Inventory
1,317,412
1,065,457
14,480,459
77,838
17,772
124,933
839
2,285
12,846
9,946
21,973
Air Quality
Inventory
1,367,429
1,260,883
14,434,283
82,138
18,742
150,592
852
2,288
13,454
10,143
22,310
Difference
AQ vs. NI
4%
18%
0%
6%
6%
21%
2%
0%
5%
2%
2%
Control
National
Inventory
993,724
900,680
11,085,515
70,048
13,092
105,324
714
1,619
11,588
7,909
9,734
Air Quality
Inventory
1,018,962
1,079,042
10,780,871
73,620
13,603
129,002
717
1,572
12,092
7,957
10,040
Difference
AQ vs. NI
3%
20%
3%
5%
4%
23%
1%
3%
4%
1%
3%
Because the reference and control case emissions rate inputs were the same for the
national inventory and air quality inventory runs, the percent reductions due to the Tier 3 rule are
very similar, considering the differences described earlier, as shown in Table 7-39.
7-55
-------
Table 7-39 Comparison of Emission Reductions from Reference to Control Case in
"National" and "Air Quality" Onroad Inventories
Pollutant
NOX
VOC
CO
PM2.5
Benzene
Ethanol
Acrolein
1,3 -Butadiene
Formaldehyde
Acetaldehyde
S02
2018
National
Inventory
Reduction
-9.6%
-2.8%
-1.6%
-0.1%
-5.8%
-1.7%
-2.6%
-5.2%
-2.3%
-3.3%
-56.3%
Air Quality
Inventory
Reduction
-9.9%
-2.4%
-1.6%
-0.4%
-6.4%
-1.3%
-2.8%
-5.7%
-2.3%
-3.5%
-55.9%
2030
National Inventory
Reduction
-24.6%
-15.5%
-23.4%
-10.0%
-26.3%
-15.7%
-14.9%
-29.2%
-9.8%
-20.5%
-55.7%
Air Quality
Inventory
Reduction
-25.5%
-14.4%
-25.3%
-10.4%
-27.4%
-14.3%
-15.8%
-31.3%
-10.1%
-21.5%
-55.0%
inventories.
The following sections summarize the analysis done to generate the air quality
iries.
7.2.1.1.1 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 not available for all counties. 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 grouped counties together with counties
with similar characteristics to generate emission rates that can be used in all of the counties in the
grouping.
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 representing counties are chosen so that they can be used to compute rates, such as
g/mi factors, that will be representative across the group of counties. To assure this, the counties
are grouped based on vehicle age, fuel parameters, emission standards, I/M programs and
altitude. However, representative counties are not meant to represent VMT. VMT is estimated
for every Continental U.S. county. As explained in Section 7.2.1.1.3, the SMOKE model
calculates emissions by multiplying the county-specific VMT by the county-group specific g/mi
emission rates produced in the MOVES run. The characteristics used to group the counties are
summarized in Table 7-40 below.
7-56
-------
Table 7-40 Characteristics for Representing County Groupings
County Grouping Characteristic
State
Fuel Parameters
Emission Standards
Inspection/Maintenance Programs
Altitude
Vehicle Age
Description
Counties in each group must be in the same state as the
representing county.
Average gasoline fuel properties for January and July
2005, including RVP, sulfur level, ethanol fraction and
percent benzene
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.
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.
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 average age of passenger vehicles is calculated for
each county. The counties in each group must be in the
same average age category as the representing county.
The result is a set of 146 county groups with similar ages, fuel, emission standards,
altitude and I/M programs in each state. For each group, the county with the highest VMT was
chosen as the representing county. Only these 146 counties 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.
39
7.2.7.7.2 SMOKE-MOVES
The official EPA highway vehicle emissions model (MOVES) was updated as described
in Section 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.40 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
7-57
-------
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.41
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. A series of post-
processing scripts was developed to take the MOVES emission rate tables and translate them 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. 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.
7.2.1.1.3 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 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, VMT, age distributions, and inspection-maintenance programs for each of the
representing counties. The source data for each of these inputs is described below.
7.2.1.1.3.1 Temperature and Humidity
Ambient temperature can have a large impact on emissions. Low 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.
The 12-km gridded meteorological input data for the entire year of 2007 covering the
continental United States were derived from simulations of version 3.1 of the Weather Research
and Forecasting Model42, Advanced Research WRF43 core. The WRF Model is a mesoscale
numerical weather prediction system developed for both operational forecasting and atmospheric
research applications. The Meteorology-Chemistry Interface Processor (MCIP)44 version 3.6
was used as the software for maintaining dynamic consistency between the meteorological
model, the emissions model, and air quality chemistry model.
EPA applied the SMOKE-MOVES tool, Met4moves, to the gridded, hourly
meteorological data (output from MCIP) to generate a list of the maximum temperature ranges,
average relative humidity, and temperature profiles that are needed for MOVES to create the
emission-factor lookup tables. "Temperature profiles" are arrays of 24 temperatures that
describe how temperatures change over a day, and they are used by MOVES to estimate vapor
7-58
-------
venting emissions. The hourly gridded meteorological data (output from MCIP) was also used
directly by SMOKE.
The temperature lists were organized based on the representative counties and fuel
months as described in Sections 4.6.4.1 and 4.6.4.2, respectively in the documentation for the
2008 National Emission Inventory.45 Temperatures were analyzed for all of the counties that are
mapped to the representative counties, i.e., for the county groups, and for all the months that
were mapped to the fuel months. EPA used Met4moves to determine the minimum and
maximum temperatures in a county group for the January fuel month and for the July fuel month,
and the minimum and maximum temperatures for each hour of the day. Met4moves also
generated idealized temperature profiles using the minimum and maximum temperatures and 10
degree intervals. In addition to the meteorological data, the representative counties and the fuel
months, Met4moves uses spatial surrogates to determine which grid cells from the
meteorological data to collect temperature and relative humidity statistics. For example, if a
county had a mountainous area with no roads, this would be excluded from the meteorological
statistics. The output for the daily mode is one temperature range per county per day and is a
more detailed approach for modeling the vapor venting emissions. EPA ran Met4moves in daily
mode for 2007 base year.
The treatment of humidity was simpler. The humidity values that correspond to each
temperature value in each temperature bin are averaged and used as the humidity for calculations
for that temperature bin. Each set of temperature bins for a grouping of counties will have its
own set of corresponding humidity values. Humidity affects the formation of oxides of nitrogen
(NOx) during combustion and the calculation of air conditioning load effects.
2007 calendar year temperatures and humidity values described above were used for the
2018 and 2030 projection years as well.
7.2.1.1.3.2 Required Vehicle Population and VMT Inputs
Vehicle population and vehicle miles traveled (VMT) data are required input for MOVES
when modeling on a county basis. Using the technical guidance provided to states by EPA, a
contractor generated appropriate national estimates for vehicle populations and VMT 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."46
7.2.1.1.3.3 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
7-59
-------
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 2008 NEI. These inputs
were then converted to a format consistent with MOVES.
7.2.1.1.4 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.
The annual vehicle miles traveled (VMT) values calculated for calendar year 2007 are
estimated using VMT estimates from the Federal Highways Administration (FHWA) for 2007
and 2008, combined with the state supplied VMT estimates submitted for the 2008 calendar year.
The FHWA estimates, found in the Vehicle-miles of travel by functional system table (VM-2)
can be obtained from the web at:
http://www.fhwa.dot.gov/policyinformation/statistics/2007/
http://www.fhwa.dot.gov/policyinformation/statistics/2008/
The VMT data from the VM2 tables are broken out by state and HPMS road type. The
VMT values from both 2007 and 2008 are placed into a single table (matched on state and road
type) and an adjustment factor (2007 VMT / 2008 VMT) is calculated for each state and road
type.
The VMT used for the 2008 NEI is obtained from the FF10 format file (by county and
SCC) used with SMOKE for Version 3 of the 2008 NEI
(VMT_NEI_2008_updated2_18jan2012_v3.csv). The development of the 2008 VMT estimates
using state supplied data and FHWA estimates is described in the technical support document for
2008 NEI.47
The 2007 VMT values were calculated by applying the adjustment factors calculated
from the FHWA tables to the appropriate rows in the 2008 VMT data, matching on state and
HPMS road type. The same adjustment was used for all counties in a state and that all Source
Classification Code (SCC) vehicle types used the same adjustment for each road type.
The 2018 and 2030 VMT was created by multiplying the base year 2007 platform VMT
by growth adjustment factors. The adjustment factors are at the state level by the 12 SCC
vehicle types. The resulting total VMT values by SCC in 2018 are normalized to match the total
VMT by SCC from MOVES runs using VMT projections from the "early release" of the 2013
AEO. Adjustment factors by state were derived using the National Mobile Inventory Model
(NMIM) County Database (NCD) growth from 2007 to 2018 and 2018 to 2030 to account for the
relative growth among states. The NCD20080522 database contains the most recent county
specific VMT projections available. So, while the AEO projections are used to calculate national
totals, the relative growth among states is derived from the NCD.
7-60
-------
Once the annual VMT by county has been estimated, the population can be determined
by assigning a fixed average VMT to each vehicle. The VMT for each SCC is calculated from
the MOVES runs used to generate VMT from the AEO estimates. These MOVES runs also
generated population estimates using MOVES default growth and scrappage algorithms. A ratio
of the VMT per vehicle by SCC vehicle type is calculated using the VMT values and the
corresponding population estimates. Using this VMT/population ratio, the vehicle population for
each county by SCC can be calculated from the VMT estimates by SCC.
The average speeds provided to SMOKE for each county were derived from the default
national average speed distributions found in the default MOVES2010b database
AvgSpeedDistribution table. These average speeds are the average speeds originally developed
for the previous EPA highway vehicle emission factor model, MOBILE6.48 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.
7.2.1.2 Nonroad Emissions
The "primary" nonroad emissions used in air quality modeling are identical to those used
for national inventories as presented in Section 7.1.4 above. The NMEVI 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.49 Table 7-41 shows that the only effect of the rule
captured by the NONROAD model is a decrease in SO2 due to a drop in fuel sulfur. Not
modeled are a probable decrease in sulfate PM and possible decreases in other pollutants due to
improved catalyst performance in new equipment that may be equipped with catalysts. Both of
these un-modeled effects would be due to that decreases in gasoline fuel sulfur that are part of
this rule.
Table 7-41 National Nonroad Emissions for Calendar Year 2018 and 2030
Pollutant
NOX
VOC
CO
PM2.5
Benzene
Acrolein
1,3-Butadiene
Formaldehyde
Acetaldehyde
SO2
2018
Reference
1,076,370
1,436,324
13,566,942
105,409
26,124
613
3,298
16,225
8,094
2,729
Control
1,076,370
1,436,324
13,566,942
105,409
26,124
613
3,298
16,225
8,094
1,977
Difference
0%
0%
0%
0%
0%
0%
0%
0%
0%
-28%
2030
Reference
729,721
1,225,104
14,935,644
68,308
24,146
541
3,215
13,940
6,902
3,113
Control
729,721
1,225,104
14,935,644
68,308
24,146
541
3,215
13,940
6,902
2,251
Difference
0%
0%
0%
0%
0%
0%
0%
0%
0%
-28%
7-61
-------
7.2.1.3 Heavy-Duty Extended Idle
7.2.1.3.1 Methodology
A non-negligible amount of heavy duty vehicles emissions occur during long idling of
heavy duty diesel trucks engaged in long distance hauling during federally required "down" time.
These "extended idle" emissions are accounted for in our MOVES model. However, the amount
of extended idle that occurs in counties varies significantly, depending on the presence of
interstate highways, freight routes and the presence of truck stops.
We have developed a method to generate an extended idle adjustment using the county
specific extended idle activity and county specific combination truck vehicle populations. These
adjustments are applied in SMOKE to adjust the representing county extended idle emission
rates produced by MOVES (in units of grams per vehicle) to reflect the different amount of
extended idle activity in the represented counties.
RPVc = RPV * (POPn * EIAllocFactor)/POPc
Where:
• RPVc : Rate per vehicle for the county.
• RPV: Rate per vehicle of the representing county.
• POPn: National vehicle population.
• POPc: Vehicle population of the county.
• EIAllocFactor: Emission rate (grams per vehicle).
7.2.1.4 Portable Fuel Container and Upstream Emissions
The Tier 3 rule has no impact on portable fuel container (PFC) emissions. The standards
are also not expected to impact upstream emissions associated with fuel transport/distribution.
For fuel production, the results of our refinery permitting analysis described in Section V.B. of
the preamble and Chapter 4 of the RIA project minor emissions increases at some refineries due
to the reductions in fuel sulfur content that would be required by the Tier 3 standards. We did not
include these emission impacts in our modeling because the projected increases are small and
may be even less than projected if refineries apply emissions controls to reduce emissions
increases.
Although there is no modeled impact of the Tier 3 standards on upstream or PFC
emissions a significant number of modifications were made to the 2007 v.5 platform inventory
reflecting the renewable fuel volumes projected by AEO2013 in the reference case air quality
inventory. These modifications are described in detail in a memorandum to the docket.50
Modifications to point and nonpoint inventories include adjustments to agricultural emissions,
7-62
-------
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 changes in 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.
7.2.1.5 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.L 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. The EPA maintains a database of VOC and particulate matter (PM)
speciation profiles for various emission sources including mobile sources. This database, called
SPECIATE, maintains the record of each profile including its referenced source, testing
methods, a subjective rating of the quality of the data, and other detailed data that allow
researchers to decide which profile is most suitable for model input.M Mobile source
hydrocarbon speciation profiles used in this analysis are from EPA's SPECIATE database
(version 4.4), and additional information on the use of these profiles in air quality modeling, such
as applicable source categories, can be found in the Emissions Inventory TSD.
7.2.2 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)
L For more information, please see the website for SMOKE: http://www.smoke-model.org/index.cfm.
M For more information, please see the website for SPECIATE:
http ://www. epa. gov/ttn/chief/software/speciate/index. html.
7-63
-------
found in the docket for this rule. Results of the air quality modeling are presented in Section
7.2.4.
7.2.2.1 Modeling Methodology
A national-scale air quality modeling analysis was performed to estimate future year 8-
hour ozone concentrations, annual PM2.s concentrations, 24-hour PM2.s concentrations, annual
NO2 concentrations, air toxics concentrations, visibility levels and nitrogen and sulfur deposition
levels for 2018 and 2030. The 2007-based CMAQ modeling platform was used as the basis for
the air quality modeling for this 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 2007. 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.51'52'53 The CMAQ model version 5.0 was most
recently peer-reviewed in September of 2011 for the U.S. EPA.54 The CMAQ model is a well-
known and well-respected tool and has been used in numerous national and international
applications.55'56'57 This 2007 multi-pollutant modeling platform used the most recent multi-
pollutant CMAQ code available at the time of air quality modeling (CMAQ version 5.0.1N).
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 v5.0.1 which reflects updates to version 4.7 to improve the
underlying science. Section 7.2.3 of this RIA discusses the chemical mechanism and SOA
formation.
7.2.2.2 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 a 12 km grid as shown in Figure 7-11. The modeling domain contains 25 vertical
layers with the top of the modeling domain at about 17,600 meters, or 50 millibars (mb) of
atmospheric pressure.
N CMAQ version 5.0.1 was released on October 19, 2011. It is available from the Community Modeling and
Analysis System (CMAS) as well as previous peer-review reports at: http://www.cmascenter.org.
7-64
-------
Figure 7-11 Map of the CMAQ Modeling Domain
7.2.2.3 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 Weather Research and Forecasting Model
Ł - S &
(WRF) version 3.3, Advanced Research WRF (ARW) core for the entire year of 2007 over
model domains that are slightly larger than those shown in Figure 7-11. Previous CMAQ annual
simulations have typically utilized meteorology provided by the 5th Generation Mesoscale
Model (MM5).59 The WRF Model is a next-generation mesoscale numerical weather prediction
system developed for both operational forecasting and atmospheric research applications
(http://wrf-model.org). The meteorology for the national 36 km grid and 12 km grid were
developed by EPA and are described in more detail within the AQM TSD. The meteorological
outputs from WRF were processed to create model-ready inputs for CMAQ using the
Meteorology-Chemistry Interface Processor (MCIP) version 4.1.2. 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.60
The lateral boundary and initial species concentrations are provided by a three-
dimensional global atmospheric chemistry model, the GEOS-CHEM model (standard version 8-
03-02 with version 8-02-03 chemistry).61 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 2007 with
a grid resolution of 2 degree x 2.5 degree (latitude-longitude) and 46 vertical layers up to 0.01
hPa. 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
7-65
-------
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 2007 base year and each of the future year base cases
and control scenarios analyzed for this rule are summarized in Section 7.1.2 of this RIA.
7.2.2.4 CMAQ Evaluation
An operational model performance evaluation for ozone, PIVb.s 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 2007 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 five large subregions:0 Midwest, Northeast, Southeast, Central, and
West U.S. 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.1" Overall,
the performance for the 2007 modeling platform is within the range or close to that of these other
applications. 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 2007 CMAQ model performance evaluation
is available within the AQM TSD found in the docket of this rule.
7.2.2.5 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 NC>2
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:
- 2007 base year
- 2018 Tier 3 reference case
- 2018 Tier 3 control case
- 2030 Tier 3 reference case
0 The subregions are defined by States where: Midwest is IL, IN, MI, OH, and WI; Northeast is CT, DE,
MA, MD, ME, NH, NJ, NY, PA, RI, and VT; Southeast is AL, FL, GA, KY, MS, NC, SC, TN, VA, and
WV; Central is AR, IA, KS, LA, MN, MO, ME, OK, and TX; West is AK, CA, OR, WA, AZ, MM, CO, UT, WY,
SD, ND, MT, ID, and NV.
p 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.
7-66
-------
- 2030 Tier 3 control case
The emission inventories used in the air quality and benefits modeling are different from
the 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 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 7.2.1 of this 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 2007 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 NC>2 concentrations and visibility impairment for each of the
2018 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., 2005-2009).
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.s 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 |ig/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)".62 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.63 The projected 8-hour ozone design
values were calculated using the approach identified in EPA's guidance on air quality modeling
attainment demonstrations.64
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, naphthalene, and acrolein, as well as
7-67
-------
annual nitrate and sulfate deposition. These data were not compared in a relative sense due to
the limited observational data available.
7.2.3 Chemical Mechanisms in Modeling
This rule presents inventories for NOx, VOC, CO, PM2.5, SO2, NHa, and seven air
toxics: benzene, 1,3-butadiene, formaldehyde, acetaldehyde, ethanol, naphthalene and acrolein.
The air toxics were added as explicit model species to the carbon bond 5 (CB05) mechanisms
used in CMAQvS.0.1,65 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.66 Ethanol emissions for air
quality modeling were based on speciation of VOC using different ethanol profiles (EO, E10 and
E85) (see Section 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 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.67
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.02
NO2 + -OH + M^HNO3+M k = 1.19 x 10'11 cm3moleculeV 68
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.
1 All rate coefficients are listed at 298 K and, if applicable, 1 bar of air.
7-68
-------
Atmospheric reactions and chemical mechanisms involving several key formation
pathways are discussed in more detail in the following sections.
7.2.3.1 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 62 to form the acetyl peroxy radical
[CH3C(O)OO].R 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. s
CH3CHO + -OH -> CH3C-O + H2O k = 1.5 x 10'11 cn^molecule'V1 69
CH3C-O + O2 + M -> CH3C(O)OO + M
CH3C(O)OO + NO -> CH3C(O)O- + NO2 k = 2.0 x 10'11 cn^molecule'V1 70
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 cn^molecule'V1 71
Acetaldehyde can react with the NO3 radical, ground state oxygen atom (O3P) and
chlorine, although these reactions are much slower. Acetaldehyde can also photolyze (hv),
which predominantly produces -CH3 (which reacts as shown above to form CH3OO-) and HCO
(which rapidly forms HO2 and CO):
CH3CHO + hv +2 O2 -> CH3OO- +HO2 + CO A, = 240-3 80 nm 72
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.73 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
R 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).
s All rate coefficients are listed at 298 K and, if applicable, 1 bar of air.
7-69
-------
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 + hv + 2 O2 -> 2 HO2 + CO 1 = 240-360 nm 74
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
mechanism75'76 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.77
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 (O (3P)), hydroxyl radical (OH), or nitrate radicals.
The reaction rates are based on expert recommendations,78 and the photolysis rate is from
IUPAC recommendations.
In CMAQ v5.0, 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, CXOs). 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.79 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).
The reaction of ethanol (CH3CH2OH) with OH is slower than some other important
reactions but can be an important source of acetaldehyde if the emissions are large. Based on
kinetic data for molecular reactions, the only important chemical loss process for ethanol (and
other alcohols) is reaction with the hydroxyl radical (-OH).80 This reaction produces
7-70
-------
acetaldehyde (CH3CHO) with a 90 percent yield.81 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
T
order of a day in polluted urban areas or several days in unpolluted areas. For example, an
atmospheric lifetime for acetaldehyde under nominal oxidant conditions, OH of 1.0 x 10"6
cm3molecule"1s"1, would be 3.5 days.
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 -CHS 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
7.2.3.2 S econdary Organi c Aero sol s ( S O A)
Secondary organic aerosol (SOA) chemistry research described below has led to
implementation of new pathways for secondary organic aerosol (SOA) in CMAQ v5.0, based on
89 S'l
recommendations of Edney et al. (2007) and Carlton et al. (2008). ' In previous versions of
CMAQ, all SOA was semivolatile and resulted from the oxidation of compounds emitted entirely
in the gas-phase. In CMAQ v5.0, 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 (OA) 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 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).84'85 A
wide range of laboratory studies conducted over the past twenty years show that anthropogenic
aromatic hydrocarbons and long-chain alkanes, along with biogenic isoprene, monoterpenes, and
sesquiterpenes, contribute to SOA formation.86'87'88'89'90 Modeling studies, as well as carbon
isotope measurements, indicate that a significant fraction of SOA results from the oxidation of
biogenic hydrocarbons.91'92 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.93
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,
T All rate coefficients are listed at 298 K and, if applicable, 1 bar of air.
7-71
-------
currently accounting for almost 40 percent of anthropogenic VOC,94 mobile sources are also an
important source of SO A, particularly in populated areas.
Toluene is an important contributor to anthropogenic SOA.95'96 Mobile sources are the
most significant contributor to ambient toluene concentrations as shown by analyses done for the
2005 National Air Toxics Assessment (NATA)97 and the Mobile Source Air Toxics (MSAT)
QQ
Rule. The 2005 NATA indicates that onroad and nonroad mobile sources accounted for almost
60 percent (1.46 |ig/m3) of the total average nationwide ambient concentration of toluene (2.48
|ig/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.99 Oxidation of alkanes with longer chains as well as cyclic alkanes form SOA
with relatively higher yields than small straight-chain alkanes.100
It is unlikely that ethanol would form SOA directly or affect SOA formation indirectly
through changes in the radical populations due to 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.101 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 could form. As expected, no SOA was produced. From these
experiments, the upper limit for the aerosol yield is less than 0.01 percent based on scanning
mobility particle sizer (SMPS) data. Given the lack of aerosol formation found in these initial
smog chamber experiments, these data were not published.
In general, measurements of OA 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.102'103 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 POA from the
measured OC concentration.104 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
7-72
-------
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 complete compositional data from
the precursors has made the identification of aromatically-derived SOA in ambient samples
challenging, which in turn has prevented observation-based measurements of individual SOA
source contributions to ambient PM levels.
As a first step in estimating ambient SOA concentrations, EPA has developed a tracer-
based method.105'106 The method is based on using mass fractions of SOA tracer compounds,
measured in smog chamber-generated SOA samples, to convert ambient concentrations of SOA
tracer compounds to ambient SOA concentrations. This method consists of irradiating the SOA
precursor of interest in a smog chamber in the presence of NOx, collecting the SOA produced on
filters, and then analyzing the samples for highly polar compounds using advanced analytical
chemistry methods. Employing this method, candidate tracers have been identified for several
VOC compounds which are emitted in significant quantities and known to produce SOA in the
atmosphere. Some of these SOA-forming compounds include toluene, a variety of
monoterpenes, isoprene, and p-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.107'108'109'1 °
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.111 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, />-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 environmental
smog chambers. These unstudied compounds could produce SOA species that are being used as
tracers for other VOCs thus overestimating 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
7-73
-------
biogenic hydrocarbons.112'113 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 been used to estimate
monthly ambient aromatic SOA concentrations from March 2004 to February 2005 in five U.S.
Midwestern cities.114 The annual tracer-based SOA concentration estimates were 0.15, 0.18,
0.13, 0.15, and 0.19 ug 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 ug/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.11
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.116'117'118'119'
120,121,122 studies jiave incjicated that ambient OC levels may be underestimated by current model
parameterizations.123 While the treatment of new precursors has likely reduced the
1 9/1
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.125'126
7.2.3.3 Ozone
As mentioned above, the addition of ethanol to fuels has been shown to contribute to
PAN formation and this is one way for it to contribute therefore to ground-level ozone formation
downwind of NOx sources. PAN is a reservoir and carrier of NOx and is the product of acetyl
radicals reacting with NO2 in the atmosphere. One source of PAN is the photooxidation of
acetaldehyde (Section 7.2.3.1), but many VOCs have the potential for forming acetyl radicals
and therefore PAN or a PAN-type compound.11 PAN can undergo thermal decomposition with a
lifetime of approximately 1 hour at 298K or 148 days at 250K. v
CH3C(O)OONO2 + M -> CH3C(O)OO- + NO2 + M k = 3.3 x 10'4 s'1 127
u 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. Mclver (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.).
v All rate coefficients are listed at 298 K and, if applicable, 1 bar of air.
7-74
-------
The reaction above shows how NC>2 is released in the thermal decomposition of PAN,
along with a peroxy radical which can oxidize NO to NC>2 as previously shown in Section
7.2.3.1. NC>2 can also be formed in photochemical reactions where NO is converted to NO2 (see
OH radical reaction of acetaldehyde in Section 7.2.3.1). In both cases, NO2 further photolyzes to
produce ozone (Oj).
NO2 + hv -> NO + O(3P) 1 = 300-800 nm 128
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.129 A discussion of CB05
mechanisms for ozone formation can be found in Yarwood et al. (2005). 13°
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 7.2.3.1, the photolysis of both aldehydes results in up to two molecules of either
hydroperoxy radical or methylperoxy radical, both of which oxidize NO to NO2 leading to ozone
formation.
7.2.3.4 Uncertainties Associated with Chemical Mechanisms
A key source of uncertainty with respect to the air quality modeling results is the
photochemical mechanisms in CMAQ v5.0. Pollutants such as ozone, PM, acetaldehyde,
formaldehyde, and acrolein 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 7.2.3.
For PM, there are a number of uncertainties associated with SO A formation that should
be addressed explicitly. As mentioned in Section 7.2.3, a large number of VOCs emitted into the
atmosphere, which have the potential to form SO A, have not yet been studied in detail. Not only
have known VOCs not been studied in detail, but unknown (or unmeasured) VOCs can also
produce SOA. This makes reconciling SOA from combustion sources extremely difficult. 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.131
7.2.4 Impacts of the Rule on Air Quality
Air quality modeling performed for this rule estimates the changes in ambient
concentrations of PM2.s, ozone and NO2, as well as changes in ambient concentrations of ethanol
and the following air toxics: acetaldehyde, acrolein, benzene, 1,3-butadiene, naphthalene, and
7-75
-------
formaldehyde. The air quality modeling results also include changes in deposition of nitrogen
and sulfur and changes in visibility levels due to this rule.
This section describes current ambient levels of the modeled pollutants and presents the
projected future ambient levels resulting from the rule.
7.2.4.1 Ozone
As described in Section 6.1.2 of this 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.
7.2.4.1.1 Current Concentrations of Ozone
Figure 7-12 shows a snapshot of measured ozone concentrations in 2010. The highest
ozone concentrations were located in California.
7-76
-------
Concentration Range (ppm)
• 0.025-0.059(81 Sites)
O 0.060 - 0.075 (835 Sites)
O 0.076 - 0.095 (279 Sites)
• 0.096-0.120 (18 Sites)
Puerto Rico
Alaska
Figure 7-12 Ozone Concentrations (fourth highest daily maximum 8-hour concentration) in
ppm for 2010
w
The primary and secondary NAAQS for ozone are 8-hour standards with a level of 0.075
ppm. The most recent revision to the ozone standards was in 2008; the previous 8-hour ozone
standards, set in 1997, had a level of 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)x As of December 5,
2013, there were 39 8-hour ozone nonattainment areas for the 1997 ozone NAAQS, composed of
216 full or partial counties, with a total population of over 112 million. Nonattainment areas for
the 1997 8-hour ozone NAAQS are pictured in Figure 7-13. Nonattainment designations for the
2008 ozone standards were finalized on April 30, 2012 and May 31, 2012.132 As of December 5,
2013, there were 46 ozone nonattainment areas for the 2008 ozone NAAQS, composed of 227
full or partial counties, with a population of over 123 million. Nonattainment areas for the 2008
w 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/.
x 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.
7-77
-------
ozone NAAQS are pictured in Figure 7-14. As of December 5, 2013, over 135 million people
. • • ,. • " Y
are living in ozone nonattainment areas
8-Hour Ozone Nonattainment Areas (1997 Standard)
Nonattainment areas are indicated by color-
When only a portion of a county is shown in color.
it indicates that only that part of the county is within
a nonattainment area boundary.
8hr Ozone Classifications
^| Extreme
I I Severs 17
I | Severe 15
| | Ss'ious
I | Moderate
I | Marginal
The St. Louis. MO-IL 8-hr Ozone (1997 Standard) multi-state nonattainment area has a
state that has been redesignated but it is not considered a maintenance area until all states
in the area are redesignated The counties for this area are displayed as nonattainment areas
The South Carolina portion of the Charlotte-Gastonia-RockHill. NC-SC 8-hr Ozone
(1997 Standard) nonattainment area has been redesignated and the North Carolina
portion will be redesignated effective January 2. 2014 The entire area is not considered
in maintenance until all states in a multi-state area are redesignated
Figure 7-13 1997 8-hour Ozone Nonattainment Areas
Y The 135 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.
7-78
-------
8-Hour Ozone Nonattainment Areas (2008 Standard)
Nonattainment areas are indicated by color
When only a portion of a county is shown in color
it indicates that only that part of the county is within
a nonattainment area boundary
8-hour Ozone Claasification
| Extreme
I I Severe 15
| | Serious
I | Moderate
I I Marginal
Figure 7-14 2008 8-hour Ozone Nonattainment Areas
States with ozone nonattainment areas are required to take action to bring those areas into
attainment. The attainment date assigned to an ozone nonattainment area is based on the area's
classification. Most ozone nonattainment areas were 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 currently
working on a review of the ozone NAAQS. If EPA revises the ozone standards pursuant to that
review, the attainment dates associated with areas designated nonattainment for that NAAQS
would be 5 or more years after the final rule is promulgated, depending on the severity of the
problem in each area.
7.2.4.1.2 Projected Concentrations of Ozone Without the Rule
EPA has already adopted many mobile source emission control programs that are
expected to reduce ambient ozone levels. These control programs include the Heavy-Duty
z 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.
7-79
-------
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 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 2018, with all current controls in effect but
excluding the emissions changes expected to occur as a result of this action or any other
additional controls, at least 19 counties, with a projected population of over 37 million people,
will 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 6 counties with a population of over 19 million
people that will 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 rule 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 standards.
7.2.4.1.1 Projected Concentrations of Ozone With the Rule
This section summarizes the results of our modeling of ozone air quality impacts in the
future with the standards. Specifically, for the years 2018 and 2030 we compare a reference
scenario (a scenario without the standards) to a control scenario that includes the standards. Our
modeling indicates that ozone design value concentrations will decrease dramatically in many
areas of the country as a result of this rule rule. Additional information on the emissions
reductions that are projected with this final action is available in Section 7.2.1 of this RIA.
Figure 7-15 and Figure 7-16 present the changes in 8-hour ozone design value
concentrations in 2018 and 2030 respectively.
7-80
-------
Leqend Number of Counties
^H<-1.0ppb 33
•• 303
^B >= -° 5 to * -° 25 228
| | >=-0.2510 <-0.1 38
I >=-0,1 tO<=0.1 56
| > 0.1 to <= 0.25 0
] > 0,25 to <= 0.5 0
^H > 0.5 to <= 1 0 o
^H > 1.0 o
Difference in 8-hr Ozone DV- 201Big_ctI minus 2015rg_ref
Figure 7-15 Projected Change in 2018 8-hour Ozone Design Values Between the Reference
Case and Control Case
7-81
-------
Difference in 8-hr Ozone DV- 2030rg_ctf minus 2Q3Qrg_ref
Figure 7-16 Projected Change in 2030 8-hour Ozone Design Values Between the Reference
Case and Control Case
As can be seen in Figure 7-15, the majority of the design value decreases in 2018 are
between 0.5 and 1.0 ppb. There are also 33 counties with projected 8-hour ozone design value
decreases of more than 1 ppb; these counties are generally in urban areas in states that have not
adopted California LEV III standards. The maximum projected decrease in an 8-hour ozone
design value in 2018 is 1.56 ppb in Henry County, Georgia near Atlanta. Figure 7-16 presents
the ozone design value changes for 2030. In 2030 the ozone design value decreases are larger
than in 2018; most decreases are projected to be between 0.5 and 1.0 ppb, but over 250 more
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 2.8 ppb in Gwinnett County,
Georgia, the northeastern part of the Atlanta metropolitan area.
Table 7-42 and Table 7-43 show the average change, due to this rule, in 2018 and 2030 8-
hour ozone design values for: (1) all counties with 2007 baseline design values, (2) counties with
2007 baseline design values that exceeded the 2008 ozone standard, (3) counties with 2007
baseline design values that did not exceed the 2008 standard, but were within 10 percent of it, (4)
counties with 2018/2030 design values that exceeded the 2008 ozone standard, and (5) counties
with 2018/2030 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 2018 and 2030,
indicating in five different ways the overall improvement in air quality.
7-82
-------
On a population-weighted basis, the average modeled future-year 8-hour ozone design
values are projected to decrease by 0.49 ppb in 2018 and 0.98 ppb in 2030.
Table 7-42 Average Change in Projected 8-hour Ozone Design Value in 2018
Average"
All
All, population-weighted
Counties whose 2007 base year is violating the 2008
8 -hour ozone standard
Counties whose 2007 base year is violating the 2008
8 -hour ozone standard, population-weighted
Counties whose 2007 base year is within 10 percent
of the 2008 8-hour ozone standard
Counties whose 2007 base year is within 10 percent
of the 2008 8-hour ozone standard, population-
weighted
Counties whose 2018 control case is violating the
2008 8-hour ozone standard
Counties whose 2018 control case is violating the
2008 8-hour ozone standard, population-weighted
Counties whose 2018 control case is within 10
percent of the 2008 8-hour ozone standard
Counties whose 2018 control case is within 10
percent of the 2008 8-hour ozone standard,
population-weighted
Number
ofU.S.
Counties
658
319
241
16
97
2020
Population
234,598,095
159,636,546
52,115,093
35,732,987
55,924,864
Change in
20 18 design
value (ppb)
-0.53
-0.49
-0.60
-0.51
-0.51
-0.50
-0.24
-0.20
-0.50
-0.55
a Averages are over counties with 2007 modeled design values.
b Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2011). 2012 Complete
Economic and Demographic Data Source (CEDDS)..
Table 7-43 Average Change in Projected 8-hour Ozone Design Value in 2030
Average"
All
All, population-weighted
Counties whose 2007 base year is violating the 2008
8 -hour ozone standard
Counties whose 2007 base year is violating the 2008
8 -hour ozone standard, population-weighted
Counties whose 2007 base year is within 10 percent
of the 2008 8-hour ozone standard
Number
ofU.S.
Counties
658
319
241
2030
Population
257,693,543
175,088,003
57,122,153
Change in
2030 design
value (ppb)
-0.94
-0.98
-1.06
-1.01
-0.90
7-83
-------
Counties whose 2007 base year is within 10 percent
of the 2008 8-hour ozone standard, population-
weighted
Counties whose 2030 control case is violating the
2008 8-hour ozone standard
Counties whose 2030 control case is violating the
2008 8-hour ozone standard, population-weighted
Counties whose 2030 control case is within 10
percent of the 2008 8-hour ozone standard
Counties whose 2030 control case is within 10
percent of the 2008 8-hour ozone standard,
population-weighted
6
32
19,415,241
33,059,752
-0.97
-0.16
-0.12
-0.86
-0.95
a Averages are over counties with 2007 modeled design values
b Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2011). 2012 Complete
Economic and Demographic Data Source (CEDDS).
There are still six counties, five of them in California, that are projected to have 8-hour
ozone design values above the 2008 NAAQS in 2030 without the standards or any other
additional controls in place. Table 7-44 below presents the changes in design values for these
counties.
Table 7-44 Change in Ozone Design Values (ppb) for Counties Projected to be Above the
2008 Ozone NAAQS in 2030
County Name
San Bernardino, California
Los Angeles, California
Riverside, California
Kern, California
Westchester, New York
Fresno, California
Change in 8-hour Ozone
Design Value (ppb)
-0.10
-0.10
-0.10
-0.02
-0.64
-0.01
Population in
2030a
2,784,490
10,742,722
2,614,198
981,806
1,196,950
1,095,075
a Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2011). 2012
Complete Economic and Demographic Data Source (CEDDS).
In terms of modeling accuracy, the count of modeled nonattainment counties is much less
certain than the average changes in air quality. Bearing this in mind, our modeling predicts that
the Tier 3 standards will reduce ozone design values in some counties from above the level of the
standard to below it. In 2018, ozone design values in three counties (Harford County in
Maryland, Denton County in Texas and Fairfield County in Connecticut) are projected to move
from being above the standard to below. The projected population in these three counties in
2018 is almost 2 million people.
As described in Section 6.1.2.1 of this 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 fuel and vehicle standards. This change in ozone results from
interactions between photochemistry, background concentrations of ozone, VOC and NOx, local
7-84
-------
emissions and meteorology. There is one county in 2018 that is projected to have an increase in
modeled ozone design value concentration (Cuyahoga County, OH, where Cleveland is located).
When NOx levels are relatively high and VOC levels relatively low, NOx forms inorganic
nitrates (i.e., particles) but relatively little ozone. In addition, NOx can react directly with ozone
resulting in suppressed ozone concentrations near NOx emissions sources. 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 Cuyahoga County in 2018. 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.
7.2.4.2 Particulate Matter
As described in Section 6.1.1 of this 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.
7.2.4.2.1 Current Concentrations ofPM
Figure 7-17 and Figure 7-18 respectively show a snapshot of annual and 24-hour PM2.s
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.
7-85
-------
Annual
Concentration Range (pg/m3)
• 3.1-12.0 (680 Sites)
O 12.1-15.0 (148 Sites)
O 15.1-18.0 (5 Sites)
• 18.1 -225(1 Site)
Puerto Rico
Alaska
Figure 7-17 Annual Average PM2.s Concentrations in ug/m3 for 2010AA
^ 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/.
7-86
-------
24-hour
Concentration Range (\ig/m3)
• 6-15 (87 Sites)
O 16-35 (704 Sites)
O 36 - 55 (42 Sites)
• 55-56(1 Site)
th
Puerto Rico
Alaska
Figure 7-18 24-hour (98 percentile 24- hour concentrations) PMi.s Concentrations in
ug/m3 for 2010BB
There are two primary NAAQS for
an annual standard (12.0 ug/m3) and a 24-hour
standard (35 ug/m), and two secondary NAAQS for
an annual standard (15.0 ug/m) and
a 24-hour standard (35 ug/m). The initial PM2.5 standards were set in 1997 and revisions to the
standards were finalized in 2006 and in December 2012. The December 2012 rule revised the
level of the primary annual PM2.5 standard from 15.0 ug/m3 to 12.0 ug/m3.133
In 2005 the EPA designated 39 nonattainment areas for the 1997 PM2.5 NAAQS (70 FR
19844, April 14, 2005). As of December 5, 2013, over 68 million people lived in the 24 areas
that are still designated as nonattainment for the 1997 annual PM2.5 NAAQS. These PM2.5
nonattainment areas are comprised of 135 full or partial counties. Nonattainment areas for the
1997 annual PM2.5 NAAQS are pictured in Figure 7-19. EPA anticipates making initial area
designation decisions for the 2012 primary annual PM2.5 NAAQS in December 2014, with those
designations likely becoming effective in early 2015.134 On November 13, 2009 and February 3,
201 1, the EPA designated 32 nonattainment areas for the 2006 24-hour PM2.5 NAAQS (74 FR
58688, November 13, 2009 and 76 FR 6056, February 3, 201 1). As of December 5, 2013, 28 of
these areas remain designated as nonattainment, and they are composed of 104 full or partial
counties, with a population of over 65 million. Nonattainment areas for the 2006 PM2.5 NAAQS
BB 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/.
7-87
-------
are pictured in Figure 7 20. In total, there are currently 39 PM2.s nonattainment areas with a
population of over 84 million people.cc
States with PM2.5 nonattainment areas will be required to take action to bring those areas
into attainment in the future. Designated nonattainment areas not currently attaining the 1997
annual PM2.s NAAQS are required to attain the NAAQS by 2015 and will be required to
maintain the 1997 annual PM2.5 NAAQS thereafter. The 2006 24-hour PM2.5 nonattainment
areas are required to attain the 2006 24-hour PM2.s NAAQS in the 2015 to 2019 time frame and
will be required to maintain the 2006 24-hour PM2.5 NAAQS thereafter. Areas to be designated
nonattainment for the 2012 primary annual PM2.5 NAAQS will likely be required to attain the
2012 NAAQS in the 2021 to 2025 time frame.
cc Data come from Summary Nonattainment Area Population Exposure Report, current as of December 5, 2013 at:
http://www.epa.gov/oar/oaqps/greenbk/popexp.html and contained in Docket EPA-HQ-OAR-2011-0135. The 84
million total is calculated by summing, without double counting, the 1997 and 2006 PM25 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
2006 nonattainment areas, and they are not the same, then the larger of the two populations is included in the sum.
7-88
-------
PM-2.5 Nonattainment Areas (1997 Standard)
Nonattainment areas are indicated by color
When only a portion of a county is shown in color.
it indicates that only that part of the county is within
a nonattainment area boundary.
12 3? 2312
The New Jersey portion of the Philadelphia-Wilmington. PA-NJ-DE PM-2.5 nonattainment area (1997
Standard) has been redesignated. while the Pennsylvania and Delaware protions have not. The entire area
is not considered in maintenance until all states in a multi-state area are redesignated.
The New Jersey and Connecticut portions of the New York-N. New Jersey-Long Island. NY-NJ-CT
PM-2.5 nonattainment area (1997 Standard) have been redesignated. while the New York portion has not
The entire area is not considered in maintenance until all states in a multi-state area are redesignated
The Ohio portion of the Steubenville-Weirton. OH-WV PM-2.5 (1997 Standard) nonattainment area has been
redesignated. while the West Virginia portion has not The entire area is not considered in maintenance until
all states in a multi-state area are redesignated
Figure 7-19 1997 PM2.5 Nonattainment Areas
7-89
-------
PM-2.5 Nonattainment Areas (2006 Standard)
Nonattainment areas are indicated by color
When only a portion of a county is shown in color.
It indicates that only that part of the county is within
a nonattainment area boundary.
The New Jersey portion of the Philadelphia-Wilmington PA-NJ-DE PM-2.5 (2006 Standard)
nonattainment area has been redesignated. while the Pennsylvania and Delaware protions have
not The entire area is not considered in maintenance until all states in a multi-state area
are redesignated
The New Jersey and Connecticut portions of the New York-N. New Jersey-Long Island. NY-NJ-CT
PM-2.5 (2006 Standard) nonattainment area have been redesignated. while the New York portion has not
The entire area is not considered in maintenance until all states in a multi-state area are redesignated.
The Ohio portion of the Steubenville-Weirton. OH-WVPM-25 (2006 Standard) nonattainment area has been
redesignated. while the West Virginia portion has notThe entire area is not considered in maintenance
until all states in a multi-state area are redesignated.
Figure 7-20 2006 PM2.s Nonattainment Areas
As of December 5, 2013, over 11 million people live in the 40 areas that are designated
as nonattainment for the PMio NAAQS. There are 33 full or partial counties that make up the
PMio nonattainment areas. Nonattainment areas for the PMio NAAQS are pictured in Figure
7-21.
7-90
-------
Counties Designated Nonattainment for PM-10
New Yort Co., NY
{Moderate)
1:35:013
• Serious
I I Moderate
Classification colors are shown for whole counties and
denote the highest area classification that the county is in
Figure 7-21 PMio Nonattainment Areas
7.2.4.2.2 Projected Concentrations ofPM2.s Without the 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.s 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
7-91
-------
already stringent local regulations from some of the burden associated with adopting additional
local controls.
The air quality modeling conducted for this rule projects that in 2030, with all current
controls in effect but excluding the emissions changes expected to occur as a result of this rule or
any other additional controls, at least 13 counties, with a projected population of over 21 million
people, will have projected design values above the level of the annual standard of 12.0 |ig/m3
and at least 18 counties, with a projected population of over 12 million people, will have
projected design values above the level of the 2006 24-hour standard of 35 |ig/m3. Since the
emission changes from this action would go into effect during the period when some areas are
still working to attain the PIVb.s 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 standards.
7.2.4.2.3 Projected Annual Average Concentrations ofPM2.s With the Rule
This section summarizes the results of our modeling of annual average PM2.5 air quality
impacts in the future due to the standards in this action. Specifically, for the years 2018 and
2030 we compare a reference scenario (a scenario without the standards) to a control scenario
that includes the standards. Our modeling indicates that annual PM2.5 design values will
decrease due to the Tier 3 standards. The decreases in annual PIVb.s design values are likely due
to the projected reductions in primary PM2.5, NOx, SOx and VOC emissions. As described in
Section 7.2.1.1, the air quality modeling used inventories that included an increase in direct
PM2.5 emissions in the West and Pacific Northwest that is an artifact of a difference in fuel
properties that isn't real.DD Although in most areas this direct PM2.5 increase is outweighed by
reductions in secondary PM2.5, the air quality modeling does predict ambient PM2.5 increases in a
few places in the West and Pacific Northwest. These modeled increases are a result of the
inventory issue, and we do not expect them to actually occur. Ambient PM2.5 projections are
discussed in more detail below. Additional information on the emissions reductions that are
projected with this action is available in Section 7.2.1 of this RIA.
Figure 7-22 and Figure 7-23 present the changes in annual PM2.5 design values in 2018
and 2030 respectively.EE
DD The issue is with the way that some of the fuel property data, specifically E200/E300 and T50/T90, matched up
in the fuel compliance database in the West and Pacific Northwest, see Section 7.2.1.1 for additional information.
EE 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.
7-92
-------
Legend Number o( Counties
^H •-- -0 1 ug/m3 0
^B >= •:i h- f -°-05 °
^H >= -0.05 to < -0 01 209
[ | >= -0.01 to <= 0.01 364
I | > 0,01 to •== Q.Q5 0
^H 90.05 to <= 0.1 0
Difference it) Annual PM2.S DV~ 2013rg_ctl minus 201Srg_ref2
Figure 7-22 Projected Change in 2018 Annual PMi.s Design Values Between the Reference
Case and Control Case
7-93
-------
Legend
I >= -0.1 to < -0.05 139
| >= -0.05 to < -0.01 326
] >=-0.01 to <=0.01 100
] > 0.01 to « 0.05 0
I > 0.05 lo <= 0,1 0
> 0.1
0/ff«r«nce In Annual PM2.S DV- 2030rg_al minus 2
-------
counties with 2018/2030 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 decrease in 2018 and 2030. On a population-weighted basis, there is a 0.01
|ig/m3 reduction in the average modeled future-year annual PIVb.s design values in 2018 and a
0.04 |ig/m3 decrease in 2030.
Table 7-45 Average Change in 2018 Annual PMi.s Design Value as a Result of the Rule
AVERAGE
All
All, population-weighted
Counties whose 2007 base year is
violating the annual PM2.5 standard
Counties whose 2007 base year is
violating the annual PM2.5 standard,
popul ati on- weighted
Counties whose 2007 base year is
within 10 percent of the annual PM2.5
standard
Counties whose 2007 base year is
within 10 percent of the annual PM2.5
standard, population-weighted
Counties whose 2018 control case is
violating the annual PM2.5 standard
Counties whose 2018 control case is
violating the annual PM2.5 standard,
popul ati on- weighted
Counties whose 2018 control case is
within 10 percent of the annual PM2.5
standard
Counties whose 2018 control case is
within 10 percent of the annual PM2.5
standard, population-weighted
Number of U.S.
Counties
573
231
123
13
28
2020
Population
230,583,259
111,371,097
45,480,691
19,690,375
29,073,863
Change in 20 18
design value
(Hg/m3)
-0.01
-0.01
-0.02
-0.02
-0.01
-0.01
-0.01
0.00
-0.02
-0.02
a Averages are over counties with 2007 modeled design values
b Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2011). 2012 Complete
Economic and Demographic Data Source (CEDDS).
7-95
-------
Table 7-46 Average Change in 2030 Annual PMi.s Design Value as a Result of the Rule
AVERAGE
All
All, population-weighted
Counties whose 2007 base year is
violating the annual PIVb.s standard
Counties whose 2007 base year is
violating the annual PIVb.s standard,
popul ati on- weighted
Counties whose 2007 base year is
within 10 percent of the annual PM2.5
standard
Counties whose 2007 base year is
within 10 percent of the annual PM2.5
standard, population-weighted
Counties whose 2030 control case is
violating the annual PM2.5 standard
Counties whose 2030 control case is
violating the annual PM2.5 standard,
popul ati on- weighted
Counties whose 2030 control case is
within 10 percent of the annual PM2.5
standard
Counties whose 2030 control case is
within 10 percent of the annual PM2.5
standard, population-weighted
Number of U.S.
Counties
573
231
123
13
20
2030
Population
252,063,937
119,932,361
50,009,709
21,376,437
22,244,541
Change in 2030
design value
(Hg/m3)
-0.04
-0.04
-0.05
-0.05
-0.04
-0.05
-0.01
-0.01
-0.07
-0.09
a Averages are over counties with 2007 modeled design values
b Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2011). 2012 Complete
Economic and Demographic Data Source (CEDDS).
There are 13 counties, mostly in California, that are projected to have annual PM2.5
design values above the NAAQS in 2030 without the standards or any other additional standards
in place. Table 7-47 below presents the changes in design values for these counties.
7-96
-------
Table 7-47 Change in Annual PMi.s Design Values (jig/m ) for Counties Projected to be
Above the Annual PM2.5 NAAQS in 2030
County Name
Kern County, California
Imperial County, California
Tulare County, California
Riverside County, California
Fresno County, California
San Bernardino County, California
Santa Cruz County, Arizona
Kings County, California
Lincoln County, Montana
El Paso County, Texas
Merced County, California
Stanislaus County, California
Los Angeles County, California
Change in
Annual
PM2.5
Design
Value
(|ig/m3)
-0.01
-0.01
-0.01
-0.01
0
-0.01
-0.03
0
-0.02
-0.06
0
0
0
Population
in 2030a
981,806
174,175
528,662
2,614,198
1,196,949
2,784,489
55,393
195,067
20,454
1,080,944
313,333
688,245
10,742,722
a Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2011).
2012 Complete Economic and Demographic Data Source (CEDDS).
In terms of modeling accuracy, the count of modeled nonattainment counties is much less
certain than the average changes in air quality. Bearing this in mind, in 2018 our modeling
predicts that the Tier 3 standards will reduce annual PM2 5 design values in Allegheny County,
PA from above the level of the standard to below it. The projected population in Allegheny
County in 2018 is over one million people.
7.2.4.2.4 Projected 24-hour Average Concentrations ofPM2.s With the Rule
This section summarizes the results of our modeling of 24-hour PM25 air quality impacts
in the future due to the rule. Specifically, for the years 2018 and 2030 we compare a reference
scenario (a scenario without the standards) to a control scenario that includes the standards. Our
modeling indicates that 24-hour PM2 5 design values will decrease due to the Tier 3 standards.
The decreases in 24-hour PM2.5 design values are likely due to the projected reductions in
primary PM2.s, NOx, SOx and VOCs. As described in Section 7.2.1.1, the air quality modeling
used inventories that include an increase in direct PM2 5 emissions in the West and Pacific
Northwest that is an artifact of a difference in fuel properties that isn't real.FF Although in most
areas this direct PM2.5 increase is outweighed by reductions in secondary PM2.5, the air quality
FF The issue is with the way that some of the fuel property data, specifically E200/E300 and T50/T90, matched up in
the fuel compliance database in the West and Pacific Northwest, see Section 7.2.1.1 for additional information.
7-97
-------
modeling does predict ambient PM2.s increases in a few places in the West and Pacific
Northwest. These modeled increases are a result of the inventory issue, and we do not expect
them to actually occur. Ambient PM2.5 projections are discussed in more detail below.
Additional information on the emissions reductions that are projected with this action is available
in Section 7.2.1 of this RIA.
Figure 7-24 and Figure 7-25 present the changes in 24-hour PM2.5 design values in 2018
and 2030 respectively
Legend Number cl Counties
Hi <= -0.5 ug/m3 0
^H > .0.5 10 <= -0 25 1
^•:--0.25tn t=-0 IS 15
t >.Q.15lo<=.0.05 207
I | > -0.05 10 < 0 05 342
[ >=0.05to<015 3
|| >= 0.15 to < 0,25 0
HH >= 0.25 to < 0.5 0
^H >= 0 5 0
Difference in Daily PM2.5 DV-- 2018rg_ctl minus 201Brg_ret2
Figure 7-24 Projected Change in 2018 24-hour PMi.s Design Values Between the Reference
Case and the Control Case
7-98
-------
Legend
Ntitnlmi ii' • ' i
j^H <= -0.5 ug/m3
•• >-0.5 to <=-0.25
^H > -0.25 to <= -0.15
| '" | > -0.15 to <= -0.05
1 >-0.05 to < 0.05
] >= 0.05 to < 0.15
H >= 0.15 to < 0.25
B| -= 025to< 0.5
•• - 0.5
Difference in Daily PM2.5 DV - 2030rg_cU minus 2030rg_ref
Figure 7-25 Projected Change in 2030 24-hour PMi.s Design Values Between the Reference
Case and the Control Case
As shown in Figure 7-24, in 2018 there are 16 counties with projected 24-hour PM2.5
design value decreases greater than 0.15 |ig/m3. These counties are in urban areas in states that
have not adopted California LEV III standards. The maximum projected decrease in a 2018 24-
hour PM2.5 design value is 0.30 |ig/m3 in Utah County, Utah. There are three counties with
projected increases in their 24-hour PM2.5 design values in 2018: Washington County, Oregon;
King County, Washington; and Sheridan County, Wyoming. These projected increases are a
result of the issue with the air quality modeling emissions inventories discussed in Section
7.2.1.1, and we do not expect these increases will occur.
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 |ig/m3 and over 50 counties have projected design value decreases greater
than 0.25 |ig/m3. The maximum projected decrease in a 24-hour PM2.5 design value in 2030 is
0.8 |ig/m3 in Salt Lake County, Utah. As shown in Figure 7-25, design values in 9 counties are
projected to decrease by more than 0.5 |ig/m3. These counties are in Utah, Idaho, Colorado and
Wisconsin. There are two counties with projected increases in their 24-hour PM2.5 design values
in 2030: King County, Washington, and Pierce County, Washington. These projected increases
are a result of the issue with the air quality modeling emissions inventories discussed in Section
7-99
-------
7.2.1.1 and we do not expect these increases will occur. Additional information on the emissions
reductions that are projected with this action is available in Section 7.2.1.
Table 7-48 and Table 7-49 show the average change in 2018/2030 24-hour PM2.5 design
values for: (1) all counties with 2007 baseline design values, (2) counties with 2007 baseline
design values that exceeded the 24-hour PM2.5 standard, (3) counties with 2007 baseline design
values that did not exceed the standard, but were within 10 percent of it, (4) counties with
2018/2030 design values that exceeded the 24-hour PM2.5 standard, and (5) counties with
2018/2030 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 0.13 |ig/m3 in
2030 due to the Tier 3 standards.
Table 7-48 Average Change in Projected 24-hour PMi.s Design Values in 2018
Average"
All
All, population-weighted
Counties whose 2007 base year is violating the 2006
24-hour PM2 5 standard
Counties whose 2007 base year is violating the 2006
24-hour PM2 5 standard, population-weighted
Counties whose 2007 base year is within 10 percent
of the 2006 24-hour PM2 5 standard
Counties whose 2007 base year is within 10 percent
of the 2006 24-hour PM2 5 standard, population-
weighted
Counties whose 2030 control case is violating the
2006 24-hour PM2 5 standard
Counties whose 2030 control case is violating the
2006 24-hour PM2 5 standard, population-weighted
Counties whose 2030 control case is within 10
percent of the 2006 24-hour PM2 5 standard
Counties whose 2030 control case is within 10
percent of the 2006 24-hour PM2 5 standard,
population-weighted
Number
ofU.S.
Counties
568
55
84
24
14
2030
Population13
229,741,229
56,171,551
39,543,851
18,382,427
16,167,178
Change in
2030 design
value
(Mg/m3)
-0.04
-0.05
-0.05
-0.06
-0.06
-0.06
-0.04
-0.03
-0.07
-0.09
Note:
a Averages are over counties with 2007 modeled design values
b Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2011). 2012 Complete
Economic and Demographic Data Source (CEDDS).
7-100
-------
Table 7-49 Average Change in Projected 24-hour PMi.s Design Values in 2030
Average"
All
All, population-weighted
Counties whose 2007 base year is violating the 2006
24-hour PM2 5 standard
Counties whose 2007 base year is violating the 2006
24-hour PM2 5 standard, population-weighted
Counties whose 2007 base year is within 10 percent
of the 2006 24-hour PM2 5 standard
Counties whose 2007 base year is within 10 percent
of the 2006 24-hour PM2 5 standard, population-
weighted
Counties whose 2030 control case is violating the
2006 24-hour PM2 5 standard
Counties whose 2030 control case is violating the
2006 24-hour PM2 5 standard, population-weighted
Counties whose 2030 control case is within 10
percent of the 2006 24-hour PM2 5 standard
Counties whose 2030 control case is within 10
percent of the 2006 24-hour PM2 5 standard,
population-weighted
Number
ofU.S.
Counties
568
55
84
18
13
2030
Population13
251,240,080
60,280,914
42,498,707
12,363,252
22,759,997
Change in
2030 design
value
(Hg/m3)
-0.13
-0.13
-0.17
-0.13
-0.18
-0.16
-0.11
-0.14
-0.05
0.00
a Averages are over counties with 2007 modeled design values
b Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2011). 2012 Complete
Economic and Demographic Data Source (CEDDS).
There are 18 counties that are projected to have 24-hour PM2.5 design values above the
NAAQS in 2030 without the Tier 3 standards or any other additional controls in place. Table
7-50 below presents the changes in design values for these counties.
7-101
-------
Table 7-50 Change in 24-hour PMi.s Design Values (jig/m3) for Counties Projected to be
Above the 24-hour PM2.5 NAAQS in 2030
County Name
Sacramento County, California
Butte County, California
Klamath County, Oregon
Imperial County, California
Pierce County, Washington
Stanislaus County, California
Fresno County, California
Merced County, California
Kings County, California
Kern County, California
Lake County, Oregon
Salt Lake County, Utah
Final County, Arizona
Lane County, Oregon
Allegheny County, Pennsylvania
San Joaquin County, California
Utah County, Utah
Tulare County, California
Change in 24-hour
PM2 5 Design Value
((ig/m3)
0
0
-0.1
-0.1
0.1
0
-0.1
0
0
0
0
-0.8
-0.2
-0.1
0
-0.1
-0.6
0
Population in 203 Oa
1,856,970
287,235
77,199
174,175
1,082,578
688,245
1,196,949
313,333
195,067
981,806
9,462
1,431,946
348,831
460,992
1,234,930
833,417
661,455
528,662
a Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2011). 2012
Complete Economic and Demographic Data Source (CEDDS).
7.2.4.3 Nitrogen Di oxi de
As described in Section 6.1.3 of this 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.
7.2.4.3.1 Current Concentrations of NO 2
The EPA most recently completed a review of the primary NAAQS for NO2 in January
2010. There are two primary NAAQS for NO2: an annual standard (53 ppb) and a 1-hour
standard (100 ppb). The EPA promulgated area designations in the Federal Register on February
17, 2012. In this initial round of designations, all areas of the country were designated as
"unclassifiable/attainment" for the 2010 NO2 NAAQS based on data from the existing air quality
monitoring network. The EPA and state agencies are working to establish an expanded network
of NO2 monitors, expected to be deployed in the 2014-2017 time frame. Once three years of air
quality data have been collected from the expanded network, the EPA will be able to evaluate
NO2 air quality in additional locations.135'136
7-102
-------
7.2.4.3.2 Projected Concentrations of NO 2 without the Rule
EPA has already adopted many mobile source emission control programs that are
expected to reduce ambient NC>2 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 NC>2 in the future are expected to decrease.
7.2.4.3.3 Projected Concentrations of NO 2 with the Rule
This section summarizes the results of our modeling of annual average NC>2 air quality
impacts in the future due to the final Tier 3 standards. Specifically, for the years 2018 and 2030
we compare a reference scenario (a scenario without the standards) to a control scenario that
includes the standards. Figure 7-26 and Figure 7-27 present the changes in annual NC>2
concentrations in 2018 and 2030 respectively.
Difference in Annual Total NO2 Concentration
2018rg_ctl minus 2018rg_ref2
Figure 7-26 Projected Change in 2018 Annual NOi Concentrations Between the Reference
Case and Control Case
7-103
-------
Difference in Annual Total NO2 Concentration
2030rg ctl minus 2030rg ref
Figure 7-27 Projected Change in 2030 Annual NOi Concentrations Between the Reference
Case and Control Case
As shown in Figure 7-27, our modeling indicates that by 2030 annual NC>2 concentrations
in the majority of the country would decrease less than 0.1 ppb due to this rule. However,
decreases in annual NC>2 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.
7.2.4.4 Air Toxics
As described in Section 6.1.5 of this RIA, air toxics cause adverse health effects. In this
section we present information on current and model-projected future levels of air toxics.
7.2.4.4.1 Current Concentrations 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.137 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
I TO
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 2011.139 NATA for 2005 includes four steps:
7-104
-------
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
Figure 7-28 and 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.GG'HH'140 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.
GG 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.
HH 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.
7-105
-------
2005 NATA Estimated Tract Level Total Cancer Risk
Figure 7-28 Tract Level Average Carcinogenic Risk, 2005 NATA
2005 NATA Estimated Tract Level Total Respiratory Hazard Index
Total Respiratory
Hazard Index
0-1
1 -5
^H 5-10
^H 10- 15
^H 15-20
Zero Population Tracts
Figure 7-29 County Level Average Noncancer Hazard Index, 2005 NATA
7-106
-------
7.2.4.4.2 Projected Concentrations of Air Toxics
Since 2005, and in future years when this rulemaking takes effect, the contribution of
mobile sources to overall cancer risk is likely to change as a result of the impact of controls on
both stationary and mobile sources. Overall, EPA anticipates a substantial decrease in overall
cancer risk from ambient sources. In the following sections, we describe results of our modeling
of air toxics concentrations in the future with the Tier 3 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 2018, 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 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), 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 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 2018 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 2018. However, our 2018 modeling projects there would be small immediate reductions
in ambient concentrations of air toxics due to the sulfur controls that take effect 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.
7-107
-------
Benzene
Our modeling projects that the standards would have a notable impact on ambient
benzene concentrations. In 2018, soon after the Tier 3 standards take effect, ambient benzene
reductions are generally between 0.001 and 0.01 |ig/m3, or between 1 and 2.5 percent in some
areas (Figure 7-30). In 2030, our modeling projects that the rule will 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 (Figure 7-31). Absolute decreases in ambient
concentrations of benzene are generally between 0.001 and 0.01 |ig/m3 in rural areas and as
much as 0.1 |ig/m3 in urban areas (Figure 7-31).
7-108
-------
Figure 7-30 Changes in Benzene Ambient Concentrations Between the Reference Case and the
Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
•.-.:; ,-.-.,-if. .1 .nnl ir.- -.,1 . 1|-.i- i L,-.,|. .ilr...
Figure 7-31 Changes in Benzene Ambient Concentrations Between the Reference Case and the
Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
1,3-Butadiene
Our modeling also shows reductions of ambient 1,3-butadiene concentrations in 2018 and
2030. Figure 7-32 shows that in 2018, ambient concentrations of 1,3-butadiene generally
decrease between 1 and 5 percent across parts of the country, corresponding to small decreases in
absolute concentrations (less than 0.001 ug/m3). In 2030, reductions of 1,3-butadiene
7-109
-------
concentrations range between 1 and 25 percent, with decreases of at least 0.005 ug/m3 in urban
areas (Figure 7-33).
Figure 7-32 Changes in 1,3-Butadiene Ambient Concentrations Between the Reference Case
and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7-33 Changes in 1,3-Butadiene Ambient Concentrations Between the Reference Case
and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7-110
-------
Acrolein
Our modeling indicates the standards would reduce ambient concentrations of acrolein in
2018 and 2030. Figure 7-34 shows decreases in ambient concentrations of acrolein generally
between 1 and 2.5 percent in parts of the country in 2018, corresponding to small decreases in
absolute concentrations (less than 0.001 ug/m3). Reductions of acrolein concentrations in 2030
range between 1 and 10 percent, with decreases as high as 0.003 ug/m3 in a few urban areas.
Figure 7-34 Changes in Acrolein Ambient Concentrations Between the Reference Case and
the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7-35 Changes in Acrolein Ambient Concentrations Between the Reference Case and
the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7-111
-------
Ethanol
Our modeling projects that the standards would slightly decrease ambient ethanol
concentrations in 2018 and 2030. As shown in Figure 7-36, in 2018, annual percent changes in
ambient concentrations of ethanol are less than 1 percent across the country, with absolute
reductions of up to 0.1 ppb in some places. In 2030, some parts of the country, especially urban
areas, are projected to have reductions in ethanol concentrations on the order of 1 to 10 percent
as a result of the rule (Figure 7-37). 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 7-36 Changes in Ethanol Ambient Concentrations Between the Reference Case and
the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7-112
-------
Scales show rrodelM dianges betwwn ths i«fartnc« ar
Map CDk>rs *> riot moicsle ;he s«vemy cf expo
Figure 7-37 Changes in Ethanol Ambient Concentrations Between the Reference Case and the
Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/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 rule. As shown in Figure 7-38 and Figure
7-39, annual percent changes in ambient concentrations of formaldehyde are less than 1 percent
across much of the country for 2018 but are on the order of 1 to 5 percent in 2030 in some urban
areas as a result of the rule. Figure 7-38 and Figure 7-39 also show that absolute changes in
ambient concentrations of formaldehyde are generally between 0.001 and 0.01 |ig/m3 in both
years, with some areas as high as 0.1 |ig/m3 in 2030.
7-113
-------
Figure 7-38 Changes in Formaldehyde Ambient Concentrations Between the Reference Case
and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
v,-ik.-. -.nu/. rrod*IM csiana« bstw#»n the rahranc* m
Scale tanges and increments may not be c;.r.[ie-:.t = t
Figure 7-39 Changes in Formaldehyde Ambient Concentrations Between the Reference Case
and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7-114
-------
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 rule may decrease
acetaldehyde concentrations in some urban areas by 1 to 2.5 percent in 2030 (Figure 7-40).
Changes in ambient concentrations of acetaldehyde are generally in the range of 0.01 |ig/m3 to -
0.01 |ig/m3 with decreases happening in the more populated areas and increases happening in
more rural areas (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.141
Figure 7-40 Changes in Acetaldehyde Ambient Concentrations Between the Reference Case and
the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7-115
-------
Scales Wow rrodelM dianges betwwn ths i«fartnc« ar
Map tutors *> riot HKJKSIB ;he seveffly cf expo
Figure 7-41 Changes in Acetaldehyde Ambient Concentrations Between the Reference Case and
the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Naphthalene
Our modeling projects reductions in naphthalene concentrations in 2018 and 2030. As
shown in Figure 7-42 , annual percent changes in ambient concentrations of naphthalene are
between 1 and 2.5 percent across much of the country for 2018, with small decreases in absolute
concentrations (less than 0.001 ug/m3). In 2030, reductions of naphthalene concentrations
generally range between 1 and 10 percent but are as high as 25 percent in some areas of the
Southeast, with corresponding absolute decreases in urban areas of up to 0.005 |ig/m3(Figure
7-43).
7-116
-------
Figure 7-42 Changes in Naphthalene Ambient Concentrations Between the Reference Case and
the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
MW wloit *> nW rviKVa 111* wv»nty of
Figure 7-43 Changes in Naphthalene Ambient Concentrations Between the Reference Case
and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Population Metrics
Although the reductions in ambient air toxics concentrations expected from the Tier 3
standards are generally small, they are projected to benefit the majority of the U.S. population.
As shown in Table 7-51, over 75 percent of the total U.S. population is projected to experience a
decrease in ambient benzene and 1,3-butadiene concentrations of at least 1 percent. Table 7-51
also shows that over 60 percent of the U.S population is projected to experience at least a 1
7-117
-------
percent decrease in ambient ethanol and acrolein concentrations, and over 35 percent would
experience a similar decrease in ambient formaldehyde concentrations with the standards.
Table 7-51 Percent of Total Population Experiencing Changes in Annual Ambient
Concentrations of Toxic Pollutants in 2030 as a Result of the Standards
Percent Change
<-50
> -50 to < -25
> -25 to < -10
>-10to<-5
> -5 to < -2.5
>-2.5to<-l
> -1 to < 1
> 1 to < 2.5
>2.5 to<5
>5 to< 10
> 10 to < 25
> 25 to < 50
>50
Benzene
2.29%
20.63%
27.50%
28.60%
20.97%
Acrolein
0.75%
12.72%
25.17%
24.62%
36.74%
1,3 -Butadiene
19.07%
27.29%
15.37%
18.33%
19.93%
Formaldehyde
0.60%
35.34%
64.06%
Ethanol
5.39%
24.08%
34.10%
36.43%
Acetaldehyde
11.77%
88.23%
Naphthalene
10.74%
31.56%
20.58%
14.98%
22.14%
Of note, the 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 |ig/m3) for over 5 million people in 2030,
meaning that as a result of the 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.
7.2.4.5 Visibility
As described in Section 6.2.1 of this 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.
7.2.4.5.1 Current Visibility Levels
Designated PM2.5 nonattainment areas indicate that, as of December 5, 2013, over 84
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
7-118
-------
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.142
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.143'144'145 In summary,
visibility impairment is experienced throughout the U.S., in multi-state regions, urban areas, and
remote Mandatory Class I Federal areas.
7.2.4.5.2 Projected Visibility Levels
Air quality modeling conducted for the final action was used to project visibility
conditions in 137 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 rule would improve visibility in all these areas.11 The average visibility on the 20 percent
worst days at all modeled Mandatory Class I Federal areas is projected to improve by 0.02
deciviews, or 0.16 percent, in 2030. The greatest improvement in visibilities will be seen in
Craters of the Moon National Monument, where visibility is projected to improve by 0.7 percent
(0.09 DV) in 2030 due to the standards. Table 7-52 contains the full visibility results from 2030
for the 137 analyzed areas.
Table 7-52 Visibility Levels (in Deciviews) for Mandatory Class I Federal Areas on the 20
Percent Worst Days with and without this Rule
Class 1 Area
(20% worst days)
Sipsey Wilderness
Upper Buffalo Wilderness
Chiricahua NM
Chiricahua Wilderness
Galiuro Wilderness
Grand Canyon NP
Mazatzal Wilderness
Mount Baldy Wilderness
Petrified Forest NP
Pine Mountain Wilderness
Saguaro NM
Superstition Wilderness
State
AL
AR
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
AZ
2007
Baseline
Visibility
(dv)a
28.32
25.86
12.22
12.22
12.22
11.97
13.40
11.79
13.02
13.40
13.63
13.81
2018
Reference
20.59
20.01
11.82
11.83
11.99
11.21
12.65
10.98
12.24
12.69
13.02
13.18
2018
TierS
Control
20.55
19.98
11.82
11.82
11.98
11.20
12.65
10.98
12.23
12.69
13.00
13.18
2030
Reference
20.43
19.93
12.38
12.38
12.41
11.31
12.88
11.24
12.37
12.93
13.04
13.38
2030
TierS
Control
20.37
19.88
12.37
12.37
12.40
11.30
12.85
11.22
12.35
12.91
12.99
13.34
Natural
Background
10.99
11.57
7.20
7.20
7.20
7.04
6.68
6.24
6.49
6.68
6.46
6.54
11 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.
7-119
-------
Class 1 Area
(20% worst days)
Sycamore Canyon
Wilderness
Agua Tibia Wilderness
Ansel Adams Wilderness
(Minarets)
Caribou Wilderness
Cucamonga Wilderness
Desolation Wilderness
Dome Land Wilderness
Emigrant Wilderness
Hoover Wilderness
John Muir Wilderness
Joshua Tree NM
Kaiser Wilderness
Kings Canyon NP
Lassen Volcanic NP
Lava Beds NM
Marble Mountain
Wilderness
Mokelumne Wilderness
Pinnacles NM
Point Reyes NS
Redwood NP
San Gabriel Wilderness
San Gorgonio Wilderness
San Jacinto Wilderness
San Rafael Wilderness
Sequoia NP
South Warner Wilderness
Thousand Lakes
Wilderness
Ventana Wilderness
Yolla Bolly Middle Eel
Wilderness
Yosemite NP
Black Canyon of the
Gunnison NM
Eagles Nest Wilderness
Flat Tops Wilderness
Great Sand Dunes NM
State
AZ
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CA
CO
CO
CO
CO
2007
Baseline
Visibility
(dv)a
15.18
20.92
15.72
15.99
18.03
13.62
19.23
16.87
12.19
15.72
17.83
15.72
23.39
15.99
14.17
17.34
13.62
18.37
22.03
19.14
18.03
20.48
20.48
19.20
23.39
14.17
15.99
18.37
17.34
16.87
10.04
8.94
8.94
11.44
2018
Reference
14.94
17.67
14.57
15.54
15.37
12.89
17.89
15.84
11.49
14.76
15.75
14.80
21.56
15.52
13.78
17.02
12.88
16.44
21.04
18.72
15.71
17.68
17.76
17.46
21.28
13.60
15.53
16.79
17.06
15.98
9.21
7.98
8.26
10.57
2018
TierS
Control
14.94
17.66
14.57
15.54
15.36
12.89
17.89
15.84
11.48
14.76
15.75
14.80
21.55
15.52
13.78
17.01
12.88
16.43
21.03
18.70
15.71
17.68
17.76
17.46
21.28
13.60
15.53
16.79
17.06
15.98
9.20
7.97
8.26
10.56
2030
Reference
15.03
16.85
14.38
15.48
14.91
12.76
17.60
15.67
11.41
14.60
15.33
14.59
21.06
15.45
13.68
16.91
12.75
16.05
20.71
18.43
15.31
16.94
16.95
17.10
20.74
13.49
15.46
16.50
16.99
15.85
9.26
7.97
8.28
10.59
2030
TierS
Control
15.02
16.85
14.38
15.48
14.90
12.75
17.60
15.66
11.41
14.60
15.32
14.59
21.05
15.45
13.67
16.91
12.75
16.05
20.71
18.42
15.30
16.93
16.95
17.10
20.73
13.49
15.45
16.49
16.99
15.84
9.24
7.93
8.27
10.57
Natural
Background
6.65
7.64
7.12
7.31
6.99
6.05
7.46
7.64
7.71
7.12
7.19
7.12
7.70
7.31
7.85
7.90
6.05
7.99
15.77
13.91
6.99
7.30
7.30
7.57
7.70
7.85
7.31
7.99
7.90
7.64
6.21
6.06
6.06
6.66
7-120
-------
Class 1 Area
(20% worst days)
La Garita Wilderness
Maroon Bells-Snowmass
Wilderness
Mesa Verde NP
Mount Zirkel Wilderness
Rawah Wilderness
Rocky Mountain NP
Weminuche Wilderness
West Elk Wilderness
Chassahowitzka
Everglades NP
St. Marks
Cohutta Wilderness
Okefenokee
Wolf Island
Craters of the Moon NM
Sawtooth Wilderness
Mammoth Cave NP
Acadia NP
Moosehorn
Roosevelt Campobello
International Park
Isle Royale NP
Seney
Boundary Waters Canoe
Area
Voyageurs NP
Hercules-Glades
Wilderness
Mingo
Bob Marshall Wilderness
Cabinet Mountains
Wilderness
Glacier NP
Medicine Lake
Mission Mountains
Wilderness
Red Rock Lakes
Scapegoat Wilderness
ULBend
State
CO
CO
CO
CO
CO
CO
CO
CO
FL
FL
FL
GA
GA
GA
ID
ID
KY
ME
ME
ME
MI
MI
MN
MN
MO
MO
MT
MT
MT
MT
MT
MT
MT
MT
2007
Baseline
Visibility
(dv)a
10.04
8.94
11.28
9.72
9.72
12.62
10.04
8.94
23.68
20.41
25.58
28.01
26.00
26.00
13.63
14.76
30.68
21.45
19.92
19.92
21.76
24.21
20.05
19.78
26.05
27.08
15.32
13.47
18.70
18.02
15.32
11.53
15.32
14.86
2018
Reference
9.36
8.15
10.48
9.12
8.92
11.66
9.38
8.12
18.63
17.43
20.07
18.77
21.32
20.53
12.91
14.61
21.59
17.41
16.23
16.45
18.49
20.30
17.05
17.60
20.36
21.09
15.13
13.16
18.39
16.67
15.08
11.20
15.17
14.41
2018
TierS
Control
9.35
8.14
10.47
9.11
8.91
11.64
9.37
8.11
18.59
17.42
20.04
18.73
21.30
20.51
12.86
14.61
21.55
17.38
16.21
16.43
18.45
20.26
17.01
17.57
20.32
21.06
15.13
13.15
18.38
16.66
15.07
11.19
15.17
14.41
2030
Reference
9.44
8.18
10.57
9.10
8.88
11.55
9.45
8.18
18.38
17.28
19.86
18.59
21.33
20.45
12.63
14.58
21.47
17.22
16.14
16.34
18.21
20.17
16.77
17.35
20.21
20.88
15.06
13.01
18.23
16.47
14.98
11.13
15.12
14.37
2030
TierS
Control
9.43
8.17
10.55
9.08
8.86
11.50
9.44
8.16
18.31
17.25
19.81
18.52
21.31
20.41
12.54
14.57
21.41
17.19
16.12
16.32
18.13
20.09
16.70
17.29
20.14
20.83
15.05
13.00
18.21
16.45
14.97
11.11
15.11
14.36
Natural
Background
6.21
6.06
6.81
6.08
6.08
7.15
6.21
6.06
11.03
12.15
11.67
10.78
11.44
11.44
7.53
6.42
11.08
12.43
12.01
12.01
12.37
12.65
11.61
12.06
11.30
11.62
7.73
7.52
9.18
7.89
7.73
6.44
7.73
8.16
7-121
-------
Class 1 Area
(20% worst days)
Linville Gorge Wilderness
Shining Rock Wilderness
Lostwood
Great Gulf Wilderness
Presidential Range-Dry
River Wilderness
Brigantine
Bandelier NM
Bosque del Apache
Carlsbad Caverns NP
Gila Wilderness
Pecos Wilderness
San Pedro Parks Wilderness
Wheeler Peak Wilderness
White Mountain
Wilderness
Jarbidge Wilderness
Wichita Mountains
Crater Lake NP
Diamond Peak Wilderness
Eagle Cap Wilderness
Gearhart Mountain
Wilderness
Hells Canyon Wilderness
Kalmiopsis Wilderness
Mount Hood Wilderness
Mount Jefferson
Wilderness
Mount Washington
Wilderness
Mountain Lakes Wilderness
Strawberry Mountain
Wilderness
Three Sisters Wilderness
Cape Romain
Badlands NP
Wind Cave NP
Great Smoky Mountains
NP
State
NC
NC
ND
NH
NH
NJ
NM
NM
NM
NM
NM
NM
NM
NM
NV
OK
OR
OR
OR
OR
OR
OR
OR
OR
OR
OR
OR
OR
SC
SD
SD
TN
2007
Baseline
Visibility
(dv)a
27.39
26.60
19.56
20.19
20.19
27.32
11.84
13.40
15.85
12.49
9.13
9.89
9.13
13.20
12.42
22.97
13.79
13.79
16.23
13.79
18.15
16.45
13.72
16.18
16.18
13.79
16.23
16.18
26.45
16.55
15.50
28.50
2018
Reference
18.40
18.17
18.58
15.15
15.05
20.66
10.81
12.32
15.19
11.94
8.19
9.06
8.13
12.34
12.17
19.63
13.33
13.23
15.61
13.35
17.54
15.82
12.71
15.58
15.57
13.28
15.37
15.63
19.75
15.25
14.41
19.57
2018
TierS
Control
18.37
18.13
18.57
15.13
15.03
20.63
10.79
12.30
15.18
11.94
8.18
9.05
8.13
12.33
12.16
19.60
13.32
13.22
15.59
13.35
17.50
15.81
12.68
15.57
15.55
13.28
15.34
15.61
19.72
15.24
14.39
19.52
2030
Reference
18.33
18.04
18.45
15.08
14.97
20.59
10.89
12.54
15.88
12.40
8.34
9.28
8.25
12.74
12.13
19.52
13.22
13.07
15.22
13.27
17.20
15.63
12.25
15.33
15.32
13.16
15.00
15.45
19.61
15.19
14.26
19.44
2030
TierS
Control
18.28
17.98
18.44
15.05
14.94
20.55
10.85
12.50
15.86
12.39
8.32
9.27
8.23
12.73
12.12
19.45
13.22
13.07
15.20
13.27
17.16
15.62
12.23
15.31
15.31
13.16
14.97
15.44
19.56
15.17
14.24
19.38
Natural
Background
11.22
11.47
8.00
11.99
11.99
12.24
6.26
6.73
6.65
6.66
6.08
5.72
6.08
6.80
7.87
7.53
7.62
7.62
8.92
7.62
8.32
9.44
8.43
8.79
8.79
7.62
8.92
8.79
12.12
8.06
7.71
11.24
7-122
-------
Class 1 Area
(20% worst days)
Joyce-Kilmer- Slickrock
Wilderness
Big Bend NP
Guadalupe Mountains NP
Arches NP
Bryce Canyon NP
Canyonlands NP
Capitol Reef NP
James River Face
Wilderness
Shenandoah NP
Lye Brook Wilderness
Alpine Lake Wilderness
Glacier Peak Wilderness
Goat Rocks Wilderness
Mount Adams Wilderness
Mount Rainier NP
North Cascades NP
Olympic NP
Pasayten Wilderness
Dolly Sods Wilderness
Otter Creek Wilderness
Bridger Wilderness
Fitzpatrick Wilderness
Grand Teton NP
Teton Wilderness
Yellowstone NP
State
TN
TX
TX
UT
UT
UT
UT
VA
VA
VT
WA
WA
WA
WA
WA
WA
WA
WA
WV
WV
WY
WY
WY
WY
WY
2007
Baseline
Visibility
(dv)a
28.50
16.69
15.85
11.02
11.88
11.02
11.30
27.29
27.26
23.01
16.09
13.72
12.66
12.66
16.38
13.72
15.20
14.09
27.55
27.55
10.68
10.68
11.53
11.53
11.53
2018
Reference
19.65
16.39
15.23
10.33
11.40
10.50
10.73
19.05
17.67
16.74
14.87
12.78
11.92
12.04
15.53
12.87
14.30
13.51
17.97
18.11
10.23
10.21
11.14
11.18
11.26
2018
TierS
Control
19.61
16.38
15.22
10.32
11.40
10.48
10.72
19.02
17.63
16.70
14.84
12.77
11.90
12.02
15.52
12.86
14.28
13.50
17.94
18.07
10.22
10.21
11.13
11.18
11.26
2030
Reference
19.52
17.32
15.94
10.30
11.39
10.57
10.74
18.89
17.60
16.58
14.22
12.56
11.66
11.77
15.25
12.71
13.94
13.26
17.99
18.08
10.20
10.18
11.09
11.15
11.23
2030
TierS
Control
19.46
17.31
15.92
10.27
11.37
10.55
10.72
18.83
17.54
16.53
14.17
12.54
11.64
11.75
15.24
12.70
13.92
13.25
17.95
18.04
10.19
10.17
11.07
11.14
11.22
Natural
Background
11.24
7.16
6.65
6.43
6.80
6.43
6.03
11.13
11.35
11.73
8.43
8.39
8.35
8.35
8.54
8.01
8.44
8.25
10.39
10.39
6.45
6.45
6.44
6.44
6.44
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.
7.2.4.6 Deposition of Nitrogen and Sulfur
As described in Section 6.2.2 of this 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.
7-123
-------
7.2.4.6.1 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.JJ 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.KK
"U.S. EPA. (2012). U.S. EPA's Report on the Environment. Data accessed online February 15, 2012 at:
http://clpub.epa. gov/eroe/index.cfm?fuseaction=detail.viewPDF&ch=46&lShowInd=0&subtop=341&lv=list.listBy
Chapter&r=216610 and contained in Docket EPA-HQ-OAR-2011-0135.
KK U.S. EPA. (2012). U.S. EPA's Report on the Environment. Data accessed online February 15, 2012 at:
http ://clpub. epa. gov/eroe/index. cfm?fuseaction=detail.viewPDF&ch=46&lSho wlnd=0&subtop=3 41 &lv=list. listBy
Chapter&r=216610 and contained in Docket EPA-HQ-OAR-2011-0135.
7-124
-------
A. Average total sullur deposition, 1989-1991
OS
0.7
B. Average tola! sullur deposition, 2005-2007
1J3
c
08
0.6
0,7
06
0.5
0.4
0.9
0.6
1.0
0.7
0.8
1.2
1.4
1.6
5.1
C
5Covera|)t: 37 monitoring sites in 1989-1991
and 72 monitoring sites in 2005-2007.
Data source: MOP, 2008; U.S. EPA, 2008
-10
Numbers indicate total sullur deposition (kilograms per hectare),
averaged over a 3-year period.
Sines of circles indicate the relative magnitude at total sulfur deposition.
Colors in circles indicate the breakdown oi total sullur deposition:
Dry sulfur deposition • Wet sultur deposition
Figure 7-44 Total Sulfur Deposition in the Contiguous U.S., 1989-1991 and 2005 -2007
7-125
-------
„
A. Average total nitrogen deposition, 1989-1991
i
6.8 6.5
1.7
* "*fr '"ft*8^
ViTfe «5Ejy
^^» ^^^^
6.0
U.'
B Average total nitrogen depasition, 20D5-2007
2,9
V
1.4
V
6.1 ^ a.o «
8J rA, "4r **
1.9
" 'J*
'-
^
*~
2.7
2.6
V
2.3
'
J Coverage: 37 monitoring sites in 1989-1991
and 72 monitoring sites in 2005-2007,
Da/a source: fiADP. 2008: U.S. EPA, 2008
Numbers indicate total nitrogen deposition (kilograms per hectare),
15 averaged over a 3-year period.
Sizes of circles indicate the relative magnitude of total nitrogen deposition.
Colors in circles indicate the breakdown of total nitrogen deposition:
Dry nitrogen deposition • Wet nitrogen deposition
Figure 7-45 Total Nitrogen Deposition in the Contiguous U.S., 1989-1991 and 2005-2007
7-126
-------
7.2.4.6.2 Projected Levels of Nitrogen and Sulfur Deposition
Our air quality modeling projects decreases in both nitrogen and sulfur deposition due to
this rule. Figure 7-46 shows that for nitrogen deposition by 2030 the standards would result in
annual percent decreases of more than 2.5 percent in most urban areas with decreases of more
than 5 percent in urban areas in Nevada, Florida, Georgia and Virginia. In addition, smaller
decreases, in the 1 to 1.5 percent range, would occur over much of the rest of the country.
I
Legend
H
H
••
H
••
HH
••
= -5.0%
-5.0 to <= -2.5
-2.5 to <= -2.0
-2.0 to <= -1 ,5
-1.5 to <= -1.0
-1,0to<=-0.5
-0.5 to <= 0.5
0.5 to <= 1.0
1.0 to <= 1.5
1.5to<=2.0
2.0
Percent Change in Annual Total Nitrogen Deposition -- 2030rg_ctl minus 2030rg_ref
Figure 7-46 Percent Change in Annual Total Nitrogen Deposition over the U.S. Modeling
Domain as a Result of the Tier 3 Standards
Figure 7-47 shows that for sulfur deposition the standards will result in annual percent
decreases of more than 2 percent in some urban 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.
7-127
-------
Legend
••
••
IH
^B
CH
^B
••
- -2.5%
-2.510
-2.0 to
-1.5 to
-1.0 to
-0.5 to
0.510
1.0 to
1.5 to
2.010
= -2.0
= -1.5
= -1,0
= -0.5
= 0.5
= 1.0
= 1.5
= 2.0
= 2.5
2.5
Percent Change in Annual Total Sulfur Deposition - 2030rg_ctl minus 2030rg_ref
Figure 7-47 Percent Change in Annual Total Sulfur Deposition over the U.S. Modeling
Domain as a Result of the Tier 3 Standards
7.3 Greenhouse Gas Emission Impacts
Reductions in nitrous oxide (N2O) emissions and methane (CH/i) emissions, both potent
greenhouse gas emissions (with global warming potentials 298 and 25 times greater than CC>2,
respectively),LL are projected for gasoline cars and trucks due to the 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.146 EPA's sulfur study, detailed in Section 7.1.3.4.1, found close to a 30
percent reduction in CFLi emissions over the FTP when sulfur was reduced from 28 to 5 ppm (the
EPA study did not measure N2O emissions).
LL 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). N2O has a 100-year
GWP of 298 and CH4 has a 100-year GWP of 25 according to the 2007 IPCC AR4.
7-128
-------
Several studies have also established that reductions in tailpipe standards for NOx and
HC result in reductions in N2O and CH4, respectively. N2O is unique in that it is not formed
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.147'148 A study by Meffert, et al. established a strong correlation
between improvements in NOx catalytic conversion efficiency and reductions in N2O
emissions.149 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 (CARB) on 37
passenger cars and light-duty trucks ranging from the mid-1980s through early 2000's Low
Emission Vehicle (LEV) technology.150 Another study by Winer, et al (2005) tested 134 light-
duty vehicles, including passenger cars and light-duty trucks, as part of the CARB's vehicle
surveillance program, varying from Tier 0 to ULEV in emissions standards.151 These two studies
reported N2O:NOx ratios of 0.06 and 0.095, respectively, and supported the application of
N2O:NOx ratios to NOx emissions as a reasonable method for estimating N2O emission
inventories. Meszler, et al. subsequently analyzed the dataset by Behrentz and found that for
vehicles equipped with more modern controls, N2O emissions increased with vehicle mileage,
suggesting deterioration in N2O emissions as vehicles age.
152
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 7-25)). The estimated
methane reductions from Tier 3, using the 100-year global warming potential of 25 according to
the 2007 IPCC AR4, are shown in Table 7-53.
Table 7-53 Estimated Reduction in CH4 from Tier 3 Program (MMTCO2eq)
Reference case onroad
mobile emissions
Control case onroad
mobile emissions
Reduction
2018
1.5
1.4
0.1
2030
1.2
0.9
0.3
In contrast, 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 Winer and
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
7-129
-------
inventories that take into account the N2O:NOx link, because they do not account for the factors
affecting light-duty NOx emissions, such as deterioration, aggressive driving, and lower NOx
standards.
In an effort to estimate the N2O inventory accounting for the factors that affect N2O
emissions, we analyzed the data collected by CARB from 2000 to 2007, as a part of CARB's
vehicle surveillance program. The data described in Winer and Behrentz were also collected as a
part of the same program. However, the data we analyzed, available in the docket,153 is the most
comprehensive since it included additional tests that CARB collected since the studies by Winer
and Behrentz. A regression analysis was performed to examine the relationship between N2O
and NOx, and the resulting regression model correlating N2O emissions to NOx emissions is
shown in Equation 7-4. The details of the regression analysis are documented in a separate
memorandum to the Tier 3 docket.154
N 0 = e(>3"12854+0.62825 (ln(MOA")))
Equation 7-4 Regression Model Correlating NiO Emissions to NOx Emissions
The N2O reductions due to Tier 3 vehicle and sulfur standards are estimated by
employing two different methodologies, resulting in a range of reductions. The first method
applies the relationship between N2O and NOx from the regression model in Equation 7-4 to
NOx inventories from both Tier 3 and pre-Tier 3 vehicles. The second method applies the
regression of N2O and NOx only to Tier 3 fleet (2017 and later model year vehicles) and applies
the percent reduction in N2O from the UC Riverside study to MOVES inventory estimates to
pre-Tier 3 vehicles. These two methods are outlined in Table 7-54, along with the range of
reductions that result.
7-130
-------
Table 7-54 Estimated Reduction in NiO from the Tier 3 Program
2018
2030
Reduction from Tier 3 fleet
NOx reduction from Tier 3 fleet due to
vehicle and sulfur standards (U.S. Short Tons)
N2O reduction based on the regression
(MMTCO2eqa)
21,934
0.3
272,185
3.5
Reduction from pre-Tier 3 fleet
Method 1
Reference case onroad gasoline NOx
emissions (U.S. Short Tons)
Reference case onroad gasoline N2O
emissions based on the regression
(MMTCO2eqa)
Reduction from pre-Tier 3 fleet due to sulfur
standard (U.S. Short Tons)
N2O reduction based on the regression
(MMTCO2eqa)
Method 2
U.S. onroad gasoline N2O emissions from
pre-Tier 3 fleet using MOVES (MMTCO2eq)
Percent reduction in N2O going from 30 to 10
ppm
N2O reduction (MMTCO2eq)
Total Range of NiO Reduction
(MMTCO2eq)
2,599,284
31.6
242,434
1.9
8.9
23% b
1.9
2.2
377,811
5.0
56,324
0.5
1.5
23%
0.3
3.8-4.0
a Using GWP of 298
b 29 percent from 25ppm sulfur reduction in UC Riverside study scaled to 20 ppm reduction
The estimated N2O reduction is 2.2 million metric tons of carbon dioxide equivalent
(MMTCO2eq) in 2018, growing to a range between 3.8 and 4.0 MMTCO2eq in 2030. For 2018,
there was an agreement between the two methodologies described above, resulting in a single
estimate. Summing the results from Table 7-53 and Table 7-54, the total GHG reductions from
the Tier 3 rule is 2.3 million metric tons of carbon dioxide equivalent (MMTCO2eq) in 2018,
growing to a range between 4.1 and 4.3 MMTCO2eq in 2030.
These reductions will be offset to some degree by CO2 emissions associated with higher
energy use required in the process of removing sulfur within the refinery. As an extension of our
refinery-by-refinery cost modeling, we calculated the CO2 emission impacts of Tier 3 gasoline
sulfur control. We estimated refinery-specific changes in process energy and then applied
emission factors that correspond to those changes, on a refinery-by-refinery basis. As described
in Chapter 4.5 of the RIA, the results showed an increase of up to 1.9 MMTCO2e in 2018 and 1.6
MMTCO2e in 2030 for all U.S. refineries complying with the lower sulfur standards assuming
that the sulfur standards are fully phased-in. In 2018, the combined impact of CFLj and N2O
emission reductions from the vehicles and CO2 emission increases from the refineries shows a
7-131
-------
slight net decrease on a CC>2 equivalent basis. While still small, this net decrease grows to a
range between 2.5 to 2.7 MMTCO2e by 2030.
We do not expect the Tier 3 vehicle standards to result in any discernible changes in
vehicle CO2 emissions or fuel economy. Emissions of the pollutants that are controlled by the
Tier 3 program - NMOG, NOx, and PM - are not a function of the amount of fuel consumed,
since manufacturers need to design their catalytic emission control systems to reduce these
emissions regardless of their engine-out levels.
7-132
-------
References
1 U.S. Energy Information Administration, Annual Energy Outlook 2013 (April 15, 2013).
2 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.
3 MOVES (Motor Vehicle Emission Simulator) website: http://www.epa.gov/otaq/models/moves/index.htm.
4 U.S. EPA. 2014. Memorandum to Docket: Updates to MOVES for the Tier 3 FRM Analysis.
5 U.S. EPA, 2013. Assessing the Effect of Five Gasoline Properties on Exhaust Emissions from Light-Duty Vehicles
certified to Tier 2 Standards: Analysis of Data from EPAct Phase 3 (EPAct/V2/E-89). Final Report. EPA-420-R-
13-002.
6 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."
7 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 www.crcao.org.
9
! U.S. EPA. 2014. Memorandum to Docket: Updates to MOVES for the Tier 3 FRM Analysis.
U.S. EPA. 2013. Memorandum to Docket: Updates to MOVES for the Tier 3 NPRM Analysis.
10 U.S. EPA, 2008. Memorandum to Docket: Using MOVES to Generate Inventories for the RFS2 NPRM.
11 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/420rl 101 l.pdf.
12 U.S. EPA. 2011. Memorandum to Docket: Tier 3 Emission Rates for Gasoline Light Duty HC/CO/NOx Exhaust.
13 U.S. EPA. 2014. Memorandum to Docket: Updates to MOVES for the Tier 3 FRM Analysis.
14 U.S. EPA. 2014. Memorandum to Docket: Updates to MOVES for the Tier 3 FRM Analysis.
15 U.S. EPA. 2014. Memorandum to Docket: Updates to MOVES for the Tier 3 FRM Analysis.
16 Fulper, C. R.; Kishan, S.; Baldauf, R. W.; Sabisch, M.; Warila, I; Fujita, E. M.; Scarbro, C; Crews, W. S.; Snow,
R.; Gabele, P.; Santos, R.; Tierney, E.; Cantrell, B. Methods of Characterizing the Distribution of Exhaust Emissions
from Light-Duty, Gasoline-Powered Motor Vehicles in the U.S. Fleet. J. Air Waste Manage. Assoc. 2010, 60, 1376-
1387.
17 DeFries, T., Lindner, J., Kishan, S., Palacios, C. (2011), Investigation of Techniques for High Evaporative
Emissions Vehicle Detection: Denver Summer 2008 pilot Study at Lipan Street Station
18 DeFries, T., Palacios, C., Weatherby, M., Stanard, A., Kishan, S. (2013) Estimated Summer Hot-Soak
Distributions for Denver's Ken Caryl I/M Station Fleet.
19 CRC E-77 reports: Haskew, H., Liberty, T. (2008). Vehicle Evaporative Emission Mechanisms: A Pilot study,
CRC Project E-77; Haskew, H., Liberty, T. (2010), Enhanced EvaporativeEmission Vehicles (CRC E-77-2);
Haskew, H., Liberty, T. (2010), Evaporative Emissions from In-Use Vehicles: Test Fleet Expansion (CRC E-77-2b);
Haskew, H., Liberty, T. (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- available at http://crcao.org/reports/recentstudies.
20 Eastern Research Group, 2013. "Evaluation of the Effectiveness of On-Board Diagnostic (OBD) Systems in
Identifying Fuel Vapor Losses from Light-Duty Vehicles."
21 U.S. EPA, 2014. Memorandum to Docket: Development of Evaporative Emissions Calculations for Tier 3 FRM.
7-133
-------
22 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/420rll009.pdf.
23 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
24 See Chapter III and Appendix B of the Regulatory Impact Analysis for the Tier2/Sulfur Final Rule EPA 420-R-
99-023.
25 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.
26 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.
27 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
28 Peer Review of the Effects of Fuel Sulfur Level on Emissions from the In-Use Tier 2 Vehicles, EPA-HQ-OAR-
2011-0135-1847.
29 EPA In-Use Sulfur Report - Response to Peer-Review Comments, EPA-HQ-OAR-2011-0135-1848.
30 The Effects of Ultra-Low Sulfur Gasoline on Emissions from Tier 2 Vehicles in the In-Use Fleet, EPA-420-R-14-
002.
31 Ball, D., Clark, D., Moser, D. (2011). Effects of Fuel Sulfur on FTP NOX Emissions from a PZEV 4 Cylinder
Application. SAE Technical Paper 2011-01-0300.
32 Chapter 6 of the Regulatory Impact Analysis for the Control of Hazardous Air Pollutants from Mobile Sources
Final Rule, EPA 420-R-07-002, February 2007.
33 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.
http://www.epa.gov/otaq/models/nmim/420r05024.pdf.
34 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.htnrfl-2.
35 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.htnrf2008final.
36 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.htnrfcar-ems.
37 U.S. Energy Information Administration, Annual Energy Outlook 2013 (April 15, 2013).
38 U.S. EPA, 2014. Tier 3 FRM Emissions Inventory for Air Quality Modeling Technical Support Document.
39 U.S. EPA, 2014. Tier 3 FRM Emissions Inventory for Air Quality Modeling Technical Support Document.
40 User's Guide for the SMOKE-MOVES Integration Tool, July 14, 2010. Available at http://www.smoke-
model.org/smoke_moves_tool/.
41 For more information, please see the website for SMOKE: http://www.smoke-model.org/index.cfm.
42 Weather Research and Forcasting Model, http://wrf model.org.
43 Skamarock, W.C., Klemp, J.B., Dudhia, J., Gill, D.O., Barker, D.M., Duda, M.G., Huang, X., Wang, W., Powers,
J.G., A Description of the Advanced Research WRF Version 3, NCAR Technical Note NCAR/TN-475+STR, June
2008.
44 Meteorology-Chemistry Interface Processor (MCIP) version 3.6
(http://www.cmascenter.org/help/model_docs/mcip/3.6/ReleaseNotes).
7-134
-------
45 2008 National Emissions Inventory, Version 3 Technical Support Document (September 2013 - DRAFT)
(http://www.epa.gov/ttn/chief/net/2008neiv3/2008_neiv3_tsd_draft.pdf).
46 "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),
(http://www.epa.gov/otaq/models/moves/index.htm).
47 2008 National Emissions Inventory, Version 3 Technical Support Document draft, (September 2013)
http://www.epa.gov/ttn/chief/net/2008neiv3/2008 neiv3tsd draft.pdf.
48 "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).
49 U.S. EPA, 2014. Tier 3 FRM Emissions Inventory for Air Quality Modeling Technical Support Document.
50 Cook, R. 2012. Development of Air Quality Reference Case Upstream and Portable Fuel Container Inventories
for Tier 3 Final. Memorandum to Docket EPA-HQ-OAR-2010-0162.
51 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.
52 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.
53 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.
54 Brown, N., Allen, D., Amar, P., Kallos, G., McNider, R., Russell,, A., Stockwell, W. (September 2011). Final
Report: Fourth Peer Review of the CMAQ Model, NERL/ORD/EPA. U.S. EPA, Research Triangle Park, NC.,
http://www.epa.gov/asmdnerl/Reviews/201 l_CMAQ_Review_FinalReport.pdf.
55 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.
56 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.
57 U.S. EPA, 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.
58 Skamarock, W.C., Klemp, J.B., Dudhia, J., Gill, D.O., Barker, D.M., Duda, M.G., Huang, X., Wang, W., Powers,
J.G., 2008. A Description of the Advanced Research WRF Version 3.
59 Grell, G. A., Dudhia, A. J., and Stauffer, D. R., 1994. A description of the Fifth-Generation PennState/NCAR
Mesoscale Model (MM5). NCAR Technical Note NCAR/TN-398+STR. Available at http://www.mmm.
ucar.edu/mm5/doc 1 .html.
60 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:
http://www.cmascenter.org/.
61 Le Sager, P. Yantosca, B., Carouge, C. (2008). GEOS-CHEMv8-01-02 User's Guide, Atmospheric Chemistry
Modeling Group, Harvard University, Cambridge, MA, December 18, 2008.
62 U.S. EPA, 2004, Procedures for Estimating Future PM25 Values for the CAIR Final Rule by Application of the
(Revised) Speciated Modeled Attainment Test (SMAT)- Updated 11/8/04.
7-135
-------
63 U. S. EPA, 2011, Final Cross State Air Pollution Rule Air Quality Modeling TSD.
64 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.
65 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.
66 U.S. EPA. 2014. Memorandum to Docket: Updates to MOVES for the Tier 3 FRM Analysis.
67 Dodge, M.C., 2000. Chemical oxidant mechanisms for air quality modeling: critical review. Atmospheric
Environment 34, 2103-2130.
68 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.
69 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.
70 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.
71 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.
72 Sander, S.P., Fried!, R.R., Golden, D.M., Kurylo, M.J., Huie, R.E., Orkin, V.L., Moortgat, O.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.
73 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.
74 Sander, S.P., Fried!, R.R., Golden, D.M., Kurylo, M.J., Huie, R.E., Orkin, V.L., Moortgat, O.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.
75 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.
76 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.
77 Sander, S.P., Fried!, R.R., Golden, D.M., Kurylo, M.J., Huie, R.E., Orkin, V.L., Moortgat, O.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.
78 Sander, S.P., Fried!, R.R., Golden, D.M., Kurylo, M.J., Huie, R.E., Orkin, V.L., Moortgat, O.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.
7-136
-------
79 Luecken, D.J., Hutzell, W.T., Strum, M., Pouliot, G., 2012. Regional sources of atmospheric formaldehyde and
acetaldehyde, and implications for atmospheric modeling. Atmospheric Environment, 47, 477-490.
80 Atkinson R, Arey J (2003) Atmospheric Degradation of Volatile Organic Compounds. ChemRev 103: 4605-
4638.
81 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.
82 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.
83 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.
84 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/es0720412.html.
85 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.
86 Offenberg JH, CW Lewis, M Lewandowski, M Jaoui, TE Kleindienst, EO Edney (2007) Contributions of Toluene
and L-pinene to SOA Formed in an Irradiated Toluene/L-pinene,NOx/Air Mixture: Comparison of Results Using
14C Content and SOA Organic Tracer Methods, Environ Sci Technol 41: 3972-3976.
87 Pandis, S.N., Harley, R.A., Cass, G.R., Seinfeld, J.H. (1992) Secondary organic aerosol formation and transport.
Atmos Environ 26, 2269-2282.
88 Takekawa, H. Minoura, H. Yamazaki, S. (2003) Temperature dependence of secondary organic aerosol formation
by photo-oxidation of hydrocarbons. Atmos Environ 37: 3413-3424.
89 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.
90 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.
91 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.
92 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.
93 ByunDW, 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.
94 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.
95 Hildebrandt et al. 2009 ACP http://www.atmos-chem-phys.net/9/2973/2009/acp-9-2973-2009.html
96 Ng et al. 2007 ACP http://www.atmos-chem-phvs.net/7/3909/2007/acp-7-3909-2007.html.
97 U.S. EPA. 2011. 2005 National-Scale Air Toxics Assessment.
http://www.epa.gov/ttn/atw/nata2005/risksum.html.
7-137
-------
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.
99 Ng et al. 2007 ACP http://www.atmos-chem-phys.net/7/3909/2007/acp-7-3909-2007.html.
100 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.
101 Kleindienst, T.E. (2008) Hypothetical SOA Production from Ethanol Photooxidation. Memo to the Docket EPA-
HQ-OAR-2005-0161.
102 Turpin, B.J., Huntzicker, J.J., Larson, S.M., Cass, G.R. (1991) Los Angeles Summer Midday Paniculate Carbon:
Primary and Secondary Aerosol. Environ Sci Technol 25: 1788-1793.
103 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.
104 Bae M-S, Schauer JJ, Turner JR (2006) Estimation of the Monthly Average Ratios of Organic Mass to Organic
Carbon for Fine Paniculate Matter at an Urban Site, Aerosol Sci Technol 40(12): 1123-1139.
http://dx.doi.org/10.1080/02786820601004085.
105 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.
106 Offenberg JH, CW Lewis, M Lewandowski, M Jaoui, TE Kleindienst, EO Edney (2007) Contributions of
Toluene and L-pinene to SOA Formed in an Irradiated Toluene/L-pinene,NOx/Air Mixture: Comparison of Results
Using 14C Content and SOA Organic Tracer Methods, Environ Sci Technol 41: 3972-3976.
107 Claeys M, R Szmigielski, I Kourtchev, P Van der Veken, R Vermeylen, W Maenhaut, M Jaoui, TE Kleindienst,
M Lewandowski, JH Offenberg, EO Edney (2007) Hydroxydicarboxylic acids: Markers for secondary organic
aerosol from the photooxidation of ~-pinene. Environ Sci Technol 41(5): 1628-1634.
108 Edney EO, TE Kleindienst, M Jaoui, M Lewandowski, JH Offenberg, W Wang, M Claeys (2005) Formation of
2-methyl tetrols and 2-methylglyceric acid in secondary organic aerosol from laboratory irradiated
isoprene/NOx/SO2/air mixtures and their detection in ambient PM2 5 samples collected in the Eastern United States.
Atmos Environ 39: 5281-5289.
109 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.
110 Kleindienst TE, TS Conver, CD Mclver, 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.
111 Kleindienst TE, TS Conver, CD Mclver, 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.
112 Pye et al. 2013 ES&T Epoxide pathways improve model predictions of isoprene markers and reveal key
role of acidity in aerosol formation, http://pubs.acs.org/doi/abs/io.iO2i/es4Q2io6h
113 Carton et al. 2010 ES&T Can biogenic SOA be controlled? http://pubs.acs.org/doi/full/10.1021/es903506b
114 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/es0720412.html.
115 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.
116 Henze DK, JH Seinfeld (2006) Global secondary organic aerosol from isoprene oxidation. Geophys Res Lett 33:
L09812. doi:10.1029/2006GL025976.
7-138
-------
117 Hildebrandt, L., Donahuel, N. M, Pandisl, 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.
118 Ng, N. L., Kroll, J. H., Chan, A. W. H., Chabra, P. S., Flagan, R. C., Seinfield, J. H., Secondary organic aerosol
formation from /w-xylene, toluene, and benzene, Atmospheric Chemistry and Physics Discussion, 7, 3909-3922,
2007. Docket EPA-HQ-OAR-2011-0135.
119 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.
120 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.
121 Carton, A.G., Bhave, P.V., Napelenok, S.L., Edney, E.G., Sarwar, g., Finder, R.W., Pouliot, G.A., Houyoux, M.,
(2010). Model Representation of Secondary Organic Aerosol in CMAQv4.7. Environ Sci Technol 44(22), 8553-
8560.
122 Parikh, H.M., Carton, 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.
123 Volkamer, R., J.L. Jimenez, F. SanMartini,K.Dzepina,Q. Zhang,D. Salcedo,L. T. Molina, D. R.Worsnop, andM.
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.
124 Carton, A.G., Bhave, P.V., Napelenok, S.L., Edney, E.G., Sarwar, g., Finder, R.W., Pouliot, G.A., Houyoux, M.,
(2010). Model Representation of Secondary Organic Aerosol in CMAQv4.7. Environ Sci Technol 44(22), 8553-
8560.
125 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.
126 Carton, A.G., Bhave, P.V., Napelenok, S.L., Edney, E.G., Sarwar, g., Finder, R.W., Pouliot, G.A., Houyoux, M.,
(2010). Model Representation of Secondary Organic Aerosol in CMAQv4.7. Environ Sci Technol 44(22), 8553-
8560.
127 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.
128 Sander, S.P., Friedl, R.R., Golden, D.M., Kurylo, M.J., Huie, R.E., Orkin, V.L., Moortgat, O.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.
129 Finlayson-Pitts BJ, Pitts JN Jr. (1986) Atmospheric Chemistry: Fundamentals and Experimental Techniques,
Wiley, New York.
130 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,
131http://www.cmascenter.org/help/model_docs/cmaq/4.7/RELEASE_NOTES.txt.
132 77 FR 30088 (May 21, 2012) and 77 FR 34221 (June 11, 2012).
133 77 FR 30088 (May 21, 2012) and 77 FR 34221 (June 11, 2012).
7-139
-------
133 U.S. EPA (2012). National Ambient Air Quality Standards for Paniculate Matter.
http://www.epa.gov/PM/2012/finalrule.pdf.
134 U.S. EPA (2012). Fact Sheet: Implementing the Standards.
http://www.epa.gov/airquality/particlepollution/2012/decfsimp.pdf.
135 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/20120120FS.pdf.
136 U.S. Environmental Protection Agency (2013). Revision to Ambient Nitrogen Dioxide Monitoring
Requirements. March?, 2013. http://www.epa.gov/airquality/nitrogenoxides/pdfs/20130307fr.pdf.
137 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.
139 U.S. EPA. (2011) 2005 National-Scale Air Toxics Assessment, http://www.epa.gov/ttn/atw/nata2005/. Docket
EP A-HQ-O AR-2011-0135.
140 U.S. EPA. (2011) Summary of Results for the 2005 National-Scale Assessment.
http://www.epa.gov/ttn/atw/nata2005/05pdf/sum_results.pdf.
141 Luecken, D,J, Clmorel, AJ. 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.
142 U.S. Environmental Protection Agency (EPA). 2009. Integrated Science Assessment for Paniculate Matter. U.S.
Environmental Protection Agency. Research Triangle Park. EPA/600/R-08/139F.
143 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.
144 Tnjonis, 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.
145 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.
146 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.
147 Michaels, H. (1998) Emissions of Nitrous Oxide fromHighway Mobile Sources, U.S. EPA EPA420-R-98-009.
148 Graham, L. (2006) Greenhouse Gas Emissions from 1997-2005 Model Year Light Duty Vehicles Environment
Canada ERMD Report #04-44.
149 Meffert, et al (2000) Analysis of Nitrous Oxide Emissions from Light Duty Passenger Cars, SAE 2000-01-1952.
150 Behrentz, et al. (2004) Measurements of nitrous oxide emissions from light-duty motor vehicles: a pilot study
Atmospheric Environment 4291-4303.
151 Winer, et al. (2005) Estimates of Nitrous Oxide Emissions and the Effects of Catalyst Composition and Aging,
State of California Air Resources Board 02-313.
152 Meszler, D. (2004) Light Duty Vehicle Methane and Nitrous Oxide Emissions: Greenhouse Gas Impacts Study
for Northeast States Center for a Clean Air Future.
153 Nitrous Oxide and Oxides of Nitrogen Emissions Data from California Air Resources Board's Vehicle
Surveillance Program.
154 U.S. EPA, 2014, Memorandum to Docket: Regression Analysis of Nitrous Oxide and Oxides of Nitrogen from
Motor Vehicles.
7-140
-------
7.A. Appendix to Chapter 7: Additional Air Toxics Emissions and
Air Quality Modeling Results
7A.1. Air Toxics Emissions
Table 7A-1 Mobile Source Air Toxics Included in Inventory Reductions
Pollutant ID number
144
132
137
145
134
140
141
146
130
139
135
24
40
143
138
136
142
170
70
171
71
26
27
172
72
63
173
73
20
174
74
175
75
Pollutant Name
1,2,3,4,6,7,8-Heptachlorodibenzofuran
1,2,3,4,6,7,8-Heptachlorodibenzo-p-Dioxin
1,2,3,4,7,8,9-Heptachlorodibenzofuran
1,2,3,4,7,8-Hexachlorodibenzofuran
1,2,3,4,7,8-Hexachlorodibenzo-p-Dioxin
1,2,3,6,7,8-Hexachlorodibenzofuran
1,2,3,6,7,8-Hexachlorodibenzo-p-Dioxin
1,2,3,7,8,9-Hexachlorodibenzofuran
1,2,3,7,8,9-Hexachlorodibenzo-p-Dioxin
1,2,3,7,8-Pentachlorodibenzofuran
1,2,3,7,8-Pentachlorodibenzo-p-Dioxin
1,3-Butadiene
2,2,4-Trimethylpentane
2,3,4,6,7,8-Hexachlorodibenzofuran
2,3,4,7,8-Pentachlorodibenzofuran
2,3,7,8-Tetrachlorodibenzofuran
2,3,7,8-Tetrachlorodibenzo-p-Dioxin
Acenaphthene gas
Acenaphthene particle
Acenaphthylene gas
Acenaphthylene particle
Acetaldehyde
Acrolein
Anthracene gas
Anthracene particle
Arsenic Compounds
Benz(a)anthracene gas
Benz(a)anthracene particle
Benzene
Benzo(a)pyrene gas
Benzo(a)pyrene particle
Benzo(b)fluoranthene gas
Benzo(b)fluoranthene particle
7A-1
-------
Pollutant ID number
176
76
177
77
64
65
178
78
168
68
21
41
169
69
181
81
25
42
182
82
66
61
60
62
22
185
23
67
133
131
183
83
43
184
84
44
45
46
Pollutant Name
Benzo(g,h,i)perylene gas
Benzo(g,h,i)perylene particle
Benzo(k)fluoranthene gas
Benzo(k)fluoranthene particle
Chromium 3+
Chromium 6+
Chrysene gas
Chrysene particle
Dibenzo(a,h)anthracene gas
Dibenzo(a,h)anthracene particle
Ethanol
Ethyl Benzene
Fluoranthene gas
Fluoranthene particle
Fluorene gas
Fluorene particle
Formaldehyde
Hexane
lndeno(l,2,3,c,d)pyrene gas
lndeno(l,2,3,c,d)pyrene particle
Manganese Compounds
Mercury Divalent Gaseous
Mercury Elemental Gaseous
Mercury Particulate
MTBE
Naphthalene gas
Naphthalene particle
Nickel Compounds
Octachlorodibenzofuran
Octachlorodibenzo-p-dioxin
Phenanthrene gas
Phenanthrene particle
Propionaldehyde
Pyrene gas
Pyrene particle
Styrene
Toluene
Xylene
7A-2
-------
7A.2. Seasonal Air Toxics Air Quality Modeling Results for 2018
The following section presents maps of seasonal changes in ambient concentrations of modeled
air toxics in 2018.
Benzene
Figure 7A-1 Winter Changes in Benzene Ambient Concentrations Between the Reference Case and
the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-2 Summer Changes in Benzene Ambient Concentrations Between the Reference Case
and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7A-3
-------
1,3-Butadiene
M f»nj»s lid inirnmtnts roiy net bs
Figure 7A-3 Winter Changes in 1,3-Butadiene Ambient Concentrations Between the Reference Case
and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-4 Summer Changes in 1,3-Butadiene Ambient Concentrations Between the Reference
Case and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7A-4
-------
Acrolein
Figure 7A-5 Winter Changes in Acrolein Ambient Concentrations Between the Reference Case and
the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-6 Summer Changes in Acrolein Ambient Concentrations Between the Reference Case
and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7A-5
-------
Ethanol
Figure 7A-7 Winter Changes in Ethanol Ambient Concentrations Between the Reference Case and
the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-8 Summer Changes in Ethanol Ambient Concentrations Between the Reference Case and
the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7A-6
-------
Formaldehyde
Figure 7A-9 Winter Changes in Formaldehyde Ambient Concentrations Between the Reference
Case and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-10 Summer Changes in Formaldehyde Ambient Concentrations Between the Reference
Case and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7A-7
-------
Acetaldehyde
Figure 7A-11 Winter Changes in Acetaldehyde Ambient Concentrations Between the Reference
Case and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-12 Summer Changes in Acetaldehyde Ambient Concentrations Between the Reference
Case and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7A-8
-------
Naphthalene
Figure 7A-13 Winter Changes in Naphthalene Ambient Concentrations Between the Reference Case
and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-14 Summer Changes in Naphthalene Ambient Concentrations Between the Reference
Case and the Control Case in 2018: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7A-9
-------
7A.3. Seasonal Air Toxics Air Quality Modeling Results for 2030
The following section presents maps of seasonal changes in ambient concentrations of
modeled air toxics in 2030.
Benzene
Figure 7A-15 Winter Changes in Benzene Ambient Concentrations Between the Reference Case and
the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-16 Summer Changes in Benzene Ambient Concentrations Between the Reference Case
and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7 A-10
-------
1,3-Butadiene
t incrinwfilt mat not M
Figure 7A-17 Winter Changes in 1,3-Butadiene Ambient Concentrations Between the Reference
Case and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
2030rg_ctl minus 20JOtg_rtt
Figure 7A-18 Summer Changes in 1,3-Butadiene Ambient Concentrations Between the Reference
Case and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7 A-11
-------
Acrolein
*s *n4 incrinwfilt mtt not M
Figure 7A-19 Winter Changes in Acrolein Ambient Concentrations Between the Reference Case
and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-20 Summer Changes in Acrolein Ambient Concentrations Between the Reference Case
and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7 A-12
-------
Ethanol
(flnj»s and inirnmtnts m*y
Figure 7A-21 Winter Changes in Ethanol Ambient Concentrations Between the Reference Case and
the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-22 Summer Changes in Ethanol Ambient Concentrations Between the Reference Case
and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7 A-13
-------
Formaldehyde
(flnj»s and inirnmtnts m*y
Figure 7A-23 Winter Changes in Formaldehyde Ambient Concentrations Between the Reference
Case and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-24 Summer Changes in Formaldehyde Ambient Concentrations Between the Reference
Case and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7 A-14
-------
Acetaldehyde
(flnj»s and inirnmtnts m*y
Figure 7A-25 Winter Changes in Acetaldehyde Ambient Concentrations Between the Reference
Case and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-26 Summer Changes in Acetaldehyde Ambient Concentrations Between the Reference
Case and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7 A-15
-------
Naphthalene
Figure 7A-27 Winter Changes in Naphthalene Ambient Concentrations Between the Reference Case
and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
Figure 7A-28 Summer Changes in Naphthalene Ambient Concentrations Between the Reference
Case and the Control Case in 2030: Percent Changes (left) and Absolute Changes in ug/m3 (right)
7 A-16
-------
Chapter 8 Comparison of Program Costs to Program Air Quality
Benefits
EPA traditionally evaluates the effectiveness of a final rule in terms of net benefits.
Section 8.1 below presents the cost-benefit analysis of the final rule.
8.1 Cost-Benefit Analysis
The net benefits of the final Tier 3 program are determined by the effects of the program
on the costs to comply with the vehicle and fuel aspects of the program along with the benefits of
improved air quality on health and the environment.
8.1.1 Program-Wide Costs
The costs that are incurred from our final program fall into three categories - for the Tier
3 exhaust standards, Tier 3 evaporative standards, and reductions in sulfur content of gasoline.
While we present these three categories of costs separately in this section, for purposes of the
calculation of cost per ton of emissions reduced analysis we have summed them to represent the
estimated costs of the final program.
All costs represent the fleet-weighted average of light-duty vehicles and trucks. All costs
are represented in 2011 dollars.
8.1.1.1 Vehicle Program Costs
The vehicle costs include the technology costs projected to meet the final exhaust and
evaporative standards including the facility and operating cost impacts, as detailed in RIA
Chapter 2 and shown in Table 8-1. The fleet mix of light-duty vehicles, light duty trucks, and
medium-duty trucks represents the MYs 2017-2025 fleet mix projected to result from the most
recent GHG and fuel economy rules. The final vehicle costs are lower than the values projected
in the proposal due to the items outlined in RIA Section 2.1 in addition to the exclusion of the
vehicle sales in California and the states that have adopted the LEV III program.
Table 8-1: Annual Vehicle Program Costs, 2011$
Year
2016
2017
2018
2019
2020
2021
2022
2023
Vehicle Exhaust
Emission Control
Costs (SMillion)
$0
$268
$539
$579
$599
$630
$640
$639
Vehicle Evaporative
Emission Control
Costs ($Million)
$0
$26
$73
$72
$98
$97
$121
$116
Operating
Costs
($Million)
$0
$0
-$1
-$2
-$3
-$5
-$6
-$8
Facility Costs
($Million)
$21
$4
$4
$4
$4
$4
$4
$4
Total Vehicle
Program Costs
($Million)a
$21
$297
$615
$653
$697
$725
$758
$751
8-1
-------
Year
2024
2025
2030
Vehicle Exhaust
Emission Control
Costs (SMillion)
$653
$668
$664
Vehicle Evaporative
Emission Control
Costs ($Million)
$114
$113
$113
Operating
Costs
($Million)
-$9
-$11
-$19
Facility Costs
($Million)
$4
$4
$4
Total Vehicle
Program Costs
($Million)a
$761
$773
$761
a These estimates include costs associated with the Tier 3 vehicle standards in all states except California and states
that have adopted the LEV III program.
8.1.1.2 Fuel Program Costs
The annual fuel costs of the Tier 3 program consist of the costs to the refiners to control
sulfur on a per gallon basis and the total number of gallons consumed annually in the U.S. except
in the state of California.
The fuel costs associated with the additional operating and capital costs to refiners to
meet the final sulfur average of 10 ppm, as described in detail in RIA Chapter 5, is 0.65 cents per
gallon. The annual fuel costs of the final Tier 3 program are lower than the proposed fuel costs,
0.89 cents, due to the reduction in cost per gallon to control sulfur.
The fuel consumption values used in the annual fuel cost projections are based on the
U.S. Energy Information Administration's Annual Energy Outlook (AEO) 2013. The latest
AEO projection used in the final rule cost projections contain lower annual fuel consumption
than the values used in the Tier 3 proposal because AEO 2013 now reflects the light-duty GHG
and fuel economy standards for the 2017-2025 model years. The light-duty standards are
projected to decrease the national gasoline fuel consumption due to more efficient vehicles.
The annual fuel consumption in AEO 2013 represents the national fuel usage in terms of
BTU. For the fuel program cost analysis, it was necessary to convert the national BTU fuel
consumption to millions of gallons of fuel consumed in the U.S. in all states except California.
This conversion process is explained here using a sample year of 2017. Based on AEO 2013, the
annual fuel consumption of motor gasoline and E85 in the transportation sector is projected to be
15,342 trillion BTU in 2017.1 AEO's million BTU/barrel conversion factors for E85 and Motor
Gasoline Average along with a conversion factor of 42 U.S. gallons of fuel per barrel were used
to calculate the annual fuel consumption in terms of million gallons.2 Based on these
conversions, the national fuel consumption in 2017 is projected to be 128,318 million gallons.
Finally, the national fuel consumption values from AEO 2013 were reduced for this analysis to
remove the fraction of gasoline sold in California. The AEO 2013 Prime Supplier Sales
Volumes of Motor Gasoline provides the sales of motor gasoline by state.3 Based on the average
between 1983 and 2012, the fraction of national gasoline sales in California was 3.6 percent.
After removing the fraction of fuel consumed in California, the annual fuel costs in 2017 are
calculated based on a fuel consumption level of 123,689 million gallons.
The Tier 3 fuel program provides flexibilities, including the ability for refiners to phase-
in the sulfur standards and earn early credits. The fuel program costs presented here assume a
start date of 2017 and do not reflect the sulfur control program phase in.
8-2
-------
A comparison of the annual fuel consumption values used in the proposal and this final
rule is shown in Figure 8-1.
Tier 3 Impacted Annual Fuel Consumption
Notice of Proposed Rulemaking vs. Final Rule
160 000
c
TO
00
'o
1
C
Q.
3
C
"3
s.
TO
C
<
—
NPRM
— Final Rule
2016 2018 2020 2022 2024 2026 2028 2030 2032
Year
Figure 8-1: Annual Fuel Consumption Comparison between NPRM and Final Rule
The projected annual fuel consumption and annual fuel costs of the final program are
listed in Table 8-2.
8-3
-------
Table 8-2: Annual Fuel Costs, 2011$
Year
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2030
Annual Fuel Consumption
Impacted by Tier 3
Program (million gallons)
0
123,698
122,962
122,121
121,015
119,625
118,190
116,645
115,066
113,335
107,065
Fuel Sulfur Control
Costs
($Million)a
$0
$804
$799
$794
$787
$778
$768
$758
$748
$737
$696
' These estimates include costs associated with the Tier 3 fuel standards in all states except California.
8.1.1.3 Total Costs
The sum of the vehicle technology costs to control exhaust and evaporative emissions, in
addition to the costs to control the sulfur level in the fuel, represent the total costs of the
program, as shown in Table 8-3.
Table 8-3: Total Annual Vehicle and Fuel Control Costs, 2011$
Year
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2030
Total Vehicle
Program Costs
($Million)
$21.4
$297
$615
$653
$697
$725
$758
$751
$761
$773
$761
Fuel Sulfur
Control Costs
($Million)
$0
$804
$799
$794
$787
$778
$768
$758
$748
$737
$696
Total Program Costs
($Million)a
$21
$1,101
$1,414
$1,447
$1,484
$1,503
$1,526
$1,509
$1,509
$1,510
$1,457
a These estimates include: (a) costs associated with the Tier 3 vehicle standards in all states except
California and states that have adopted the LEV III program and (b) the Tier 3 fuel standards in all states
except California.
8-4
-------
8.1.2 Quantified and Monetized Health and Environmental Impacts
This section presents EPA's analysis of the criteria pollutant-related health and
environmental impacts that would occur as a result of the final Tier 3 standards. The vehicles
and fuels subject to the final standards are significant sources of mobile source air pollution such
as direct PM, NOx, SOx, VOCs and air toxics. The standards will affect exhaust and evaporative
emissions of these pollutants from vehicles. Emissions of NOx (a precursor to ozone formation
and secondarily-formed PM2.s), SOx (a precursor to secondarily-formed PM2.s), VOCs (a
precursor to ozone formation and, to a lesser degree, secondarily-formed PIVb.s) and directly-
emitted PM2.5 contribute to ambient concentrations of PM2.5 and ozone. Exposure to ozone and
PM2.5 is linked to adverse human health impacts such as premature deaths as well as other
important public health and environmental effects.
The analysis in this section aims to characterize the benefits of the final standards by
answering two key questions:
1. What are the health and welfare effects of changes in ambient particulate matter
(PM2.s) and ozone air quality resulting from reductions in precursors including NOx and SO2?
2. What is the economic value of these effects?
For the final rulemaking, we have quantified and monetized the health and environmental
impacts in 2030, representing projected impacts associated with a year when the program is fully
implemented and most of the fleet is turned over. Overall, we estimate that the final standards
will lead to a net decrease in PIVb.s- and ozone-related health impacts in 2030. The estimated
decrease in population-weighted national average PM2.5 exposure results in a net decrease in
adverse PM-related human health impacts (the decrease in national population-weighted annual
average PM2.5 is 0.04 ug/m3 in 2030).A The estimated decrease in population-weighted national
average ozone exposure results in a net decrease in ozone-related health impacts (population-
weighted maximum 8-hour average ozone decreases by 0.32 ppb in 2030).
Using the lower end of EPA's range of preferred premature mortality estimates (Krewski
et al., 2009 for PM2.5 and Bell et al., 2004 for ozone),4'5 we estimate that by 2030,
implementation of the standards will reduce approximately 770 premature mortalities annually
and will yield between $6.7 and $7.4 billion in total annual benefits, depending on the discount
rate used.B The upper end of the range of avoided premature mortality estimates associated with
the final standards (based on Lepeule et al., 2012 for PM2.5 and Levy et al., 2005 for ozone)6'7
A Note that the national, population-weighted PM2 5 and ozone air quality metrics presented in this Chapter represent
an average for the entire, gridded U.S. CMAQ domain. These are different than the population-weighted PM25 and
ozone design value metrics presented in Chapter 7, which represent the average for areas with a current air quality
monitor.
B The monetized value of PM25-related mortality accounts for a twenty-year segmented cessation lag. To discount
the value of premature mortality that occurs at different points in the future, we apply both a 3 and 7 percent
discount rate. We also use both a 3 and 7 percent discount rate to value PM-related nonfatal heart attacks
(myocardial infarctions). Nonfatal myocardial infarctions (MI) are valued using age-specific cost-of-illness values
that reflect lost earnings and direct medical costs over a 5-year period following a nonfatal MI.
8-5
-------
results in approximately 2,000 premature mortalities avoided in 2030 and will yield between $18
and $19 billion in total benefits. Thus, even using the lower end of the range of premature
mortality estimates, the health impacts of the final standards presented in this rule are projected
to be substantial.
We note that of necessity, decisions on the emissions and other elements used in the air
quality modeling were made early in the analytical process for the final rulemaking. For this
reason, the modeled changes in emissions used to support the air quality and benefits analyses
are slightly different than those used to represent the final emissions impacts of the Tier 3
standards. The magnitude of the differences is small, however, and for that reason we do not
expect these differences to materially impact our cost-benefit conclusions. See Chapter 7.2.1.1
for more details.
8.1.2.1 Overview
This analysis reflects the impacts of the final Tier 3 rule in 2030 compared to a future-
year reference scenario without the program in place. Overall, we estimate that the final rule will
lead to a net decrease in PM2.5-related health and environmental impacts (see Section 7.2.4 for
more information about the air quality modeling results). The estimated decrease in population-
weighted national average PM2.5 exposure results in a net decrease in adverse PM-related human
health and environmental impacts (the decrease in national population weighted annual average
PM2.5 is 0.04 ug/m3 in 2030).
The air quality modeling also projects decreases in ozone concentrations (see Section
7.2.4). The overall estimated decrease in population-weighted national average ozone exposure
results in decreases in ozone-related health and environmental impacts (population weighted
maximum 8-hour average ozone decreases by 0.32 ppb in 2030).
We base our analysis of the program's impact on human health and the environment on
peer-reviewed studies of air quality and human health effects.8'9'10 Our benefits methods are
consistent with the RIA that accompanied the final revisions to the National Ambient Air Quality
Standards (NAAQS) for Particulate Matter, the 2008 final ozone NAAQS, and the 2010 ozone
NAAQS reconsideration. To model the ozone and PM air quality impacts of the final standards,
we used the Community Multiscale Air Quality (CMAQ) model (see Chapter 7.2.2). The
modeled ambient air quality data serves as an input to the Environmental Benefits Mapping and
Analysis Program version 4.065 (BenMAP).c BenMAP is a computer program developed by the
U.S. EPA that integrates a number of the modeling elements used in previous analyses (e.g.,
interpolation functions, population projections, health impact functions, valuation functions,
analysis and pooling methods) to translate modeled air concentration estimates into health effects
incidence estimates and monetized benefits estimates.
The range of total monetized ozone- and PM-related health impacts projected in 2030 is
presented in Table 8-4. We present total benefits based on the PM- and ozone-related premature
c Information on BenMAP, including downloads of the software, can be found at http://www.epa.gov/ttn/ecas/
benmodels.html.
8-6
-------
mortality function used. The benefits ranges therefore reflect the addition of each estimate of
ozone-related premature mortality (each with its own row in Table 8-4) to estimates of PM-
related premature mortality.
Table 8-4: Estimated 2030 Monetized PM-and Ozone-Related Health Benefits
2030 Total Ozone and PM Benefits - PM Mortality Derived from American Cancer Society Analysis and
Six-Cities Analysis3
Premature Ozone
Mortality Function
Multi-city analyses
Meta-analyses
Reference
Bell et al., 2004
Huang et al., 2005
Schwartz, 2005
Bell et al., 2005
Ito et al., 2005
Levy etal., 2005
Total Benefits
(Billions, 2011$, 3%
Discount Rate)b>c
Total: $7.4 - $15
PM:$6.0-$14
Ozone: $1.1
Total: $7.9 - $16
PM:$6.0-$14
Ozone: $1.7
Total: $8.0 - $16
PM:$6.0-$14
Ozone: $1.7
Total: $9.8 - $18
PM:$6.0-$14
Ozone: $3.6
Total: $11 -$19
PM:$6.0-$14
Ozone: $4.9
Total: $11 -$19
PM:$6.0-$14
Ozone: $5.0
Total Benefits
(Billions, 2011$, 7%
Discount Rate) b'c
Total: $6.7 - $14
PM:$5.4-$12
Ozone: $1.1
Total: $7.3 - $14
PM:$5.4-$12
Ozone: $1.7
Total: $7.3 - $14
PM:$5.4-$12
Ozone: $1.7
Total: $9.2 - $16
PM:$5.4-$12
Ozone: $3.6
Total: $11 -$18
PM:$5.4-$12
Ozone: $4.9
Total: $11 -$18
PM:$5.4-$12
Ozone: $5.0
"Total includes premature mortality-related and morbidity-related ozone and PM25 estimated benefits. Range
was developed by adding the estimate from the ozone premature mortality function to the estimate of PM2 5-
related premature mortality derived from either the ACS study (Krewski et al., 2009) or the Six-Cities study
(Lepeule et al., 2012). Range also reflects alternative estimates of non-fatal heart attacks avoided based on either
Peters et al. (2001) or a pooled estimate of four studies.
b Note that total benefits presented here do not include a number of unquantified benefits categories. A detailed
listing of unquantified health effects is provided in Table 8-5.
0 Results reflect the use of both a 3 and 7 percent discount rate, as recommended by EPA's Guidelines for
Preparing Economic Analyses and OMB Circular A-4. Results are rounded to two significant digits for ease of
presentation and computation. Totals may not sum due to rounding.
The benefits analysis presented in this chapter incorporates an array of policy and
technical changes that the Agency has adopted since the Tier 3 proposal's draft RIA. These
changes reflect EPA's work to update PM-related benefits reflected in the most recent PM
NAAQS.11 Below we note the aspects of this analysis that differ from the Tier 3 proposal's draft
RIA:12
• Incorporation of the newest American Cancer Society (ACS) mortality study and
newest Harvard Six Cities mortality study. In 2012, Lepeule et al. published an
extended analysis of the Six Cities cohort.13 Compared to the study it replaces
(Laden et al., 2006),14 this new analysis follows the cohort for a longer time and
includes more years of PM2.5 monitoring data. The all-cause PIVb.s mortality risk
coefficient drawn from Lepeule et al. is roughly similar to the Laden et al. risk
8-7
-------
coefficient applied in the EPA's recent analyses of long-term PM2.5 mortality and
has narrower confidence intervals.
In 2009, the Health Effects Institute published an extended analysis of the ACS
cohort (Krewski et al., 2009).15 Compared to the study it replaces (Pope et al.,
2002),16 this new analysis incorporates a number of methodological
improvements.0 The all-cause PM2.5 mortality risk estimate drawn from Krewski
et al. (2009) is identical to the Pope et al. (2002) risk estimate applied in recent
EPA analyses of long-term PM2.5 mortality but has narrower confidence intervals.
Updated health endpoints. We have removed the quantification of chronic
bronchitis from our main analysis. This change is consistent with the findings of
the PM Integrated Science Assessment (ISA) that the evidence for an association
between long-term exposure to PM2.5 and respiratory effects is more tenuous.17
Updated demographic data. We updated the population demographic data in
BenMAP to reflect the 2010 Census and future projections based on economic
1 &
forecasting models developed by Woods and Poole, Inc. These data replace the
earlier demographic projection data from Woods and Poole (2011).19
Incorporation of new morbidity studies. Since the publication of the 2004 Criteria
90 91
Document for Parti culate Matter, the publication of the more recent PM ISA,
and the Provisional Assessment of Recent Studies on Health Effects of Parti culate
Matter Exposure ("Provisional Assessment"),22 the epidemiological literature has
produced several new studies examining the association between short-term PM2.5
exposure and acute myocardial infarctions, respiratory and cardiovascular
hospitalizations, respiratory and cardiovascular emergency department visits,
acute respiratory symptoms and exacerbation of asthma, respiratory and
cardiovascular hospitalizations. Upon careful evaluation of this new literature, we
added several new studies to our health impact assessment; in many cases we
have replaced older single-city time-series studies with newer multi-city time-
series analyses.
Updated the survival rates for non-fatal acute myocardial infarctions. Based on
recent data from Agency for Healthcare Research and Quality's Healthcare
Utilization Project National Inpatient Sample database,23 we identified death rates
for adults hospitalized with acute myocardial infarction stratified by age. These
rates replace the survival rates from Rosamond et al. (1999).24
Updated hospital cost-of-illness (COI), including median wage data. In previous
benefits analyses, estimates of hospital charges and lengths of hospital stays were
based on discharge statistics provided by the Agency for Healthcare Research and
Quality's Healthcare Utilization Project National Inpatient Sample (NIS) database
D Refer to the 2012 PM NAAQS RIA for more detail regarding the studies themselves.
-------
for 2000. The version of BenMAP (version 4.0.65) used in this analysis updated
this information to use the 2007 database. The data source for the updated median
annual income is the 2007 American Community Survey.
The benefits in Table 8-4 include all of the estimated human health impacts we are able
to quantify and monetize at this time. However, the full complement of human health effects
associated with PM and ozone remain unquantified because of current limitations in methods or
available data. We have not quantified a number of known or suspected health effects linked
with ozone and PM for which appropriate health impact functions are not available or which do
not provide easily interpretable outcomes (e.g., changes in heart rate variability). These are
listed in Table 8-5. As a result, the health benefits quantified in this analysis are likely
underestimates of the total benefits attributable to the final program.
Table 8-5: Estimated Quantified and Unquantified Health Effects
BENEFITS
CATEGORY
SPECIFIC EFFECT
EFFECT HAS
BEEN
QUANTIFIED
EFFECT HAS
BEEN
MONETIZED
MORE
INFORMATION
Improved Human Health
Reduced
incidence of
premature
mortality and
morbidity from
exposure to PM2 5
Adult premature mortality based on
cohort study estimates and expert
elicitation estimates (age >25 or age
>30)
Infant mortality (age <1)
Non-fatal heart attacks (age > 18)
Hospital admissions — respiratory (all
ages)
Hospital admissions — cardiovascular
(age >20)
Emergency department visits for
asthma (all ages)
Acute bronchitis (age 8-12)
Lower respiratory symptoms (age 7-
14)
Upper respiratory symptoms
(asthmatics age 9-11)
Asthma exacerbation (asthmatics age
6-18)
Lost work days (age 18-65)
^
•/
v'
•/
•/
•/
v'
-------
BENEFITS
CATEGORY
Reduced
incidence of
premature
mortality and
morbidity from
exposure to
ozone
SPECIFIC EFFECT
Minor restricted-activity days (age
18-65)
Chronic Bronchitis (age >26)
Emergency department visits for
cardiovascular effects (all ages)
Strokes and cerebrovascular disease
(age 50-79)
Other cardiovascular effects (e.g.,
other ages)
Other respiratory effects (e.g.,
pulmonary function, non-asthma ER
visits, non-bronchitis chronic
diseases, other ages and populations)
Reproductive and developmental
effects (e.g., low birth weight, pre-
term births, etc.)
Cancer, mutagenicity, and
genotoxicity effects
Premature mortality based on short-
term study estimates (all ages)
Premature mortality based on long-
term study estimates (age 30-99)
Hospital admissions — respiratory
causes (age > 65)
Hospital admissions — respiratory
causes (age <2)
Emergency department visits for
asthma (all ages)
Minor restricted-activity days (age
18-65)
School absence days (age 5-17)
Decreased outdoor worker
productivity (age 18-65)
Other respiratory effects (e.g.,
premature aging of lungs)
EFFECT HAS
BEEN
QUANTIFIED
V
—
—
—
—
—
•/
—
•/
•/
v'
-------
BENEFITS
CATEGORY
Reduced
incidence of
morbidity from
exposure to air
toxics
SPECIFIC EFFECT
Cardiovascular and nervous system
effects
Reproductive and developmental
effects
Cancer (benzene, 1,3-butadiene,
formaldehyde, acetaldehyde)
Anemia (benzene)
Disruption of production of blood
components (benzene)
Reduction in the number of blood
platelets (benzene)
Excessive bone marrow formation
(benzene)
Depression of lymphocyte counts
(benzene)
Reproductive and developmental
effects (1,3-butadiene)
Irritation of eyes and mucus
membranes (formaldehyde)
Respiratory irritation (formaldehyde)
Asthma attacks in asthmatics
(formaldehyde)
Asthma-like symptoms in non-
asthmatics (formaldehyde)
Irritation of the eyes, skin, and
respiratory tract (acetaldehyde)
Upper respiratory tract irritation and
congestion (acrolein)
EFFECT HAS
BEEN
QUANTIFIED
—
—
EFFECT HAS
BEEN
MONETIZED
—
—
MORE
INFORMATION
Ozone ISAb
Ozone ISAb
IRISa'b
a We assess these benefits qualitatively because we do not have sufficient confidence in available data or
methods.
b We assess these benefits qualitatively because current evidence is only suggestive of causality or there are
other significant concerns over the strength of the association.
0 We assess these benefits qualitatively due to time and resource limitations for this analysis.
While there will be impacts associated with reductions in air toxic pollutant emissions
that result from the final program, we do not attempt to monetize those impacts. This is
primarily because currently available tools and methods to assess air toxics risk from mobile
sources at the national scale are not adequate for extrapolation to incidence estimations or
benefits assessment. The best suite of tools and methods currently available for assessment at
the national scale are those used in the National-Scale Air Toxics Assessment (NATA). The
EPA Science Advisory Board specifically commented in their review of the 1996 NATA that
these tools were not yet ready for use in a national-scale benefits analysis, because they did not
consider the full distribution of exposure and risk, or address sub-chronic health effects.26 While
EPA has since improved these tools, there remain critical limitations for estimating incidence
and assessing benefits of reducing mobile source air toxics.
8-11
-------
As part of the second prospective analysis of the benefits and costs of the Clean Air
Act,27 EPA conducted a case study analysis of the estimated health effects associated with
reducing exposure to benzene in Houston from implementation of the Clean Air Act. While
reviewing the draft report, EPA's Advisory Council on Clean Air Compliance Analysis
concluded that "the challenges for assessing progress in health improvement as a result of
reductions in emissions of hazardous air pollutants (HAPs) are daunting... due to a lack of
exposure-response functions, uncertainties in emissions inventories and background levels, the
difficulty of extrapolating risk estimates to low doses and the challenges of tracking health
progress for diseases, such as cancer, that have long latency periods."28 EPA continues to work
to address these limitations; however, we did not have the methods and tools available for
national-scale application in time for the analysis of the final program.E
The reduction in air pollution emissions that will result from the final program is
projected to have "welfare" co-benefits in addition to human health benefits, including changes
in visibility, materials damage, ecological effects from PM deposition, ecological effects from
nitrogen and sulfur emissions, vegetation effects from ozone exposure, and climate effects.F
Despite our goal to quantify and monetize as many of the benefits as possible for the final
rulemaking, the welfare co-benefits of the Tier 3 standards remain unquantified and
nonmonetized in this RIA due to data, methodology, and resource limitations. As a result, the
benefits quantified in this analysis are likely underestimates of the total benefits attributable to
the final program. We refer the reader to Chapter 6 of the PM NAAQS RIA for a complete
discussion of these welfare co-benefits.29
8.1.2.2 Human Health Impacts
Table 8-6 and Table 8-7 present the core estimates of annual PM2.5 and ozone health
impacts in the 48 contiguous U.S. states associated with the final Tier 3 program. For each
endpoint presented in Table 8-6 and Table 8-7, we provide both the point estimate and the 90
percent confidence interval.
Using EPA's preferred estimates, based on the American Cancer Society (ACS) and Six-
Cities studies and no threshold assumption in the model of mortality, we estimate that the final
program would result in between 660 and 1,500 cases of avoided PM2.s-related premature deaths
annually in 2030. A sensitivity analysis was conducted to understand the impact of alternative
concentration response functions suggested by experts in the field. As shown in Table 8-8, when
E In April, 2009, EPA hosted a workshop on estimating the benefits or reducing hazardous air pollutants. This
workshop built upon the work accomplished in the June 2000 Science Advisory Board/EPA Workshop on the
Benefits of Reductions in Exposure to Hazardous Air Pollutants, which generated thoughtful discussion on
approaches to estimating human health benefits from reductions in air toxics exposure, but no consensus was
reached on methods that could be implemented in the near term for a broad selection of air toxics. Please visit
http://epa.gov/air/toxicair/2009workshop.html for more information about the workshop and its associated materials.
F We project that the Tier 3 vehicle and fuel standards will reduce nitrous oxide (N2O) and methane (CH4) emissions
from vehicles. The reductions in these potent greenhouse gases will be offset to some degree by the increase in CO2
emissions from refineries. The combined impact is a net decrease on a CO2-equivalent basis and would yield a net
benefit if these reductions were monetized.
8-12
-------
the range of expert opinion is used, we estimate between 130 and 2,200 fewer premature
mortalities in 2030.
The range of ozone impacts is based on changes in risk estimated using several sources of
ozone-related mortality effect estimates. This analysis presents six alternative estimates for the
association based upon different functions reported in the scientific literature, derived from both
the National Morbidity, Mortality, and Air Pollution Study (NMMAPS)30'31'32 and from a series
of meta-analyses.33'34'35 This approach is not inconsistent with recommendations provided by
the NRC in their report (NRC, 2008) on the estimation of ozone-related mortality risk reductions,
"The committee recommends that the greatest emphasis be placed on estimates from new
systematic multicity analyses that use national databases of air pollution and mortality, such as in
the NMMAPS, without excluding consideration of meta-analyses of previously published
studies."36 For ozone-related premature mortality in 2030, we estimate a range of between 110
to 500 fewer premature mortalities.
Following these tables, we also provide a more comprehensive presentation of the
distributions of incidence generated using the available information from empirical studies and
expert elicitation.
Table 8-8 presents the distributions of the reduction in PM2.s-related premature mortality
based on the C-R distributions provided by each expert, as well as that from the data-derived
health impact functions, based on the statistical error associated with the ACS study and the Six-
Cities study. The 90 percent confidence interval for each separate estimate of PM-related
mortality is also provided.
In 2030, the effect estimates of nine of the twelve experts included in the elicitation panel
fall within the empirically-derived range provided by the ACS and Six-Cities studies. Only one
expert falls below this range, while two of the experts are above this range. Although the overall
range across experts is summarized in these tables, the full uncertainty in the estimates is
reflected by the results for the full set of 12 experts. The twelve experts' judgments as to the
likely mean effect estimate are not evenly distributed across the range illustrated by arraying the
highest and lowest expert means.
8-13
-------
Table 8-6: Estimated PM2.5-Related Health Impacts
Health Effect
Premature Mortality - Derived from epidemiology literature15
Adult, age 30+, ACS Cohort Study (Krewski et al., 2009)
Adult, age 25+, Six-Cities Study (Lepeule et al., 2012)
Infant, age <1 year (Woodruff et al., 1997)
Non-fatal myocardial infarction (adult, age 18 and over)
Peters etal. (2001)
Pooled estimate of 4 studies
Hospital admissions - respiratory (all ages)°'e
Hospital admissions - cardiovascular (adults, age >18)d
Emergency room visits for asthma (age 18 years and younger)6
Acute bronchitis, (children, age 8-12)e
Lower respiratory symptoms (children, age 7-14)
Upper respiratory symptoms (asthmatic children, age 9-18)
Asthma exacerbation (asthmatic children, age 6-18)
Work loss days
Minor restricted activity days (adults age 18-65)
2030 Annual
Reduction in
Incidence
(5tho/o . 95tho/oile)
660
(480 - 840)
1,500
(860-2,100)
790
(290 - 1,300)
85
(42 - 190)
210
(-38 - 380)
250
(130-440)
340
(-58 - 660)
980
(-35 - 2,000)
13,000
(6,000 - 19,000)
18,000
(5,600 - 30,000)
19,000
(2,300 - 37,000)
81,000
(70,000-91,000)
480,000
(400,000 - 550,000)
a Incidence is rounded to two significant digits. Estimates represent incidence within the 48 contiguous United
States.
b PM-related adult mortality based upon the most recent American Cancer Society (ACS) Cohort Study
(Krewski et al., 2009) and the most recent Six-Cities Study (Lepeule et al., 2012). Note that these are two
alternative estimates of adult mortality and should not be summed. PM-related infant mortality based upon a
study by Woodruff, Grillo, and Schoendorf, (1997).37
0 Respiratory hospital admissions for PM include admissions for chronic obstructive pulmonary disease
(COPD), pneumonia and asthma.
d Cardiovascular hospital admissions for PM include total cardiovascular and subcategories for ischemic heart
disease, dysrhythmias, and heart failure.
e The negative estimates at the 5th percentile confidence estimates for these morbidity endpoints reflect the
statistical power of the study used to calculate these health impacts. These results do not suggest that reducing
air pollution results in additional health impacts.
8-14
-------
Table 8-7: Estimated Ozone-Related Health Impacts
Health Effect
Premature Mortality, All agesb
Multi-City Analyses
Bell et al. (2004) - Non-accidental
Huang et al. (2005) - Cardiopulmonary
Schwartz (2005) - Non-accidental
Meta-analyses:
Bell et al. (2005) - All cause
Ito et al. (2005) - Non-accidental
Levy et al. (2005) - All cause
Hospital admissions- respiratory causes (adult, 65 and
older)c
Hospital admissions -respiratory causes (children, under 2)
Emergency room visit for asthma (all ages)d
Minor restricted activity days (adults, age 18-65)
School absence days
2030 Annual Reduction
in Incidence
(5th% - 95th%ile)
110
(46 - 170)
160
(74 - 250)
170
(68 - 270)
350
(190-510)
490
(320 - 660)
500
(360 - 630)
740
(87 - 1,400)
310
(160 - 450)
330
(-8 - 990)
600,000
(290,000 - 910,000)
210,000
(92,000 - 300,000)
a Incidence is rounded to two significant digits. Estimates represent incidence within the 48 contiguous U.S.
b Estimates of ozone-related premature mortality are based upon incidence estimates derived from several
alternative studies: Bell et al. (2004); Huang et al. (2005); Schwartz (2005); Bell et al. (2005); Ito et al.
(2005); Levy et al. (2005). The estimates of ozone-related premature mortality should therefore not be
summed.
0 Respiratory hospital admissions for ozone include admissions for all respiratory causes and subcategories for
COPD and pneumonia.
d The negative estimate at the 5th percentile confidence estimate for this morbidity endpoint reflects the
statistical power of the study used to calculate this health impact. This result does not suggest that reducing air
pollution results in additional health impacts.
Table 8-8: Results of Application of Expert Elicitation: Annual Reductions in Premature
Mortality in 2030 Associated with the Final Program
Source of Mortality
Estimate
Krewski et al. (2009)
Lepeuleetal. (2012)
Expert A
Expert B
Expert C
Expert D
Expert E
2030 Tier 3 Control
5th Percentile
480
860
340
88
480
58
1,100
Mean
660
1,500
1,700
1,400
1,300
950
2,200
95th Percentile
840
2,100
3,300
2,900
2,200
1,500
3,300
8-15
-------
Source of Mortality
Estimate
Expert F
Expert G
Expert H
Expert I
Expert J
Expert K
Expert L
2030 Tier 3 Control
5th Percentile
790
-1
-2
110
160
0
1
Mean
1,200
790
980
1,300
1,100
130
860
95th Percentile
1,700
1,400
2,300
2,300
2,300
590
1,800
8.1.2.3 Monetized Estimates of Human Health and Environmental Impacts
Table 8-9 presents the estimated monetary value of changes in the incidence of ozone and
PM2.s-related health and environmental effects. Total aggregate monetized benefits are
presented in Table 8-10. All monetized estimates are presented in 2011$. Where appropriate,
estimates account for growth in real gross domestic product (GDP) per capita between 2000 and
2030.° The monetized value of PM2.s-related mortality also accounts for a twenty-year
segmented cessation lag.H To discount the value of premature mortality that occurs at different
points in the future, we apply both a 3 and 7 percent discount rate. We also use both a 3 and 7
percent discount rate to value PM-related nonfatal heart attacks (myocardial infarctions).1
In addition to omitted benefits categories such as air toxics and various welfare effects,
not all known PM2.5- and ozone-related health and welfare effects could be quantified or
monetized. The estimate of total monetized health benefits of the final program is thus equal to
the subset of monetized PM2.5- and ozone-related health impacts we are able to quantify plus the
sum of the nonmonetized health and welfare benefits. Our estimate of total monetized benefits
in 2030 for the final program, using the ACS and Six-Cities PM mortality studies and the range
of ozone mortality assumptions, is between $7.4 and $19 billion billion, assuming a 3 percent
G Our analysis accounts for expected growth in real income over time. Economic theory argues that WTP for most
goods (such as environmental protection) will increase if real incomes increase. Benefits are therefore adjusted by
multiplying the unadjusted benefits by the appropriate adjustment factor to account for income growth over time.
For growth between 2000 and 2030, this factor is 1.23 for long-term mortality, 1.27 for chronic health impacts, and
1.08 for minor health impacts. For a complete discussion of how these adjustment factors were derived, we refer the
reader to the PM NAAQS regulatory impact analysis. Note that similar adjustments do not exist for cost-of-illness-
based unit values. For these, we apply the same unit value regardless of the future year of analysis.
H Based in part on prior SAB advice, EPA has typically assumed that there is a time lag between changes in
pollution exposures and the total realization of changes in health effects. Within the context of benefits analyses,
this term is often referred to as "cessation lag". The existence of such a lag is important for the valuation of
premature mortality incidence because economic theory suggests that benefits occurring in the future should be
discounted. In this analysis, we apply a twenty-year distributed lag to PM mortality reductions. This method is
consistent with the most recent recommendation by the EPA's Science Advisory Board. Refer to: EPA - Science
Advisory Board, 2004. Advisory Council on Clean Air Compliance Analysis Response to Agency Request on
Cessation Lag. Letter from the Health Effects Subcommittee to the U.S. Environmental Protection Agency
Administrator, December.
1 Nonfatal myocardial infarctions (MI) are valued using age-specific cost-of-illness values that reflect lost earnings
and direct medical costs over a 5-year period following a nonfatal MI.
8-16
-------
discount rate, or between $6.7 and $18 billion, assuming a 7 percent discount rate. As the results
indicate, total benefits are driven primarily by the reduction in PM2.5- and ozone-related
premature fatalities each year and represent the benefits of the Tier 3 program anticipated to
occur annually when the program is fully implemented and most of the fleet turned over.
The next largest benefit is for reductions in nonfatal heart attacks, although this value is
more than an order of magnitude lower than for premature mortality. Hospital admissions for
respiratory and cardiovascular causes, minor restricted activity days, and work loss days account
for the majority of the remaining benefits. The remaining categories each account for a small
percentage of total benefit; however, they represent a large number of avoided incidences
affecting many individuals. A comparison of the incidence table to the monetary benefits table
reveals that there is not always a close correspondence between the number of incidences
avoided for a given endpoint and the monetary value associated with that endpoint. For
example, there are many more work loss days than PM-related premature mortalities, yet work
loss days account for only a very small fraction of total monetized benefits. This reflects the fact
that many of the less severe health effects, while more common, are valued at a lower level than
the more severe health effects. Also, some effects, such as hospital admissions, are valued using
a proxy measure of willingness-to-pay (e.g., cost-of-illness). As such, the true value of these
effects may be higher than that reported here.
Table 8-9: Estimated Monetary Value of Changes in Incidence of Health and Welfare
Effects (millions of 2010$)
HEALTH ENDPOINTS
2030
(5THAND95TH
PERCENTILE)
PM2.5-Related Health Effects
Premature Mortality - Derived
from Epidemiology Studies'3'0
Adult, age 30+ - ACS study
(Krewski et al., 2009)
3% discount rate
7% discount rate
Adult, age 25+ - Six-Cities study
(Lepeule et al., 2012)
3% discount rate
7% discount rate
Infant Mortality, <1 year -
(Woodruff etal. 1997)
Non-fatal acute myocardial infarctions
Peters etal., 2001
3% discount rate
7% discount rate
Pooled estimate of 4 studies
3% discount rate
7% discount rate
$6,100
($910-$14,000)
$5,500
($820-$13,OOP)
$14,000
($2,000 - $33,000)
$12,000
($1,800-$30,000)
$13
($1.8-$32)
$96
($21 -$230)
$93
($19-$220)
$10
($2.6 - $27)
$10
($2.4 - $27)
8-17
-------
Hospital admissions for respiratory causes'1
Hospital admissions for cardiovascular causes
Emergency room visits for asthmad
Acute bronchitis (children, age 8-12)d
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthma, 9-11)
Asthma exacerbations
Work loss days
Minor restricted-activity days (MRADs)
$5.9
(-$1.6 -$11)
$9.9
($5.0 -$17)
$0.15
(-$0.02 - $0.29)
$0.49
(-$0.02 -$1.2)
$0.27
($0.11 -$0.51)
$0.62
($0.18 -$1.4)
$1.1
($0.14 -$2.7)
$12
($11 -$14)
$34
($20 - $49)
Ozone-Related Health Effects
Premature Mortality, All ages -
Derived from Multi-city analyses
Premature Mortality, All ages -
Derived from Meta-analyses
Bell etal, 2004
Huang etal., 2005
Schwartz, 2005
Bell etal., 2005
Ito et al., 2005
Levy et al., 2005
Hospital admissions- respiratory causes (adult, 65 and older)
Hospital admissions- respiratory causes (children, under 2)
Emergency room visit for asthma (all ages)
Minor restricted activity days (adults, age 18-65)
School absence days
$1,100
($150 -$2,800)
$1,600
($220 -$4,100)
$1,700
($220 - $4,400)
$3,600
($5 10 -$8,800)
$5,000
($740 -$12,000
$5,100
($760 -$12,000)
$21
($2.5 - $39)
$3.7
($1.9 -$5.4)
$0.14
(-$0.003 -$0.41)
$43
($19 -$73)
$21
($9.3 -$31)
a Monetary benefits are rounded to two significant digits for ease of presentation and computation. PM and
ozone benefits are nationwide.
b Monetary benefits adjusted to account for growth in real GDP per capita between 1990 and the analysis
year (2030).
0 Valuation assumes discounting over the SAB recommended 20 year segmented lag structure. Results
reflect the use of 3 percent and 7 percent discount rates consistent with EPA and OMB guidelines for
preparing economic analyses.
dThe negative estimate at the 5th percentile confidence estimate for this morbidity endpoint reflects the
statistical power of the study used to calculate this health impact. This result does not suggest that reducing
air pollution results in additional health impacts.
8-18
-------
Table 8-10: Total Monetized Ozone and PM-related Benefits Associated with the Final
Program in 2030
Total Ozone and PM Benefits (billions, 2011$) -
PM Mortality Derived from the ACS and Six-Cities Studies
3% Discount Rate
Ozone
Mortality
Function
Multi-city
Meta-analysis
Reference
Bell et al.,
2004
Huang et al.,
2005
Schwartz,
2005
Bell et al.,
2005
Ito et al.,
2005
Levy et al.,
2005
Mean Total
Benefits
$7.4 -$15
$7.9 -$16
$8.0 -$16
$9.8 -$18
$11-$19
$11-$19
7% Discount Rate
Ozone
Mortality
Function
Multi-city
Meta-analysis
Reference
Bell et al.,
2004
Huang et al.,
2005
Schwartz,
2005
Bell et al.,
2005
Ito et al.,
2005
Levy etal.,
2005
Mean Total
Benefits
$6.7 -$14
$7.3 -$14
$7.3 -$14
$9.2 -$16
$11-$18
$11-$18
Total Ozone and PM Benefits (billions, 2011$) -
PM Mortality Derived from Expert Elicitation (Lowest and Highest Estimate)
3% Discount Rate
Ozone
Mortality
Function
Multi-city
Meta-analysis
Reference
Bell et al.,
2004
Huang et al.,
2005
Schwartz,
2005
Bell et al.,
2005
Ito et al.,
2005
Levy etal.,
2005
Mean Total
Benefits
$2.4 - $22
$2.9 - $22
$3.0 -$22
$4.9 - $24
$6.3 - $26
$6.3 - $26
7% Discount Rate
Ozone
Mortality
Function
Multi-city
Meta-analysis
Reference
Bell et al.,
2004
Huang et al.,
2005
Schwartz,
2005
Bell et al.,
2005
Ito et al.,
2005
Levy etal.,
2005
Mean Total
Benefits
$2.4 - $20
$2.9 - $20
$3.0 -$20
$4.9 - $22
$6.3 - $24
$6.3 - $26
8.1.2.4 Methodology
We follow a "damage-function" approach in calculating total benefits of the modeled
changes in environmental quality. This approach estimates changes in individual health
endpoints (specific effects that can be associated with changes in air quality) and assigns values
to those changes assuming independence of the values for those individual endpoints. Total
benefits are calculated simply as the sum of the values for all non-overlapping health endpoints.
The "damage-function" approach is the standard method for assessing costs and benefits of
environmental quality programs and has been used in several recent published analyses.38'39'40
To assess economic value in a damage-function framework, the changes in environmental
quality must be translated into effects on people or on the things that people value. In some
cases, the changes in environmental quality can be directly valued. In other cases, such as for
8-19
-------
changes in ozone and PM, an impact analysis must first be conducted to convert air quality
changes into effects that can be assigned dollar values. For the purposes of this RIA, the health
impacts analysis (HIA) includes those health effects that are directly linked to ambient levels of
air pollution and specifically to those linked to ozone and PM2.5.
We note at the outset that the EPA rarely has the time or resources to perform extensive
new research to measure directly either the health outcomes or their values for regulatory
analyses. Thus, similar to Kunzli et al. (2000)41 and other, more recent health impact analyses,
our estimates are based on the best available methods of benefits transfer. Benefits transfer is the
science and art of adapting primary research from similar contexts to obtain the most accurate
measure of benefits for the environmental quality change under analysis. Adjustments are made
for the level of environmental quality change, the socio-demographic and economic
characteristics of the affected population, and other factors to improve the accuracy and
robustness of benefits estimates.
8. 1. 2. 4. 1 Human Health Impact Assessment
The health impact assessment (HIA) quantifies the changes in the incidence of adverse
health impacts resulting from changes in human exposure to PM2.5 and ozone air quality. HIAs
are a well-established approach for estimating the retrospective or prospective change in adverse
health impacts expected to result from population-level changes in exposure to pollutants. 42 PC-
based tools such as the environmental Benefits Mapping and Analysis Program (BenMAP) can
systematize health impact analyses by applying a database of key input parameters, including
health impact functions and population projections — provided that key input data are available,
including air quality estimates and risk coefficients.43 Analysts have applied the HIA approach to
estimate human health impacts resulting from hypothetical changes in pollutant levels.44' 45,46
The EPA and others have relied upon this method to predict future changes in health impacts
expected to result from the implementation of regulations affecting air quality.47 For this
assessment, the HIA is limited to those health effects that are directly linked to ambient ozone
and PM2.5 concentrations.
The HIA approach used in this analysis involves three basic steps: (1) utilizing
projections of PIVb.s air qualityj and estimating the change in the spatial distribution of the
ambient air quality; (2) determining the subsequent change in population-level exposure; (3)
calculating health impacts by applying concentration-response relationships drawn from the
epidemiological literature to this change in population exposure.
A typical health impact function might look like:
where yo is the baseline incidence (the product of the baseline incidence rate times the potentially
affected population), p is the effect estimate, and Ax is the estimated change in the summary
1 Projections of ambient PM25 concentrations for this analysis were generated using the Community Multiscale Air
Quality model (CMAQ). See Chapter 7 of this RIA for more information on the air quality modeling.
8-20
-------
pollutant measure. There are other functional forms, but the basic elements remain the same.
The following subsections describe the sources for each of the first three elements: size of the
potentially affected populations; PM2.s and ozone effect estimates; and baseline incidence rates.
We also describe the treatment of potential thresholds in PM-related health impact functions in
Section 8.1.2.5.3. Section 8.1.2.4.6 describes the ozone and PM air quality inputs to the health
impact functions.
8.1.2.4.2 Potentially Affected Populations
Quantified and monetized human health impacts depend on the demographic
characteristics of the population, including age, location, and income. We use population
projections based on economic forecasting models developed by Woods and Poole, Inc.48 The
Woods and Poole (WP) database contains county-level projections of population by age, sex, and
race out to 2040, relative to a baseline using the 2010 Census data. Projections in each county are
determined simultaneously with every other county in the United States to take into account
patterns of economic growth and migration. The sum of growth in county-level populations is
constrained to equal a previously determined national population growth, based on Bureau of
Census estimates.49 According to WP, linking county-lev el growth projections together and
constraining to a national-level total growth avoids potential errors introduced by forecasting
each county independently. County projections are developed in a four-stage process:
• First, national-level variables such as income, employment, and populations are
forecasted.
Second, employment projections are made for 179 economic areas defined by the
Bureau of Economic Analysis,50 using an "export-base" approach, which relies on
linking industrial-sector production of non-locally consumed production items, such
as outputs from mining, agriculture, and manufacturing with the national economy.
The export-based approach requires estimation of demand equations or calculation of
historical growth rates for output and employment by sector.
• Third, population is projected for each economic area based on net migration rates
derived from employment opportunities and following a cohort-component method
based on fertility and mortality in each area.
• Fourth, employment and population projections are repeated for counties, using the
economic region totals as bounds. The age, sex, and race distributions for each region
or county are determined by aging the population by single year of age by sex and
race for each year through 2040 based on historical rates of mortality, fertility, and
migration.
8.1.2.4.3 Effect Estimate Sources
The first step in selecting effect coefficients is to identify the health endpoints to be
quantified. We base our selection of health endpoints on consistency with the EPA's Integrated
Science Assessments (which replace previous Criteria Documents), with input and advice from
the SAB-HES, a scientific review panel specifically established to provide advice on the use of
the scientific literature in developing benefits analyses for the EPA's Report to Congress on The
Benefits and Costs of the Clean Air Act 1990 to 2020.51 In addition, we have included more
8-21
-------
CO CQ
recent epidemiology studies from the PM ISA and the Provisional Assessment. ' In general,
we follow a weight of evidence approach, based on the biological plausibility of effects,
availability of concentration-response functions from well conducted peer-reviewed
epidemiological studies, cohesiveness of results across studies, and a focus on endpoints
reflecting public health impacts (like hospital admissions) rather than physiological responses
(such as changes in clinical measures like Forced Expiratory Volume [FEV1]).
There are several types of data that can support the determination of types and magnitude
of health effects associated with air pollution exposures. These sources of data include
toxicological studies (including animal and cellular studies), human clinical trials, and
observational epidemiology studies. All of these data sources provide important contributions to
the weight of evidence surrounding a particular health impact. However, only epidemiology
studies provide direct concentration-response relationships that can be used to evaluate
population-level impacts of reductions in ambient pollution levels in a health impact assessment.
For the data-derived estimates, we relied on the published scientific literature to ascertain
the relationship between PM2.5;, ozone, and adverse human health effects. We evaluated
epidemiological studies using the selection criteria summarized in Table 8-11. These criteria
include consideration of whether the study was peer-reviewed, the match between the pollutant
studied and the pollutant of interest, the study design and location, and characteristics of the
study population, among other considerations. In general, the use of concentration-response
functions from more than a single study can provide a more representative distribution of the
effect estimate. However, there are often differences between studies examining the same
endpoint, making it difficult to pool the results in a consistent manner. For example, studies may
examine different pollutants or different age groups. For this reason, we consider very carefully
the set of studies available examining each endpoint and select a consistent subset that provides a
good balance of population coverage and match with the pollutant of interest. In many cases,
either because of a lack of multiple studies, consistency problems, or clear superiority in the
quality or comprehensiveness of one study over others, a single published study is selected as the
basis of the effect estimate.
When several effect estimates for a pollutant and a given health endpoint have been
selected, they are quantitatively combined or pooled to derive a more robust estimate of the
relationship. The BenMAP Manual Technical Appendices provides details of the procedures
used to combine multiple impact functions.54 In general, we used fixed or random effects models
to pool estimates from different single city studies of the same endpoint. Fixed effects pooling
simply weights each study's estimate by the inverse variance, giving more weight to studies with
greater statistical power (lower variance). Random effects pooling accounts for both within-study
variance and between-study variability, due, for example, to differences in population
susceptibility. We used the fixed effects model as our null hypothesis and then determined
whether the data suggest that we should reject this null hypothesis, in which case we would use
the random effects model. K Pooled impact functions are used to estimate hospital admissions
K EPA recently changed the algorithm BenMAP uses to calculate study variance, which is used in the pooling
process. Prior versions of the model calculated population variance, while the version used here calculated sample
8-22
-------
and asthma exacerbations. When combining evidence across multi-city studies (e.g.,
cardiovascular hospital admission studies), we use equal weights pooling. The effect estimates
drawn from each multi-city study are themselves pooled across a large number of urban areas.
For this reason, we elected to give each study an equal weight rather than weighting by the
inverse of the variance reported in each study. For more details on methods used to pool
incidence estimates, see the BenMAP Manual Appendices.
Effect estimates selected for a given health endpoint were applied consistently across all
locations nationwide. This applies to both impact functions defined by a single effect estimate
and those defined by a pooling of multiple effect estimates. Although the effect estimate may, in
fact, vary from one location to another (e.g., because of differences in population susceptibilities
or differences in the composition of PM), location-specific effect estimates are generally not
available.
Table 8-11: Criteria Used When Selecting C-R Functions
Consideration
Peer-Reviewed
Research
Study Type
Study Period
Population Attributes
Study Size
Study Location
Pollutants Included in
Model
Comments
Peer-reviewed research is preferred to research that has not undergone the peer-review
process.
Among studies that consider chronic exposure (e.g., over a year or longer), prospective
cohort studies are preferred over ecological studies because they control for important
individual -level confounding variables that cannot be controlled for in ecological studies.
Studies examining a relatively longer period of time (and therefore having more data) are
preferred, because they have greater statistical power to detect effects. Studies that are
more recent are also preferred because of possible changes in pollution mixes, medical
care, and lifestyle over time. However, when there are only a few studies available, studies
from all years will be included.
The most technically appropriate measures of benefits would be based on impact functions
that cover the entire sensitive population but allow for heterogeneity across age or other
relevant demographic factors. In the absence of effect estimates specific to age, sex,
preexisting condition status, or other relevant factors, it may be appropriate to select effect
estimates that cover the broadest population to match with the desired outcome of the
analysis, which is total national-level health impacts. When available, multi-city studies
are preferred to single city studies because they provide a more generalizable
representation of the concentration-response function.
Studies examining a relatively large sample are preferred because they generally have
more power to detect small magnitude effects. A large sample can be obtained in several
ways, including through a large population or through repeated observations on a smaller
population (e.g., through a symptom diary recorded for a panel of asthmatic children).
U.S. studies are more desirable than non-U.S. studies because of potential differences in
pollution characteristics, exposure patterns, medical care system, population behavior, and
lifestyle. National estimates are most appropriate when benefits are nationally distributed;
the impact of regional differences may be important when benefits only accrue to a single
area.
When modeling the effects of ozone and PM (or other pollutant combinations) jointly, it is
important to use properly specified impact functions that include both pollutants. Using
single-pollutant models in cases where both pollutants are expected to affect a health
variance. This change did not affect the selection of random or fixed effects for the pooled incidence estimates
between the proposal and final RIA.
8-23
-------
Measure of PM
Economically Valuable
Health Effects
Non-overlapping
Endpoints
outcome can lead to double-counting when pollutants are correlated.
For this analysis, impact functions based on PM25 are preferred to PM10 because of the
focus on reducing emissions of PM25 precursors, and because air quality modeling was
conducted for this size fraction of PM. Where PM2 5 functions are not available, PM10
functions are used as surrogates, recognizing that there will be potential downward
(upward) biases if the fine fraction of PM10 is more (less) toxic than the coarse fraction.
Some health effects, such as forced expiratory volume and other technical measurements
of lung function, are difficult to value in monetary terms. These health effects are not
quantified in this analysis.
Although the benefits associated with each individual health endpoint may be analyzed
separately, care must be exercised in selecting health endpoints to include in the overall
benefits analysis because of the possibility of double -counting of benefits.
It is important to note that we are unable to separately quantify all of the possible PM and
ozone health effects that have been reported in the literature for three reasons: (1) the possibility
of double counting (such as hospital admissions for specific respiratory diseases versus hospital
admissions for all or a sub-set of respiratory diseases); (2) uncertainties in applying effect
relationships that are based on clinical studies to the potentially affected population; or (3) the
lack of an established concentration-response (CR) relationship. Table 8-12 lists the health
endpoints included in this analysis.
Table 8-12: Health Impact Functions Used in BenMAP to Estimate Impacts of PM2.s and
Ozone Reductions
ENDPOINT
POLLUTANT
STUDY
STUDY POPULATION
Premature Mortality
Premature mortality -
daily time series
Premature mortality
— cohort study, all-
cause
Premature mortality,
total exposures
Premature mortality
— all-cause
03
PM25
PM25
PM25
Multi-city
Bell et al (2004) (NMMAPS study)55 - Non-
accidental
CO
Huang et al (2005) - Cardiopulmonary
Schwartz (2005)57 - Non-accidental
Meta-analyses:
Bell et al (2005)58 - All cause
Ito et al (2005)59 - Non-accidental
Levy et al (2005)60 - All cause
Krewski et al. (2009)61
Lepeuleetal. (2012)62
Expert Elicitation (lEc, 2006)63
Woodruff etal. (1997)64
All ages
>29 years
>25 years
>24 years
Infant (<1 year)
Chronic Illness
8-24
-------
ENDPOINT
Nonfatal heart attacks
POLLUTANT
PM25
STUDY
Peters etal. (200 1)65
Pooled estimate:
Pope et al. (2006)66
Sullivan et al. (2005)67
Zanobettietal. (2009)68
Zanobetti and Schwartz (2006)69
STUDY POPULATION
Adults (>18 years)
Hospital Admissions
Respiratory
Cardiovascular
Asthma-related ER
visits
Asthma-related ER
visits (cont'd)
03
PM25
PM25
PM25
PM25
PM25
03
PM25
Pooled estimate:
Schwartz (1995) - ICD 460-519 (all resp)70
Schwartz (1994a; 1994b) - ICD 480-486
(pneumonia)71'72
Moolgavkar et al. (1997) - ICD 480-487
(pneumonia)73
Schwartz (1994b) - ICD 491-492, 494-496
(COPD)
Moolgavkar et al. (1997) - ICD 490-496
(COPD)
Burnett etal. (200 1)74
Pooled estimate:
Zanobetti etal. (2009)— ICD 460-519 (All
respiratory)
Kloog et al. (2012)— ICD 460-519 (All
Respiratory)75
Moolgavkar (2000)— ICD 490-496 (Chronic
lung disease)76
Pooled estimate:
Babin et al. (2007)— ICD 493 (asthma)77
Sheppard (2003)— ICD 493 (asthma)78
Pooled estimate: Zanobetti et al. (2009)— ICD
390-459 (all cardiovascular)
Peng et al. (2009)— ICD 426-427; 428; 430-
438; 410-414; 429; 440-449 (Cardio-, cerebro-
and peripheral vascular disease)79
Peng et al. (2008)— ICD 426-427; 428; 430-
438; 410-414; 429; 440-449 (Cardio-, cerebro-
and peripheral vascular disease)80
Bell et al. (2008)— ICD 426-427; 428; 430-
438; 410-414; 429; 440-449 (Cardio-, cerebro-
and peripheral vascular disease)81
Moolgavkar (2000)— ICD 390-429 (all
cardiovascular)
Pooled estimate:
Peel et al (2005)82
Wilson etal(2005)83
Pooled estimate:
Mar etal. (2010)84
Slaughter et al. (2005)85
Glad etal. (2012)86
>64 years
<2 years
>64 years
18-64 years
<18 years
>64 years
20-64 years
All ages
All ages
All ages
8-25
-------
ENDPOINT
POLLUTANT
STUDY
STUDY POPULATION
Other Health Endpoints
Acute bronchitis
Upper respiratory
symptoms
Lower respiratory
symptoms
Asthma exacerbations
Work loss days
School absence days
Minor Restricted
Activity Days
(MRADs)
PM25
PM25
PM25
PM25
PM25
03
03
PM25
Dockeryetal. (1996)87
Popeetal. (199 1)88
Schwartz and Neas (2000)89
Pooled estimate:
Ostro et al. (2001)90 (cough, wheeze and
shortness of breath)
Mar et al. (2004) (cough, shortness of breath)
Ostro (1987)91
Pooled estimate:
Gilliland et al. (2001)92
Chenetal. (2000)93
Ostro and Rothschild (1989)94
Ostro and Rothschild (1989)
8-12 years
Asthmatics, 9-11
years
7-14 years
6-18 years3
18-65 years
5-17 yearsb
18-65 years
18-65 years
Notes:
" The original study populations were 8 to 13 for the Ostro et al. (2001) study and 7 to 12 for the Mar et al. (2004)
study. Based on advice from the SAB-HES, we extended the applied population to 6-18, reflecting the common
biological basis for the effect in children in the broader age group. See: U.S. EPA-SAB (2004) and NRC (2002).
* Gilliland et al. (2001) studied children aged 9 and 10. Chen et al. (2000) studied children 6 to 11. Based on advice
from the National Research Council and the EPA SAB-HES, we have calculated reductions in school absences for
all school-aged children based on the biological similarity between children aged 5 to 17.
For detailed descriptions of each of the individual studies referenced in Table 8-12,
please refer to the RIAs that accompanied the final revisions to the National Ambient Air Quality
Standards (NAAQS) for Particulate Matter, the 2008 final ozone NAAQS, and the 2010 ozone
NAAQS reconsideration.95'96'97
8.1.2.4.4
Baseline Incidence Rates
Epidemiological studies of the association between pollution levels and adverse health
effects generally provide a direct estimate of the relationship of air quality changes to the relative
risk of a health effect, rather than estimating the absolute number of avoided cases. For example,
a typical result might be that a 10 |ig/m3 decrease in daily PM2.5 levels might be associated with
a decrease in hospital admissions of 3%. The baseline incidence of the health effect is necessary
to convert this relative change into a number of cases. A baseline incidence rate is the estimate of
the number of cases of the health effect per year in the assessment location, as it corresponds to
baseline pollutant levels in that location. To derive the total baseline incidence per year, this rate
must be multiplied by the corresponding population number. For example, if the baseline
incidence rate is the number of cases per year per million people, that number must be multiplied
by the millions of people in the total population.
Table 8-13 summarizes the sources of baseline incidence rates and provides average
incidence rates for the endpoints included in the analysis. For both baseline incidence and
prevalence data, we used age-specific rates where available. We applied concentration-response
functions to individual age groups and then summed over the relevant age range to provide an
8-26
-------
estimate of total population benefits. In most cases, we used a single national incidence rate, due
to a lack of more spatially disaggregated data. Whenever possible, the national rates used are
national averages, because these data are most applicable to a national assessment of benefits.
For some studies, however, the only available incidence information comes from the studies
themselves; in these cases, incidence in the study population is assumed to represent typical
incidence at the national level. County, state and regional incidence rates are available for
hospital admissions, and county-level data are available for premature mortality.
We projected mortality rates such that future mortality rates are consistent with our
projections of population growth.98 To perform this calculation, we began first with an average
of 2004-2006 cause-specific mortality rates. Using Census Bureau projected national-level
annual mortality rates stratified by age range, we projected these mortality rates to 2050 in 5-year
increments. 99'100
The baseline incidence rates for hospital admissions and emergency department visits
reflect the revised rates first applied in the CSAPR RIA.101 In addition, we have revised the
baseline incidence rates for acute myocardial infarction. These revised rates are more recent,
which provides a better representation of the rates at which populations of different ages, and in
different locations, visit the hospital and emergency department for air pollution-related
illnesses. Also, the new baseline incidence rates are more spatially refined. For many locations
within the U.S., these data are resolved at the county- or state-level, providing a better
characterization of the geographic distribution of hospital and emergency department visits than
the previous national rates. Lastly, these rates reflect unscheduled hospital admissions only,
which represent a conservative assumption that most air pollution-related visits are likely to be
unscheduled. If air pollution-related hospital admissions are scheduled, this assumption would
underestimate these benefits.
For the set of endpoints affecting the asthmatic population, in addition to baseline
incidence rates, prevalence rates of asthma in the population are needed to define the applicable
population. Table 8-14 lists the prevalence rates used to determine the applicable population for
asthma symptoms. Note that these reflect current asthma prevalence and assume no change in
prevalence rates in future years. We updated these rates in the CSAPR RIA.
Table 8-13: Baseline Incidence Rates and Population Prevalence Rates for Use in Impact
Functions, General Population
Endpoint
Mortality
Hospitalizations
Asthma ER Visits
Nonfatal Myocardial
Parameter
Daily or annual mortality
rate projected to 2020
Daily hospitalization rate
Daily asthma ER visit rate
Daily nonfatal myocardial
Rates
Value
Age-, cause-, and
county-specific rate
Age-, region-, state-,
county- and cause-
specific rate
Age-, region-, state-,
county- and cause-
specific rate
Age-, region-, state-,
Source
CDC Wonder (2006-2008)102
U.S. Census bureau
2007 HCUP data files3'103
2007 HCUP data files3
2007 HCUP data files;3 adjusted by
8-27
-------
Endpoint
Infarction (heart
attacks)
Asthma Exacerbations
Acute Bronchitis
Lower Respiratory
Symptoms
Upper Respiratory
Symptoms
Work Loss Days
School Loss Days
Minor Restricted-
Activity Days
Parameter
infarction incidence rate
per person, 18+
Incidence among asthmatic
African- American children
daily wheeze
daily cough
daily shortness of breath
Annual bronchitis
incidence rate, children
Daily lower respiratory
symptom incidence among
children15
Daily upper respiratory
symptom incidence among
asthmatic children
Daily WLD incidence rate
per person (18-65)
Aged 18-24
Aged 25-44
Aged 45-64
Rate per person per year,
assuming 180 school days
per year
Daily MRAD incidence
rate per person
Rates
Value
and county- specific
rate
0.173
0.145
0.074
0.043
0.0012
0.3419
0.00540
0.00678
0.00492
9.9
0.02137
Source
0.93 for probability of surviving after
28 days (Rosamond et al., 1999)104
Ostroetal. (2001)
American Lung Association (2002,
Table II)105
Schwartz et al. (1994, Table 2)
Pope etal. (1991, Table 2)
1996 HIS (Adams, Hendershot, and
Marano, 1999, Table 41);106 U.S.
Bureau of the Census (2000)107
National Center for Education
Statistics (1996)108 and 1996 HIS
(Adams et al., 1999, Table 47);
Ostro and Rothschild (1989, p. 243)
Notes:
" Healthcare Cost and Utilization Program (HCUP) database contains individual
hospital and emergency department discharges for a variety of ICD codes.
* Lower respiratory symptoms are defined as two or more of the following: couj
level, state and regional-level
jh, chest pain, phlegm, and wheeze.
Table 8-14: Asthma Prevalence Rates Used for this Analysis
Population Group
All Ages
<18
5-17
18-44
45-64
65+
African American, 5 to 17
African American, <18
Asthma Prevalence Rates
Value
0.0780
0.0941
0.1070
0.0719
0.0745
0.0716
0.1776
0.1553
Source
American Lung Association (2010, Table 7)109
American Lung Association (2010, Table 9)
American Lung Association3
Notes:
ab Calculated by ALA for U.S. EPA, based on NHIS data (CDC, 2008).110
8.1.2.4.5
Economic Values for Health Outcomes
After quantifying the change in adverse health impacts, we estimate the economic value
of these avoided impacts. The appropriate economic value for a change in a health effect depends
8-28
-------
on whether the health effect is viewed ex ante (before the effect has occurred) or ex post (after
the effect has occurred). Reductions in ambient concentrations of air pollution generally lower
the risk of future adverse health effects by a small amount for a large population. The appropriate
economic measure is therefore ex ante willingness to pay (WTP) for changes in risk. m
Epidemiological studies generally provide estimates of the relative risks of a particular health
effect for a given increment of air pollution (often per 10 |ig/m3 for PM^.s). These relative risks
can be used to develop risk coefficients that relate a unit reduction in PM2.5 to changes in the
incidence of a health effect. In order to value these changes in incidence, WTP for changes in
risk need to be converted into WTP per statistical incidence. This measure is calculated by
dividing individual WTP for a risk reduction by the related observed change in risk. For
example, suppose a measure is able to reduce the risk of premature mortality from 2 in 10,000 to
1 in 10,000 (a reduction of 1 in 10,000). If individual WTP for this risk reduction is $100, then
the WTP for an avoided statistical premature mortality amounts to $1 million ($100/0.0001
change in risk). Using this approach, the size of the affected population is automatically taken
into account by the number of incidences predicted by epidemiological studies applied to the
relevant population. The same type of calculation can produce values for statistical incidences of
other health endpoints.
For some health effects, such as hospital admissions, WTP estimates are generally not
available. In these cases, we use the cost of treating or mitigating the effect. For example, for the
valuation of hospital admissions we use the avoided medical costs as an estimate of the value of
avoiding the health effects causing the admission. These cost-of-illness (COI) estimates
generally (although not necessarily in every case) understate the true value of reductions in risk
of a health effect. They tend to reflect the direct expenditures related to treatment but not the
value of avoided pain and suffering from the health effect.112'113
We provide unit values for health endpoints (along with information on the distribution
of the unit value) in Table 8-15. All values are in constant year 2011 dollars, adjusted for growth
in real income out to 2030 using projections provided by Standard and Poor's. Economic theory
argues that WTP for most goods (such as environmental protection) will increase if real income
increases. Many of the valuation studies used in this analysis were conducted in the late 1980s
and early 1990s. Because real income has grown since the studies were conducted, people's
willingness to pay for reductions in the risk of premature death and disease likely has grown as
well. We did not adjust cost of illness-based values because they are based on current costs.
Similarly, we did not adjust the value of school absences, because that value is based on current
wage rates. For details on valuation estimates for PM-related endpoints, see the the RIA that
accompanied the final revisions to the National Ambient Air Quality Standards (NAAQS) for
Particulate Matter.114 For details on valuation estimates for ozone-related endpoints, see the
RIAs for the 2008 Ozone NAAQS RIA and the 2010 ozone NAAQS reconsideration.115'116
8-29
-------
Table 8-15: Unit Values for Economic Valuation of Health Endpoints (2011$)
Health Endpoint
Premature Mortality
(Value of a Statistical
Life)
Nonfatal Myocardial
Infarction (heart
attack)
3% discount rate
Age 0-24
Age 25-44
Age 45-54
Age 55-65
Age 66 and over
7% discount rate
A f\ /•*
Age 0-24
Age 25-44
Age 45-54
Age 55-65
Age 66 and over
Hospital Admissions
Chronic Obstructive
Pulmonary Disease
(COPD)
Asthma Admissions
All Cardiovascular
Age 18-64
Age 65-99
All respiratory (ages
65+)
Central Estimate of
Value Per Statistical
Incidence
1990
Income
Level
$8,300,000
$100,000
$110,000
$120,000
$210,000
$100,000
$100,000
$110,000
$110,000
$200,000
$100,000
$22,000
$17,000
$43,000
$42,000
$37,000
2030
Income
Level3
$10,200,000
$100,000
$110,000
$120,000
$210,000
$100,000
$100,000
$110,000
$110,000
$200,000
$100,000
$22,000
$17,000
$43,000
$42,000
$37,000
Derivation of Distributions of Estimates
EPA currently recommends a central VSL of $6. 3m (2000$, 1990
income) based on a Weibull distribution fitted to 26 published VSL
estimates (5 contingent valuation and 21 labor market studies). The
underlying studies, the distribution parameters, and other useful
information are available in Appendix B of EPA's current Guidelines
for Preparing Economic Analyses (U.S. EPA, 2010).117
No distributional information available. Age-specific cost-of-illness
values reflect lost earnings and direct medical costs over a 5 -year
period following a nonfatal MI. Lost earnings estimates are based on
Cropper and Krupnick (1990). 118 Direct medical costs are based on
simple average of estimates from Russell et al. (1998)119 and Wittels et
al. (1990).120
Lost earnings:
Cropper and Krupnick (1990). Present discounted value of 5 years of
lost earnings:
age of onset: at 3% at 7%
25-44 $9,000 $8,000
45-54 $13,000 $12,000
55-65 $77,000 $69,000
Direct medical expenses: An average of:
1. Wittels et al. (1990) ($100,000— no discounting)
2. Russell et al. (1998), 5-year period ($22,000 at 3% discount rate;
$21,000 at 7% discount rate)
No distributional information available. The COI estimates (lost
earnings plus direct medical costs) are based on ICD-9 code-level
information (e.g. , average hospital care costs, average length of
hospital stay, and weighted share of total COPD category illnesses)
reported in Agency for Healthcare Research and Quality (2007)
(www.ahrq.gov).
No distributional information available. The COI estimates (lost
earnings plus direct medical costs) are based on ICD-9 code-level
information (e.g. , average hospital care costs, average length of
hospital stay, and weighted share of total asthma category illnesses)
reported in Agency for Healthcare Research and Quality (2007)
(www.ahrq.gov).
No distributional information available. The COI estimates (lost
earnings plus direct medical costs) are based on ICD-9 code-level
information (e.g. , average hospital care costs, average length of
hospital stay, and weighted share of total cardiovascular category
illnesses) reported in Agency for Healthcare Research and Quality
(2007) (www.ahrq.gov).
No distributions available. The COI point estimates (lost earnings plus
direct medical costs) are based on ICD-9 code level information (e.g. ,
average hospital care costs, average length of hospital stay, and
weighted share of total COPD category illnesses) reported in Agency
8-30
-------
for Healthcare Research and Quality, 2007 (www.ahrq.gov).
All respiratory (ages
0-2)
$13,000
$13,000
No distributions available. The COI point estimates (lost earnings plus
direct medical costs) are based on ICD-9 code level information (e.g.,
average hospital care costs, average length of hospital stay, and
weighted share of total COPD category illnesses) reported in Agency
for Healthcare Research and Quality, 2007 (www.ahrq.gov).
Emergency Room
Visits for Asthma
$440
$440
No distributional information available. Simple average of two unit
COI values:
(1) $311.55, from Smith etal. (1997)122 and
(2) $260.67, from Stanford et al. (1999).123
Respiratory Ailments Not Requiring Hospitalization
Upper Respiratory
Symptoms (URS)
$32
Combinations of the three symptoms for which WTP estimates are
available that closely match those listed by Pope et al. result in seven
different "symptom clusters," each describing a "type" of URS. A
dollar value was derived for each type of URS, using mid-range
estimates of WTP (Neumann et al., 1994)124 to avoid each symptom in
the cluster and assuming additivity of WTPs. In the absence of
information surrounding the frequency with which each of the seven
types of URS occurs within the URS symptom complex, we assumed a
uniform distribution between $9.2 and $43 (2000$).
Lower Respiratory
Symptoms (LRS)
$21
$22
Combinations of the four symptoms for which WTP estimates are
available that closely match those listed by Schwartz et al. result in 11
different "symptom clusters," each describing a "type" of LRS. A
dollar value was derived for each type of LRS, using mid-range
estimates of WTP (Neumann et al., 1994) to avoid each symptom in
the cluster and assuming additivity of WTPs. The dollar value for LRS
is the average of the dollar values for the 11 different types of LRS. In
the absence of information surrounding the frequency with which each
of the 11 types of LRS occurs within the LRS symptom complex, we
assumed a uniform distribution between $6.9 and $25 (2000$).
Asthma
Exacerbations
$57
$59
Asthma exacerbations are valued at $45 per incidence, based on the
mean of average WTP estimates for the four severity definitions of a
"bad asthma day," described inRowe and Chestnut (1986).125 This
study surveyed asthmatics to estimate WTP for avoidance of a "bad
asthma day," as defined by the subjects. For purposes of valuation, an
asthma exacerbation is assumed to be equivalent to a day in which
asthma is moderate or worse as reported in the Rowe and Chestnut
(1986) study. The value is assumed have a uniform distribution
between $16 and $71 (2000$).
Acute Bronchitis
$470
$510
Assumes a 6-day episode, with the distribution of the daily value
specified as uniform with the low and high values based on those
recommended for related respiratory symptoms in Neumann et al.
(1994). The low daily estimate of $10 is the sum of the mid-range
values recommended by lEc 1994 for two symptoms believed to be
associated with acute bronchitis: coughing and chest tightness. The
high daily estimate was taken to be twice the value of a minor
respiratory restricted-activity day, or $110 (2000$).
Work Loss Days
(WLDs)
Variable
(U.S.
median =
$150)
Variable
(U.S.
median =
$150)
No distribution available. Point estimate is based on county-specific
median annual wages divided by 50 (assuming 2 weeks of vacation)
and then by 5—to get median daily wage. U.S. Year 2000 Census,
compiled by Geolytics, Inc. (Geolytics, 2002)126
8-31
-------
Minor Restricted
Activity Days
(MRADs)
School Absence Days
$66
$98
$71
$98
Median WTP estimate to avoid one MRAD from Tolley et al.
(1986). 127 Distribution is assumed to be triangular with a minimum of
$22 and a maximum of $83, with a most likely value of $52 (2000$).
Range is based on assumption that value should exceed WTP for a
single mild symptom (the highest estimate for a single symptom — for
eye irritation — is $16.00) and be less than that for a WLD. The
triangular distribution acknowledges that the actual value is likely to be
closer to the point estimate than either extreme.
No distribution available
Note:
a Willingness-to-pay-based benefits are adjusted by multiplying the unadjusted benefits by the appropriate
adjustment factor to account for income growth over time using projections provided by Standard and Poor's. Due
to a lack of reliable projections of income growth past 2024, we assume constant WTP from 2024 through 2030.
This results in an underestimate of benefits occurring between 2024 and 2030. For growth between 2000 and 2024,
this factor is 1.23 for long-term mortality, 1.27 for chronic health impacts, and 1.08 for minor health impacts. Note
that similar adjustments do not exist for cost-of-illness-based unit values. For these, we apply the same unit value
regardless of the future year of analysis.
8.1.2.4.6 Manipulating Air Quality Modeling Data for Health
Impacts Analysis
In Chapter 7, we summarized the methods for and results of estimating air quality for the
program. These air quality results are in turn associated with human populations to estimate
changes in health effects. For the purposes of this analysis, we focus on the health effects that
have been linked to ambient changes in ozone and PM2.5 related to emission reductions estimated
to occur due to the implementation of the program. We estimate ambient PM2.5 and ozone
concentrations using the Community Multiscale Air Quality model (CMAQ). This section
describes how we converted the CMAQ modeling output into full-season profiles suitable for the
health impacts analysis.
General Methodology
First, we extracted hourly, surface-layer PM and ozone concentrations for each grid cell
from the standard CMAQ output files. For ozone, these model predictions are used in
conjunction with the observed concentrations obtained from the Aerometric Information
Retrieval System (AIRS) to generate ozone concentrations for the entire ozone season.L'M The
predicted changes in ozone concentrations from the future-year base case to future-year control
scenario serve as inputs to the health and welfare impact functions of the benefits analysis (i.e.,
BenMAP).
To estimate ozone-related health effects for the contiguous United States, full-season
ozone data are required for every BenMAP grid-cell. Given available ozone monitoring data, we
generated full-season ozone profiles for each location in two steps: (1) we combined monitored
L The ozone season for this analysis is defined as the 5-month period from May to September.
M Based on AIRS, there were 961 ozone monitors with sufficient data (i.e., 50 percent or more days reporting at
least nine hourly observations per day [8 am to 8 pm] during the ozone season).
8-32
-------
observations and modeled ozone predictions to interpolate hourly ozone concentrations to a grid
of 12-km by 12-km population grid cells for the contiguous 48 states, and (2) we converted these
full-season hourly ozone profiles to an ozone measure of interest, such as the daily 8-hour
maximum.N'°
For PM2.5, we also use the model predictions in conjunction with observed monitor data.
CMAQ generates predictions of hourly PM species concentrations for every grid. The species
include a primary coarse fraction (corresponding to PM in the 2.5 to 10 micron size range), a
primary fine fraction (corresponding to PM less than 2.5 microns in diameter), and several
secondary particles (e.g., sulfates, nitrates, and organics). PM2.5 is calculated as the sum of the
primary fine fraction and all of the secondarily formed particles. Future-year estimates of PM2.5
were calculated using relative reduction factors (RRFs) applied to 2005 ambient PM2.s and PM2.s
species concentrations. A gridded field of PM2.5 concentrations was created by interpolating
Federal Reference Monitor ambient data and IMPROVE ambient data. Gridded fields of PM2.5
species concentrations were created by interpolating EPA speciation network (ESPN) ambient
data and IMPROVE data. The ambient data were interpolated to the CMAQ 12 km grid.
The procedures for determining the RRFs are similar to those in EPA's draft guidance for
modeling the PM2.5 standard (EPA, 2001).128 The guidance recommends that model predictions
be used in a relative sense to estimate changes expected to occur in each major PM2.5 species.
The procedure for calculating future-year PM2.s design values is called the "Speciated Modeled
Attainment Test (SMAT)." EPA used this procedure to estimate the ambient impacts of the final
program.
Table 8-16 provides those ozone and PM2.5 metrics for grid cells in the modeled domain
that enter the health impact functions for health benefits endpoints. The population-weighted
average reflects the baseline levels and predicted changes for more populated areas of the nation.
This measure better reflects the potential benefits through exposure changes to these populations.
Table 8-16: Summary of CMAQ-Derived Population-Weighted Ozone and PMi.s Air
Quality Metrics for Health Benefits Endpoints Associated with the Final Tier 3 Program
Statistic51
2030
Baseline
Change6
Ozone Metric: National Population-Weighted Average (ppb)°
Daily Maximum 8 -Hour Average
Concentration
43.65
0.32
PM2.s Metric: National Population-Weighted Average (|ig/mj)
Annual Average Concentration
7.94
0.04
Notes:
The 12-km grid squares contain the population data used in the health benefits analysis model, BenMAP.
0 This approach is a generalization of planar interpolation that is technically referred to as enhanced Voronoi
Neighbor Averaging (EVNA) spatial interpolation. See the BenMAP manual for technical details, available for
download at http://www.epa.gov/auYbenmap.
8-33
-------
" Ozone and PM2 5 metrics are calculated at the CMAQ grid-cell level for use in health effects
estimates. Ozone metrics are calculated over relevant time periods during the daylight hours of
the "ozone season" (i.e., May through September). Note that the national, population-weighted
PM2 5 and ozone air quality metrics presented in this chapter represent an average for the entire,
gridded U.S. CMAQ domain. These are different than the population-weighted PM2 5 and ozone
design value metrics presented in Chapter 7, which represent the average for areas with a current
air quality monitor.
* The change is defined as the base-case value minus the control-case value.
0 Calculated by summing the product of the projected CMAQ grid-cell population and the
estimated CMAQ grid cell seasonal ozone concentration and then dividing by the total population.
Emissions and air quality modeling decisions are made early in the analytical process.
For this reason, the emission control scenarios used in the air quality and benefits modeling are
slightly different than the final emission inventories estimated for the Tier 3 standards. Please
refer to Section 7.2.1.1 for more information about the inventories used in the air quality
modeling that supports the health impacts analysis.
8.1.2.5 Methods for Describing Uncertainty
In any complex analysis using estimated parameters and inputs from numerous models,
there are likely to be many sources of uncertainty. This analysis is no exception. As outlined both
in this and preceding chapters, this analysis includes many data sources as inputs, including
emission inventories, air quality data from models (with their associated parameters and inputs),
population data, population estimates, health effect estimates from epidemiology studies,
economic data for monetizing benefits, and assumptions regarding the future state of the world
(i.e., regulations, technology, and human behavior). Each of these inputs may be uncertain and
would affect the benefits estimate. When the uncertainties from each stage of the analysis are
compounded, even small uncertainties can have large effects on the total quantified benefits.
After reviewing the EPA's approach, the National Research Council (NRC) (2002, 2008),
129,130 w^c^ js part Of ^e National Academies of Science, concluded that the EPA's general
methodology for calculating the benefits of reducing air pollution is reasonable and informative
in spite of inherent uncertainties. The NRC also highlighted the need to conduct rigorous
quantitative analyses of uncertainty and to present benefits estimates to decision makers in ways
that foster an appropriate appreciation of their inherent uncertainty. Since the publication of these
reports, the EPA has continued work to improve the characterization of uncertainty in both
health incidence and benefits estimates. In response to these recommendations, we have
expanded our previous analyses to incorporate additional quantitative and qualitative
characterizations of uncertainty. Although we have not yet been able to make as much progress
towards a full, probabilistic uncertainty assessment as envisioned by the NAS as we had hoped,
we have added a number of additional quantitative and qualitative analyses to highlight the
impact that uncertain assumptions may have on the benefits estimates. In addition, for some
inputs into the benefits analysis, such as the air quality data, it is difficult to address uncertainty
probabilistically due to the complexity of the underlying air quality models and emission inputs.
Therefore, we decline to make up alternative assumptions simply for the purpose of probabilistic
uncertainty characterization when there is no scientific literature to support alternate
assumptions.
8-34
-------
To characterize uncertainty and variability, the 2012 PM NAAQS RIA utilized an
approach that combined elements from two recent analyses by the EPA,131'132 and used a tiered
approach developed by the World Health Organization (WHO) for characterizing uncertainty.133
We refer the reader to this tiered assessment as an assessment of the potential impact and
magnitude of each aspect of uncertainty that is also present in the final Tier 3 RIA (see Appendix
5c of the 2012 PM NAAQS RIA).
Data limitations prevent us from treating each source of uncertainty quantitatively and
from reaching a full-probabilistic simulation of our results, but we were able to consider the
influence of uncertainty in the risk coefficients and economic valuation functions by
incorporating four quantitative analyses described in more detail below:
1. A Monte Carlo assessment that accounts for random sampling error and between
study variability in the epidemiological and economic valuation studies;
2. The quantification of PM-related mortality using alternative PM2.5 mortality effect
estimates drawn from two long-term cohort studies and an expert elicitation;
3. A concentration benchmark assessment that characterizes the distribution of
avoided PM2.s-related deaths relative to specific concentrations in the long-term epidemiological
studies used to estimate PM2.s-related mortality;
4. The quantification of ozone-related mortality using alternative ozone mortality
effect estimates drawn from six short-term studies: three multi-city studies and three meta-
analyses of the existing literature;
5. An analysis of the influence of various parameters on total monetized benefits.
8.1.2.5.1 Monte Carlo Assessment
Similar to other recent RIAs, we used Monte Carlo methods for characterizing random
sampling error associated with the concentration response functions from epidemiological
studies and random effects modeling to characterize both sampling error and variability across
the economic valuation functions. The Monte Carlo simulation in the BenMAP software
randomly samples from a distribution of incidence and valuation estimates to characterize the
effects of uncertainty on output variables. Specifically, we used Monte Carlo methods to
generate confidence intervals around the estimated health impact and monetized benefits. The
reported standard errors in the epidemiological studies determined the distributions for individual
effect estimates for endpoints estimated using a single study. For endpoints estimated using a
pooled estimate of multiple studies, the confidence intervals reflect both the standard errors and
the variance across studies. The confidence intervals around the monetized benefits incorporate
the epidemiology standard errors as well as the distribution of the valuation function. These
confidence intervals do not reflect other sources of uncertainty inherent within the estimates,
such as baseline incidence rates, populations exposed and transferability of the effect estimate to
diverse locations. As a result, the reported confidence intervals and range of estimates give an
incomplete picture about the overall uncertainty in the benefits estimates.
8-35
-------
8.1.2.5.2 Alternative Concentration-Response Functions for PM2.5-
Related Mortality
We assign the greatest economic value to the reduction in PM2.s related mortality risk.
Therefore, it is particularly important to attempt to characterize the uncertainties associated with
reductions in premature mortality. To better understand the concentration-response relationship
between PM2.5 exposure and premature mortality, the EPA conducted an expert elicitation in
2006.134'135 In general, the results of the expert elicitation support the conclusion that the
benefits of PM2.5 control are very likely to be substantial.
Alternative concentration-response functions are useful for assessing uncertainty beyond
random statistical error, including uncertainty in the functional form of the model or alternative
study design. Thus, we include the expert elicitation results as well as standard errors approaches
to provide insights into the likelihood of different outcomes and about the state of knowledge
regarding the benefits estimates. In this analysis, we present the results derived from the expert
elicitation as indicative of the uncertainty associated with a major component of the health
impact functions, and we provide the independent estimates derived from each of the twelve
experts to better characterize the degree of variability in the expert responses.
In previous RIAs, the EPA presented benefits estimates using concentration response
functions derived from the PM2.5 Expert Elicitation as a range from the lowest expert value
(Expert K) to the highest expert value (Expert E). However, this approach did not indicate the
agency's judgment on what the best estimate of PIVb.s benefits may be, and the EPA's
independent Science Advisory Board (SAB) recommended refinements to the way EPA
presented the results of the elicitation.136 As a result of this recommendation, we have presented
the ACS and Six-Cities cohort-based studies as our core premature mortality estimates, such as
in the RIA for the final PM NAAQS. Using alternate relationships between PM2.5 and premature
mortality supplied by experts, higher and lower benefits estimates are plausible, but most of the
expert-based estimates of the mean PM2.5 effect on mortality fall between the two epidemiology-
based estimates. Please note that the benefits estimates results presented are not the direct results
from the studies or expert elicitation; rather, the estimates are based in part on the effect
coefficients provided in those studies or by experts. In addition, the experts provided
distributions around their mean PM2.5 effect estimates, which provide more information
regarding the overall range of uncertainty, and this overall range is larger than the range of the
mean effect estimates from each of the experts.
Even these multiple characterizations with confidence intervals omit the contribution to
overall uncertainty from uncertainty in air quality changes, baseline incidence rates, and
populations exposed. Furthermore, the approach presented here does not yet include methods for
addressing correlation between input parameters and the identification of reasonable upper and
lower bounds for input distributions characterizing uncertainty in additional model elements. As
a result, the reported confidence intervals and range of estimates give an incomplete picture
about the overall uncertainty in the estimates. This information should be interpreted within the
context of the larger uncertainty surrounding the entire analysis.
8-36
-------
8.1.2.5.3 Concentration Benchmark Analysis for PM2.5
In this analysis, we estimate the number of avoided PM2.s-related deaths occurring due to
PM2.5 reductions down to various PM2.5 concentration benchmarks, including the Lowest
Measured Level (LML) of each long-term PM2.5 mortality study. This analysis is one of several
sensitivities that the EPA has historically performed that characterize the uncertainty associated
with the PM-mortality relationship and the economic value of reducing the risk of premature
death.137'138'139
Our review of the current body of scientific literature indicates that a log-linear no-
threshold model provides the best estimate of PM-related long-term mortality. The PM ISA,140
which was twice reviewed by the EPA's Clean Air Scientific Advisory Committee,141>142
concluded that the evidence supports the use of a no-threshold log-linear model while also
recognizing potential uncertainty about the exact shape of the concentration-response function.1"
Consistent with this finding, we estimate benefits associated with the full range of PIVb.s
exposure in conjunction with sensitivity analyses to recognize the potential uncertainty at lower
concentrations. Specifically, we incorporated a LML assessment, a method the EPA has
employed in several recent RIAs.143'1 4'145 In addition, we have incorporated an assessment using
specific concentration benchmarks identified in the EPA's Policy Assessment for P articulate
Matter1^
These two approaches summarize the distribution of avoided PM2.s-related mortality
impacts relative to baseline (i.e., pre-rule) annual mean PM2.s levels. The LML approach
compares the percentage of avoided premature deaths estimated to occur above and below the
minimum observed air quality level of each long-term cohort study we use to quantify PM. In the
air quality benchmark approach, we summarize the impacts occurring at different points in the
distribution of the air quality data used in these same epidemiology studies.
Our confidence in the estimated number of premature deaths avoided (but not in the
existence of a causal relationship between PM and premature mortality) diminishes as we
estimate these impacts in locations where PM2.5 levels are below the LML. This interpretation is
consistent with the Policy Assessment and advice from SAB-CASAC during their peer review.147
The Policy Assessment concludes that the range from the 25th to the 10th percentile is a
reasonable range of the air quality distribution below which we start to have appreciably less
confidence in the magnitude of the associations observed in the epidemiological studies. In
general, we are more confident in the magnitude of the risks we estimate from simulated PM2.5
concentrations that coincide with the bulk of the observed PM concentrations in the
epidemiological studies at are used to estimate the benefits. Likewise, we are less confident in
the risk we estimate from simulated PM2.5 concentrations that fall below the bulk of the observed
data in these studies. However, there are uncertainties inherent in identifying any particular point
at which our confidence in reported associations becomes appreciably less, and the scientific
evidence provides no clear dividing line.
P For a summary of the scientific review statements regarding the lack of a threshold in the PM2 5-mortality
relationship, see the Technical Support Document (TSD) entitled Summary of Expert Opinions on the Existence of
a Threshold in the Concentration-Response Function for PM2.s-related Mortality (U.S. EPA, 2010f).
8-37
-------
For these reasons, we consider the LML as well as one standard deviation below the
meanQ air quality levels when characterizing the distribution of mortality impacts. It is important
to emphasize that "less confidence" does not mean "no confidence." In addition, while we may
have less confidence in the magnitude of the risk, we still have high confidence that PIVb.s is
causally associated with risk at those lower air quality concentrations. To clarify this concept,
Figure 8-2graphically displays the spectrum of confidence using illustrative concentration
benchmarks from the major epidemiology studies cited in this chapter.
Most Confidence
Mean of PM2.5 data in epidemiology study
1 standard deviation below the mean PM25 data in epidemiology study
Below LML of PM2.5 data in epidemiology study (extrapolation)
Less Confidence
Figure 8-2: Relationship between the Size of the PM Mortality Estimates and the PM2.s
Concentration Observed in the Epidemiology Study
Although these types of concentration benchmark analyses (e.g., LML, one standard
deviation below the mean, etc.) provide some insight into the level of uncertainty in the
estimated PIVb.s mortality benefits, the EPA does not view these concentration benchmarks as a
concentration threshold below which we would not quantify health benefits of air quality
improvements. Rather, the core benefits estimates reported in this RIA (i.e., those based on
Krewski et al. (2009) and Lepeule et al. (2012)) are the best measures because they reflect the
full range of modeled air quality concentrations associated with the emission reduction
strategies. In reviewing the Policy Assessment, SAB-CASAC confirmed that "[although there is
increasing uncertainty at lower levels, there is no evidence of a threshold (i.e., a level below
which there is no risk for adverse health effects)". 148 In addition, in reviewing the Costs and
Benefits of the Clean Air Act,149 the SAB-HES noted that "[t]his [no-threshold] decision is
supported by the data, which are quite consistent in showing effects down to the lowest measured
levels. Analyses of cohorts using data from more recent years, during which time PM
concentrations have fallen, continue to report strong associations with mortality".150 Therefore,
Q A range of one standard deviation around the mean represents approximately 68 percent of normally distributed
data, and, below the mean falls between the 25th and 10th percentiles.
8-38
-------
the best estimate of benefits includes estimates below and above these concentration benchmarks
but uncertainty is higher in the magnitude of health benefits estimated at lower concentrations,
with the lowest confidence below the LML. Estimated health impacts reflecting air quality
improvements below and above these concentration benchmarks are appropriately included in
the total benefits estimate. In other words, our higher confidence in the estimated benefits above
these concentration benchmarks should not imply an absence of confidence in the benefits
estimated below these concentration benchmarks.
We estimate that most of the avoided PM-related impacts quantified in this analysis occur
among populations exposed at or above the LML of the Lepeule et al. (2012) study, while a
majority of the impacts occur at or above the LML of the Krewski et al. (2009) study. We show
the estimated reduction in incidence of premature mortality above and below the LML or air
quality benchmarks of these studies in Table 8-17, and we graphically display the distribution of
PM2.s-related mortality impacts for the final standard in Figure 8-3 and Figure 8-4.
Table 8-17: Estimated Reduction in Incidence of Adult Premature Mortality Occurring
Above and Below Various Concentration Benchmarks in the Underlying Epidemiology
Studies3
Allocation of Reduced Mortality
Epidemiology
Study
Krewski et al.
(2009)
Lepeule et al.
(2012)
Total
Reduced
Mortality
Incidence
660
1,500
Below 1 Std.
Dev.
Below AQ
Mean
620
(94%)
N/A
At or
Above 1
Std. Dev.
Below AQ
Mean
38
(6%)
N/A
Below
LML
37
(6%)
630
(42%)
Incidence
At or Above
LML
620
(94%)
860
(58%)
Mortality incidence estimates are rounded to whole numbers and two significant digits, so estimates may not
sum across columns. One standard deviation below the mean is equivalent to the middle of the range between
the 10th and 25th percentile. For Krewski, the LML is 5.8 ug/m3 and one standard deviation below the mean is
11.0 ug/m3. For Lepeule et al., the LML is 8 ug/m3 and we do not have the data for one standard deviation
below the mean. It is important to emphasize that although we have lower levels of confidence in levels below
the LML for each study, the scientific evidence does not support the existence of a level below which health
effects from exposure to PM2 5 do not occur.
8-39
-------
LMLof Krewskietal. (2009) study
_QJ
O
5
Q.
O
>
•6
0
0
_ •
LMLof Lepeuleetal. (2012)study
1
<1 1-2 2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12 12-13 13-14 14-15 15-16 16-17 17-18 18-19 19-20
Baseline Annual Mean PM25 Level (ug/m3)
Of total PM2.s-Related deaths avoided for 12 jig/m3:
94% occur among populations exposed to PM2 5 levels at or above the LML of the Krewski et al. (2009) study.
58% occur among populations exposed to PM2 5 levels at or above the LML of the Lepeule et al. (2012) study.
Figure 8-3. Number of Premature PMi.s-related Deaths Avoided for the Final Tier 3
Standards in 2030 According to the Baseline Level of PMi.s and the Lowest Measured Air
Quality Levels of Each Mortality Study
8-40
-------
Cumulative percentage of avoided PM Mortalities
i-»N)UJ.eii-ncn-vJCouDo
OOOOOOOOOOO
Cumulative Percentage of Total PM-related Mortalities Avoided by
Baseline Air Quality Level
LML of Krewski et al. (2009) study
_x
/
/
/
'
X^
f
1
LMLof Lepeuleetal. (2012)study
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Baseline Annual Mean PM2 5 Level (|4g/m3)
Of total PM2 s-Related deaths avoided for 12 fig/nr:
94% occur among populations exposed to PM2 5 levels at or above the LML of the Krewski et al. (2009) study.
58% occur among populations exposed to PM2 5 levels at or above the LML of the Lepeule et al. (2012) study.
Figure 8-4. Number of Premature PMi.s-related Deaths Avoided for the Final Tier 3
Standards in 2030 According to the Baseline Level of PMi.s and the Lowest Measured Air
Quality Levels of Each Mortality Study
While the LML of each study is important to consider when characterizing and
interpreting the overall level of PM2.s-related benefits, as discussed earlier in this chapter, the
EPA believes that both of the cohort-based mortality estimates are suitable for use in air
pollution health impact analyses. When estimating PM-related premature deaths avoided using
risk coefficients drawn from the Lepeule et al. (2012) analysis of the Harvard Six Cities and the
Krewski et al. (2009) analysis of the ACS cohorts there are innumerable other attributes that may
affect the size of the reported effect estimates—including differences in population
demographics, the size of the cohort, activity patterns and particle composition among others.
The LML assessment presented here provides a limited representation of one key difference
between the two studies.
8-41
-------
8.1.2.5.4 Alternative Concentration-Response Functions for Ozone-
Related Mortality
In 2006 the EPA requested an NAS study to evaluate the extent to which the
epidemiological literature to that point improved the understanding of ozone-related mortality.
The NAS found that short-term ozone exposure was likely to contribute to ozone-related
mortality151 and issued a series of recommendations to EPA, including that the Agency should
present multiple short-term ozone mortality estimates, including those based on multi-city
analyses such as the National Morbidity, Mortality and Air Pollution Study (NMMAPS) as well
as meta-analytic studies. The NAS also recommended that EPA remove reference to a no-causal
relationship between ozone exposure and premature mortality. The quantification and
presentation of ozone-related premature mortality in this analysis is responsive to these
recommendations.
8.1.2.5.5 Influence Analysis - Quantitative Assessment of
Uncertainty
In the past few years, the EPA has initiated several projects to improve the
characterization of uncertainty for benefits analysis. In particular, the EPA recently completed
the first phase of a quantitative uncertainty analysis of benefits, hereafter referred to as the
"Influence Analysis".152 The Influence Analysis diagramed the uncertain components of each
step within the benefits analysis process, identified plausible ranges for a sensitivity analysis, and
assessed the sensitivity to total benefits to changes in each component. Although this analysis
does not quite fulfill the goal of a full probabilistic assessment, it accomplished the necessary
first steps and identified the challenges to accomplishing that goal. Below are some of the
preliminary observations from the first phase of the project.
• The components that contribute the most to uncertainty of the monetized benefits
and mortality incidence (in order of importance) are the value-of-a-statistical-life
(VSL), the concentration-response (C-R) function for mortality, and change in
PM2.5 concentration.
• The components that contribute the least to uncertainty of the monetized benefits
and mortality incidence are population, morbidity valuation, and income
elasticity.
• The choice of a C-R function for mortality affects the mortality incidence and
monetized benefits more than other sources of uncertainty within each C-R
function.
• Alternative cessation lag structures for mortality have a moderate effect on the
monetized benefits.
• Because the health impact function is essentially linear, the key components show
the same sensitivity across all mortality C-R functions even if the midpoints differ
significantly from one expert to another.
8-42
-------
8.1.2.5.6 Qualitative Assessment of Uncertainty and Other Analysis
Limitations
Although we strive to incorporate as many quantitative assessments of uncertainty as
possible, there are several aspects we are only able to address qualitatively. These aspects are
important factors to consider when evaluating the benefits of the final Tier 3 standards.
The total monetized benefit estimates presented in this chapter are based on our
interpretation of the best available scientific literature and methods and supported by the EPA's
independent SAB (Health Effects Subcommittee) (SAB-HES) and the National Academies of
Science (NAS).153'154 The benefit estimates are subject to a number of assumptions and
uncertainties. For example, the key assumptions underlying the estimates for premature
mortality, which account for over 98% of the total monetized benefits in this analysis, include
the following:
1. We assume that all fine particles, regardless of their chemical
composition, are equally potent in causing premature mortality. This is an
important assumption, because PM2.5 varies considerably in composition across
sources, but the scientific evidence is not yet sufficient to allow differentiation of
effect estimates by particle type. The 2009 PM ISA, which was twice reviewed by
Clean Air Scientific Advisory Committee (SAB-CASAC), concluded that "many
constituents of PM2.5 can be linked with multiple health effects, and the evidence
is not yet sufficient to allow differentiation of those constituents or sources that
are more closely related to specific outcomes".155
2. We assume that the health impact function for fine particles is log-linear
without a threshold in this analysis. Thus, the estimates include health benefits
from reducing fine particles in areas with varied concentrations of PM2.5,
including both areas that do not meet the fine particle standard and those areas
that are in attainment, down to the lowest modeled concentrations.
3. We assume that there is a "cessation" lag between the change in PM
exposures and the total realization of changes in mortality effects. Specifically,
we assume that some of the incidences of premature mortality related to PM2.5
exposures occur in a distributed fashion over the 20 years following exposure
based on the advice of the SAB-HES,156 which affects the valuation of mortality
benefits at different discount rates.
4. To characterize the uncertainty in the relationship between PM2.5 and
premature mortality, we include a set of twelve estimates based on results of the
expert elicitation study in addition to our core estimates. Even these multiple
characterizations omit the uncertainty in air quality estimates, baseline incidence
rates, populations exposed and transferability of the effect estimate to diverse
locations. As a result, the reported confidence intervals and range of estimates
give an incomplete picture about the overall uncertainty in the PIVb.s estimates.
8-43
-------
This information should be interpreted within the context of the larger uncertainty
surrounding the entire analysis.
5. There is uncertainty in the magnitude of the association between ozone
and premature mortality. The range of ozone benefits associated with the final
standards is estimated based on the risk of several sources of ozone-related
mortality effect estimates. In a report on the estimation of ozone-related
premature mortality published by the National Research Council, a panel of
experts and reviewers concluded that short-term exposure to ambient ozone is
likely to contribute to premature deaths and that ozone-related mortality should be
included in estimates of the health benefits of reducing ozone exposure.11
As previously described, we strive to monetize as many of the benefits anticipated from
the revised and alternative standards as possible given data and resource limitations, but the
monetized benefits estimated in this RIA inevitably only reflect a portion of the benefits.
Specifically, only certain benefits attributable to the health impacts associated with exposure to
ambient fine particles have been monetized in this analysis. Data and methodological limitations
prevented the EPA from quantifying or monetizing the benefits from several important health
benefit categories (see Table 8-5 for more information). If we could fully monetize all of the
benefit categories, the total monetized benefits would exceed the costs by an even greater margin
than we currently estimate.
To more fully address these uncertainties, including those we cannot quantify, the 2012
PM NAAQS RIA utilized a four-tiered approach using the WHO uncertainty framework,157
which provides a means for systematically linking the characterization of uncertainty to the
sophistication of the underlying risk assessment. The EPA has applied similar approaches in
other analyses.158'159 Using this framework, the PM NAAQS summarized the key uncertainties in
the health benefits analysis, including our assessment of the direction of potential bias,
magnitude of impact on the monetized benefits, degree of confidence in our analytical approach,
and our ability to assess the source of uncertainty. We refer the reader to this tiered assessment as
an assessment of the potential impact and magnitude of each aspect of uncertainty that is also
present in the final Tier 3 RIA (see Appendix 5b of the 2012 PM NAAQS RIA).160
8.1.3 Comparison of Costs and Benefits
This section presents the cost-benefit comparison related to the expected impacts of the
final Tier 3 program. In estimating the net benefits of the program, the appropriate cost measure
is 'social costs.' Social costs represent the welfare costs of a rule to society and do not consider
transfer payments (such as taxes) that are simply redistributions of wealth. For this analysis, we
estimate that the social costs of the program are equivalent to the estimated vehicle and fuel
compliance costs of the program. While vehicle manufacturers and fuel producers would see
their costs increase by the amount of those compliance costs, they are expected to pass them on
R National Research Council (NRC), (2008). Estimating Mortality Risk Reduction and Economic Benefits from
Controlling Ozone Air Pollution. The National Academies Press: Washington, D.C.
8-44
-------
in their entirety to vehicle and fuel consumers in the form of increased prices. Ultimately, these
costs will be borne by the final consumers of these goods. The social benefits of the program are
represented by the monetized value of health and welfare improvements experienced by the U.S.
population. Table 8-18 contains the estimated social costs and the estimated monetized benefits
of the program.
The results in Table 8-18 suggest that the 2030 monetized benefits of the final standards
are greater than the expected costs. Specifically, the annual benefits of the total program will
range between $7.4 to $19 billion annually in 2030 using a three percent discount rate, or
between $6.7 to $18 billion assuming a 7 percent discount rate, compared to estimated social
costs of approximately $1.5 billion in that same year. Though there are a number of health and
environmental effects associated with the final standards that we are unable to quantify or
monetize (see Table 8-5), the benefits of the final standards outweigh the projected costs.
Using a conservative benefits estimate, the 2030 benefits outweigh the costs by a factor
of 4.5. Using the upper end of the benefits range, the benefits could outweigh the costs by a
factor of 13. Thus, even taking the most conservative benefits assumptions, benefits of the final
standards clearly outweigh the costs.
Table 8-18: Summary of Annual Benefts and Costs Associated with the Final Tier 3
Program (Billions, 2011$)a
Description
Vehicle Program Costs
Fuels Program Costs
Total Estimated Costsb
Total Estimated Health Benefits°'d'e'f
3 percent discount rate
7 percent discount rate
Annual Net Benefits (Total Benefits - Total Costs)
3 percent discount rate
7 percent discount rate
2030
$0.76
$0.70
$1.5
$7.4 -$19
$6.7 -$18
$5.9 -$18
$5.2 -$17
Notes:
a All estimates represent annual benefits and costs anticipated for the year 2030. Totals are rounded to two
significant digits and may not sum due to rounding.
b The calculation of annual costs does not require amortization of costs over time. Therefore, the estimates of annual
cost do not include a discount rate or rate of return assumption (see Chapter 2 of the RIA for more information on
vehicle costs, Chapter 5 for fuel costs, and Section 8.1.1 for a summary of total program costs).
0 Total includes ozone and PM2 5 benefits. Range was developed by adding the estimate from the Bell et al., 2004
ozone premature mortality function to PM2 5-related premature mortality derived from the American Cancer Society
cohort study (Krewski et al., 2009) for the low estimate and ozone premature mortality derived from the Levy et al.,
2005 study to PM25-related premature mortality derived from the Six-Cities (Lepeule et al., 2012) study for the high
estimate.
d Annual benefits analysis results reflect the use of a 3 percent and 7 percent discount rate in the valuation of
premature mortality and nonfatal myocardial infarctions, consistent with EPA and OMB guidelines for preparing
economic analyses.
e Valuation of premature mortality based on long-term PM exposure assumes discounting over the SAB
recommended 20-year segmented lag structure described in the Regulatory Impact Analysis for the 2012 PM
National Ambient Air Quality Standards (December, 2012).
8-45
-------
all possible benefits or disbenefits are quantified and monetized in this analysis. Potential benefit categories
that have not been quantified and monetized are listed in Table 8-5.
8.1.4 Illustrative Analysis of Estimated Quantified and Monetized Impacts Associated with the
Rule in 2018
For illustrative purposes, this section presents the quantified and monetized impacts
associated with the final standards in 2018. As presented in Section 7.1.5, the emissions impacts
of the final standards in 2018 are primarily due to the effects of sulfur on the existing (pre-Tier 3)
fleet.
This analysis reflects the impacts of the final Tier 3 rule in 2018 compared to a future-
year reference scenario without the program in place. Overall, we estimate that the final rule will
lead to a net decrease in PM2.s-related health and environmental impacts in 2018 (see Section
7.2.4 for more information about the air quality modeling results). The decrease in population-
weighted national average PIVb.s exposure results in a net decrease in adverse PIVb.s-related
human health and environmental impacts (the decrease in national population-weighted annual
average PM2.5 is 0.012 ug/m3 in 2018). The air quality modeling also projects decreases in
ozone concentrations. The overall decrease in population-weighted national average ozone
exposure results in decreases in ozone-related health and environmental impacts (population-
weighted maximum 8-hour average ozone decreases by 0.15 ppb in 2018).
Table 8-19 and Table 8-20 present the annual PM2.5 and ozone health impacts in the 48
contiguous U.S. states associated with the final Tier 3 program. For each endpoint presented in
Table 8-19 and Table 8-20, we provide both the point estimate and the 90 percent confidence
interval. Using EPA's preferred estimates, based on the American Cancer Society (ACS) and
Six-Cities studies and no threshold assumption in the model of mortality, we estimate that the
final standards would result in between 180 and 400 cases of avoided PlV^.s-related premature
mortalities annually in 2018. For ozone-related premature mortality in 2018, we estimate a range
of between 45 to 210 fewer premature mortalities.
Table 8-21 presents the estimated monetary value of changes in the incidence of ozone
and PM2.s-related health and environmental effects. Total aggregate monetized benefits are
presented in Table 8-22. All monetized estimates are presented in 2011$. Where appropriate,
estimates account for growth in real gross domestic product (GDP) per capita between 2000 and
2018. The monetized value of PM2.5-related mortality also accounts for a twenty-year segmented
cessation lag.8 To discount the value of premature mortality that occurs at different points in the
s Based in part on prior SAB advice, EPA has typically assumed that there is a time lag between changes in
pollution exposures and the total realization of changes in health effects. Within the context of benefits analyses,
this term is often referred to as "cessation lag". The existence of such a lag is important for the valuation of
premature mortality incidence because economic theory suggests that benefits occurring in the future should be
discounted. In this analysis, we apply a twenty-year distributed lag to PM mortality reductions. This method is
consistent with the most recent recommendation by the EPA's Science Advisory Board. Refer to: EPA - Science
Advisory Board, 2004. Advisory Council on Clean Air Compliance Analysis Response to Agency Request on
8-46
-------
future, we apply both a 3 and 7 percent discount rate. We also use both a 3 and 7 percent
T
discount rate to value PM2.s-related nonfatal heart attacks (myocardial infarctions).
In addition to omitted benefits categories such as air toxics and various welfare effects,
not all known PM2.5- and ozone-related health and welfare effects could be quantified or
monetized. The estimate of total monetized health benefits of the final standards is thus equal to
the subset of monetized PM2.5- and ozone-related health impacts we are able to quantify plus the
sum of the nonmonetized health and welfare benefits. Our estimate of total monetized benefits
associated with the final standards in 2018, using the ACS and Six-Cities PM mortality studies
and the range of ozone mortality assumptions, is between $2.1 and $5.6 billion, assuming a 3
percent discount rate, or between $1.9 and $5.3 billion, assuming a 7 percent discount rate.
The results in Table 8-21 demonstrate that the gasoline sulfur standards provide large
immediate benefits in the program's first year, related to emission reductions from existing
gasoline vehicles. The benefits increase substantially after 2018, as the vehicle standards phase
in after 2018 and as the fleet turns over.
Table 8-19: Estimated PM2.5-Related Health Impacts"
Health Effect
Premature Mortality - Derived from epidemiology literature15
Adult, age 30+, ACS Cohort Study (Krewski et al., 2009)
Adult, age 25+, Six-Cities Study (Lepeule et al., 2012)
Infant, age <1 year (Woodruff et al., 1997)
Non-fatal myocardial infarction (adult, age 18 and over)
Peters etal. (2001)
Pooled estimate of 4 studies
Hospital admissions - respiratory (all ages)°'e
Hospital admissions - cardiovascular (adults, age >18)d
Emergency room visits for asthma (age 18 years and younger)6
Acute bronchitis, (children, age 8-12)e
Lower respiratory symptoms (children, age 7-14)
2018 Annual
Reduction in
Incidence
(5tho/o . 95tho/oile:)
180
(130-220)
400
(230 - 570)
0
(0-1)
200
(74 - 330)
22
(11-48)
51
(-6-93)
64
(33-110)
100
(-18-200)
280
(-10 - 560)
3,500
(1,700-5,300)
Cessation Lag. Letter from the Health Effects Subcommittee to the U.S. Environmental Protection Agency
Administrator, December.
T Nonfatal myocardial infarctions (MI) are valued using age-specific cost-of-illness values that reflect lost earnings
and direct medical costs over a 5-year period following a nonfatal MI.
8-47
-------
Upper respiratory symptoms (asthmatic children, age 9-18)
Asthma exacerbation (asthmatic children, age 6-18)
Work loss days
Minor restricted activity days (adults age 18-65)
5,000
(1,600 - 8,500)
5,200
(650 - 10,000)
23,000
(20,000 - 27,000)
140,000
(120,000 - 160,000)
a Incidence is rounded to two significant digits. Estimates represent incidence within the 48 contiguous United
States.
b PM-related adult mortality based upon the most recent American Cancer Society (ACS) Cohort Study
(Krewski et al., 2009) and the most recent Six-Cities Study (Lepeule et al., 2012). Note that these are two
alternative estimates of adult mortality and should not be summed. PM-related infant mortality based upon a
study by Woodruff, Grillo, and Schoendorf, (1997)u
0 Respiratory hospital admissions for PM include admissions for chronic obstructive pulmonary disease
(COPD), pneumonia and asthma.
d Cardiovascular hospital admissions for PM include total cardiovascular and subcategories for ischemic heart
disease, dysrhythmias, and heart failure.
e The negative estimates at the 5th percentile confidence estimates for these morbidity endpoints reflect the
statistical power of the study used to calculate these health impacts. These results do not suggest that reducing
air pollution results in additional health impacts.
Table 8-20: Estimated Ozone-Related Health Impacts"
Health Effect
Premature Mortality, All agesb
Multi-City Analyses
Bell et al. (2004) - Non-accidental
Huang et al. (2005) - Cardiopulmonary
Schwartz (2005) - Non-accidental
Meta-analyses:
Bell et al. (2005) - All cause
Ito et al. (2005) - Non-accidental
Levy et al. (2005) - All cause
Hospital admissions- respiratory causes (adult, 65 and older)c
Hospital admissions -respiratory causes (children, under 2)
Emergency room visit for asthma (all ages)
2018 Annual Reduction in
Incidence (5th -95th
percentile)
45
(17-73)
65
(27 - 100)
69
(25-110)
150
(75 - 220)
200
(130-280)
210
(150-270)
260
(130-280)
130
(60-210)
140
u Woodruff, T.J., J. Grillo, and K.C. Schoendorf. (1997). The Relationship Between Selected Causes of
Postneonatal Infant Mortality and Particulate Air Pollution in the United States. Environmental Health Perspectives
105(6):608-612.
8-48
-------
Health Effect
Minor restricted activity days (adults, age 18-65)
School absence days
2018 Annual Reduction in
Incidence (5th -95th
percentile)
(-19-430)
270,000
(120,000-420,000)
92,000
(36,000 - 140,000)
Notes:
a Incidence is rounded to two significant digits. Estimates represent incidence within the 48 contiguous U.S.
b Estimates of ozone-related premature mortality are based upon incidence estimates derived from several alternative
studies: Bell et al. (2004); Huang et al. (2005); Schwartz (2005) ; Bell et al. (2005); Ito et al. (2005); Levy et al. (2005).
The estimates of ozone-related premature mortality should therefore not be summed.
0 Respiratory hospital admissions for ozone include admissions for all respiratory causes and subcategories for
COPD and pneumonia.
Table 8-21: Estimated Monetary Value of Changes in Incidence of Health and Welfare
Effects (millions of 2011$) a'b
HEALTH ENDPOINTS
2018
(5THAND95TH
PERCENTILE)
PM2.5-Related Health Effects
Premature Mortality - Derived
from Epidemiology Studies'3'0
Adult, age 30+ - ACS study
(Krewski et al., 2009)
3% discount rate
7% discount rate
Adult, age 25+ - Six-Cities study
(Lepeule etal., 2012)
3% discount rate
7% discount rate
Infant Mortality, <1 year -
(Woodruff etal. 1997)
Non-fatal acute myocardial infarctions
Peters etal., 2001
3% discount rate
7% discount rate
Pooled estimate of 4 studies
3% discount rate
7% discount rate
Hospital admissions for respiratory causes'1
Hospital admissions for cardiovascular causes
Emergency room visits for asthmad
$1,600
($230 - $3,700)
$1,400
($210 -$3,300)
$3,500
($5 10 -$8,600)
$3,200
($460 - $7,700)
$4.0
($0.55 -$10)
$26
($5.9 - $60)
$25
($5. 5 -$59)
$2.8
($0.74 - $7.3)
$2.7
($0.68 - $7.2)
$1.4
(-$0.32 - $2.7)
$2.6
($1.3 -$4.4)
$0.045
(-$0.007 - $0.087)
8-49
-------
Acute bronchitis (children, age 8-12)d
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthma, 9-11)
Asthma exacerbations
Work loss days
Minor restricted-activity days (MRADs)
$0.14
(-$0.005 - $0.34)
$0.076
($0.030 -$0.14)
$0.17
($0.049 - $0.38)
$0.31
($0.039 - $0.74)
$3.5
($3.1 -$4.0)
$9.8
($5.7 -$14)
Ozone-Related Health Effects
Premature Mortality, All ages -
Derived from Multi-city analyses
Premature Mortality, All ages -
Derived from Meta-analyses
Bell etal, 2004
Huang etal., 2005
Schwartz, 2005
Bell etal., 2005
Ito et al., 2005
Levy et al., 2005
Hospital admissions- respiratory causes (adult, 65 and older)
Hospital admissions- respiratory causes (children, under 2)
Emergency room visit for asthma (all ages)
Minor restricted activity days (adults, age 18-65)
School absence days
$440
($56 -$1,100)
$670
($95 -$1,700)
$710
($96 -$1,800)
$1,400
($200 - $3,600)
$2,000
($290 - $4,800)
$2,000
($300 - $4,800)
$7.5
($0.62 -$14)
$1.6
($0.74 - $2.5)
$0.061
(-$0.008 -$0.18)
$19
($7.7 -$33)
$9.4
($3.6 -$14)
a Monetary benefits are rounded to two significant digits for ease of presentation and computation. PM and
ozone benefits are nationwide.
b Monetary benefits adjusted to account for growth in real GDP per capita between 1990 and the analysis
year (2018).
0 Valuation assumes discounting over the SAB recommended 20 year segmented lag structure. Results
reflect the use of 3 percent and 7 percent discount rates consistent with EPA and OMB guidelines for
preparing economic analyses.
dThe negative estimate at the 5th percentile confidence estimate for this morbidity endpoint reflects the
statistical power of the study used to calculate this health impact. This result does not suggest that reducing
air pollution results in additional health impacts.
8-50
-------
Table 8-22: Total Estimated Monetized Ozone and PM-related Benefits Associated with
the Final Program in 2018
Total Ozone and PM Benefits (billions, 2011$) -
PM Mortality Derived from the ACS and Six-Cities Studies
3% Discount Rate
Ozone
Mortality
Function
Multi-city
Meta-analysis
Reference
Bell et al.,
2004
Huang et al.,
2005
Schwartz,
2005
Bell et al.,
2005
Ito et al.,
2005
Levy etal.,
2005
Mean Total
Benefits
$2.1 -$4.1
$2.3 - $4.2
$2.3 - $4.3
$3.1 -$5.0
$3.6 -$5.6
$3.6 -$5.6
7% Discount Rate
Ozone
Mortality
Function
Multi-city
Meta-analysis
Reference
Bell et al.,
2004
Huang et al.,
2005
Schwartz,
2005
Bell et al.,
2005
Ito et al.,
2005
Levy etal.,
2005
Mean Total
Benefits
$1.9 -$3.7
$2.1 -$3.9
$2.1 -$3.9
$2.9 - $4.7
$3.4 -$5.2
$3.5 -$5.3
8-51
-------
References
1 U.S. Energy Information Administration. Annual Energy Outlook 2013. Energy Consumption by Sector and
Source, United States, Reference Case. Last viewed on October 23, 2013 at
http://www.eia. gov/oiaf/aeo/tablebrowser/#release=AEO2013&subject=2-AEO2013&table=2-AEO2013®ion=l-
0&cases=ref2013 -d!02312a.
2 U.S. Energy Information Administration. Annual Energy Outlook 2013. Conversion Factors, Reference Case.
Last viewed on October 23, 2013 at http://www.eia. gov/oiaf/aeo/tablebrowser/#release=AEO2013&subject=0-
AEO2013&table=20-AEO2013®ion=0-0&cases=ref2013-dl02312a.
3 U.S. Energy Information Administration. Prime Supplier Sales Volumes. Last viewed on October 23, 2013 at
http://www.eia.gov/dnav/pet/pet cons_prim a EPMO POO Mgalpda.htm.
4 Krewski, D., M. Jerret, R.T. Burnett, R. Ma, E. Hughes, Y. Shi, et al. (2009). Extended follow-up and spatial
analysis of the American Cancer Society study linking paniculate air pollution and mortality. HEIResearch Report,
7Ą0,HealthEffects Institute, Boston, MA.
5 Bell, M.L., et al. (2004). Ozone and short-term mortality in 95 US urban communities, 1987-2000. Journal of the
American Medical Association, 292(19), 2372-2378.
6 Lepeule J, Laden F, Dockery D, Schwartz J (2012). Chronic Exposure to Fine Particles and Mortality: An
Extended Follow-Up of the Harvard Six Cities Study from 1974 to 2009. Environ Health Perspect. Jul;120(7):965-
70.
7 Levy, J.I., S.M. Chemerynski, and J.A. Sarnat. (2005). Ozone exposure and mortality: an empiric bayes
metaregression analysis. Epidemiology. 16(4), 458-68.
U.S. Environmental Protection Agency. (2012). Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Particulate Matter. Prepared by: Office of Air and Radiation, EPA-
452/R-12-005. Retrieved November 22, 2013 at http://www.epa.gov/ttn/ecas/ria.html.
9 U.S. Environmental Protection Agency. (2008). Final Ozone NAAQSRegulatory Impact Analysis. Prepared by:
Office of Air and Radiation, Office of Air Quality Planning and Standards. Retrieved November 22, 2013 at
http://www.epa.gov/ttn/ecas/ria.html. EPA-HQ-OAR-2009-0472-0238.
10 U.S. Environmental Protection Agency. (2010). Summary of the updated Regulatory Impact Analysis (RIA) for
the Reconsideration of the 2008 Ozone National Ambient Air Quality Standard (NAAQS). Prepared by: Office of
Air and Radiation, Office of Air Quality Plannin and Standards. Retrieved November 22, 2013 at
http://www.epa.gov/ttnecasl/regdata/RIAs/sl-supplemental_analysis_full.pdf.
11 U.S. Environmental Protection Agency. (2012). Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Particulate Matter. Prepared by: Office of Air and Radiation, EPA-
452/R-12-005. Retrieved August 14, 2013 at http://www.epa.gov/ttn/ecas/ria.html.
12 U.S. Environmental Protection Agency. (2013). Draft Regulatory Impact Analysis: Tier 3 Motor Vehicle
Emission and Fuel Standards. Prepared by: Office of Air and Radiation. EPA-420-D-13-002. Retrieved October
18, 2013 at http://www.epa.gov/otaq/documents/tier3/420dl3002.pdf.
13 Lepeule J, Laden F, Dockery D, Schwartz J. 2012. Chronic Exposure to Fine Particles and Mortality: An
Extended Follow-Up of the Harvard Six Cities Study from 1974 to 2009. Environ Health Perspect. Jul;120(7):965-
70.
14 Laden, F., J. Schwartz, F.E. Speizer, and D.W. Dockery. 2006. "Reduction in Fine Particulate Air Pollution and
Mortality." American Journal of Respiratory and Critical Care Medicine 173:667-672.
15 Krewski, D., M. Jerret, R.T. Burnett, R. Ma, E. Hughes, Y. Shi, et al. (2009). Extended follow-up and spatial
analysis of the American Cancer Society study linking paniculate air pollution and mortality. HEI Research Report,
140,RealthEffects Institute, Boston, MA.
16 Pope, C.A., III, R.T. Burnett, M.J. Thun, E.E. Calle, D. Krewski, K. Ito, and G.D. Thurston. 2002. "Lung Cancer,
Cardiopulmonary Mortality, and Long-term Exposure to Fine Particulate Air Pollution." Journal of the American
Medical Association 287:1132-1141.
8-52
-------
17 U.S. Environmental Protection Agency (U.S. EPA). 2009. Integrated Science Assessment for Paniculate Matter
(Final Report). EPA-600-R-08-139F. National Center for Environmental Assessment—RTF Division. December.
Available on the Internet at .
18 Woods & Poole Economics Inc. 2012. Population by Single Year of Age. CD-ROM. Woods & Poole Economics,
Inc. Washington, D.C.
19 Woods & Poole Economics, Inc. 2011. 2012 Complete Economic and Demographic Data Source (CEDDS).
20 U.S. Environmental Protection Agency (U.S. EPA). 2004. Air Quality Criteria for Paniculate Matter. National
Center for Environmental Assessment, Office of Research and Development, U.S. Environmental Protection
Agency, Research "Wangle Park, NC EPA/600/P-99/002bF. Available on the Internet at
.
21 U.S. Environmental Protection Agency (U.S. EPA). 2009. Integrated Science Assessment for Paniculate Matter
(Final Report). EPA-600-R-08-139F. National Center for Environmental Assessment—RTF Division. December.
Available on the Internet at .
22 U.S. Environmental Protection Agency (U.S. EPA). 2012. Provisional Assessment of Recent Studies on Health
Effects of Particulate Matter Exposure. EPA/600/R-12/056A. National Center for Environmental Assessment—RTF
Division. December.
23 Agency for Healthcare Research and Quality (AHRQ). 2009. HCUPnet, Healthcare Cost and Utilization Project.
Rockville, MD. Available on the Internet at . American Lung Association (ALA). 1999.
Chronic Bronchitis. Available on the Internet at .
24 Rosamond, W., G. Broda, E. Kawalec, S. Rywik, A. Pajak, L. Cooper, and L. Chambless. 1999. "Comparison of
Medical Care and Survival of Hospitalized Patients with Acute Myocardial Infarction in Poland and the United
States." Amencan Journal of Cardiology 83:1180-1185.
25 Agency for Healthcare Research and Quality (AHRQ). 2000. HCUPnet, Healthcare Cost and Utilization Project.
Rockville, MD. Available on the Internet at .
26 Science Advisory Board. 2001. NATA - Evaluating the National-Scale Air Toxics Assessment for 1996 - an
SAB Advisory, http://www.epa.gov/ttn/atw/sab/sabrev.html.
27 U.S. Environmental Protection Agency (U.S. EPA). 2011. The Benefits and Costs of the Clean Air Act from 1990
to 2020. Office of Air and Radiation, Washington, DC. March. Available on the Internet at
.
28 U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2008. Benefits of Reducing
Benzene Emissions in Houston, 1990-2020. EPA-COUNCIL-08-001. July. Available at
.
29 U.S. Environmental Protection Agency. (2012). Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Particulate Matter. Prepared by: Office of Air and Radiation, EPA-
452/R-12-005. Retrieved August 14, 2013 at http://www.epa.gov/ttn/ecas/na.html.
30 Bell, M.L., et al. (2004). Ozone and short-term mortality in 95 U.S. urban communities, 1987-2000. JAMA, 2004.
292(19): p. 2372-8. EPA-HQ-OAR-2009-0472-1662.
31 Huang, Y.; Dominici, F.; Bell, M. L. (2005) Bayesian hierarchical distributed lag models for summer ozone
exposure and cardio-respiratory mortality. Environmetrics. 16: 547-562. EPA-HQ-OAR-2009-0472-0233.
32 Schwartz, J. (2005) How sensitive is the association between ozone and daily deaths to control for temperature?
Am. J. Respir. Crit. Care Med. 171: 627-631. EPA-HQ-OAR-2009-0472-1678.
33 Bell, M.L., F. Dominici, and J.M. Samet. (2005). A meta-analysis of time-series studies of ozone and mortality
with comparison to the national morbidity, mortality, and air pollution study. Epidemiology. 16(4): p. 436-45. EPA-
HQ-OAR-2009-0472-0222.
8-53
-------
34 Ito, K., S.F. De Leon, and M. Lippmann (2005). Associations between ozone and daily mortality: analysis and
meta-analysis. Epidemiology. 16(4): p. 446-57. EPA-HQ-OAR-2009-0472-0231.
35 Levy, J.I., S.M. Chemerynski, and J.A. Sarnat. (2005). Ozone exposure and mortality: an empiric bayes
metaregression analysis. Epidemiology. 16(4): p. 458-68. EPA-HQ-OAR-2009-0472-0236.
36 National Research Council (NRC). 2008. Estimating Mortality Risk Reduction and Economic Benefits from
Controlling Ozone Air Pollution. National Academies Press. Washington, DC.
37 Woodruff, T.J., J. Grille, and K.C. Schoendorf. (1997). The Relationship Between Selected Causes of
Postneonatal Infant Mortality and Paniculate Air Pollution in the United States. Environmental Health Perspectives
105(6):608-612.
38 Levy JI, Baxter LK, Schwartz J. 2009. "Uncertainty and variability in health-related damages from coal-fired
power plants in the United States." Risk Anal. 29(7) 1000-1014.
39 Farm N, Roman HR, Fulcher CF, Gentile M, Wesson K, Hubbell BJ, Levy JI. 2012. "Maximizing Health Benefits
and Minimizing Inequality: Incorporating Local Scale Data in the Design and Evaluation of Air Quality Policies."
Risk Analysis. 32(6): 908—922.
40 Tagaris E, Liao KJ, Delucia AJ, et al. 2009. "Potential impact of climate change on air-pollution related human
health effects." Environ. Sci. Technol. 43: 4979-4988.
41 Kunzli, N., R. Kaiser, S. Medina, M. Studnicka, O. Chanel, P. Filliger, M. Herry, F. Horak Jr., V. Puybonnieux-
Texier, P. Quenel, J. Schneider, R. Seethaler, J-C Vergnaud, and H. Sommer. 2000. "Public-Health Impact of
Outdoor and Traffic-Related Air Pollution: A European Assessment." The Lancet 356:795-801.
42 Levy JI, Baxter LK, Schwartz J. 2009. "Uncertainty and variability in health-related damages from coal-fired
power plants in the United States." Risk Anal. 29(7) 1000-1014.
43 Abt Associates, Inc. 2010. Environmental Benefits and Mapping Program (Version 4.0). Bethesda, MD. Prepared
for U.S. Environmental Protection Agency Office of Air Quality Planning and Standards. Research Triangle Park,
NC. Available on the Internet at .
44 Hubbell, B., A. Hallberg, D.R. McCubbin, and E. Post. 2005. "Health-Related Benefits of Attaining the 8-Hr
Ozone Standard." Environmental Health Perspectives 113:73-82. ICD, 1979.
45 Tagaris E, Liao KJ, Delucia AJ, et al. 2009. "Potential impact of climate change on air-pollution related human
health effects." Environ. Sci. Technol. 43: 4979-4988.
46 Farm N, Roman HR, Fulcher CF, Gentile M, Wesson K, Hubbell BJ, Levy JI. 2012. "Maximizing Health Benefits
and Minimizing Inequality: Incorporating Local Scale Data in the Design and Evaluation of Air Quality Policies."
Risk Analysis. 32(6): 908—922.
47 U.S. Environmental Protection Agency (U.S. EPA). 201 Id. Regulatory Impact Analysis for the Final Mercury and
Air Toxics Standards. EPA-452/R-11-011. December. Available on the Internet at
.
48 Woods & Poole Economics Inc. 2012. Population by Single Year of Age. CD-ROM. Woods & Poole Economics,
Inc. Washington, D.C.
49 Hollman, F.W., T.J. Mulder, and J.E. Kalian. 2000. Methodology and Assumptions for the Population Projections
of the United States: 1999 to 2100. Population Division Working Paper No. 38, Population Projections Branch,
Population Division, U.S. Census Bureau, Department of Commerce. January. Available on the Internet at
.
50 U.S. Bureau of Economic Analysis (U.S. BEA). 2004. BEA Economic Areas (EAs). U.S. Department of
Commerce. Available on the Internet at .
51 U.S. Environmental Protection Agency (U.S. EPA). 2011. The Benefits and Costs of the Clean Air Act 1990 to
2020: EPA Report to Congress. Office of Air and Radiation, Office of Policy, Washington, DC. March. Available
on the Internet at .
8-54
-------
52 U.S. Environmental Protection Agency (U.S. EPA). 2009. Integrated Science Assessment for Paniculate Matter
(Final Report). EPA-600-R-08-139F. National Center for Environmental Assessment—RTF Division. December.
Available on the Internet at .
53 U.S. Environmental Protection Agency (U.S. EPA). 2012. Provisional Assessment of Recent Studies on Health
Effects of Paniculate Matter Exposure. EPA/600/R-12/056A. National Center for Environmental Assessment—RTF
Division. December.
54 Abt Associates, Inc. 2012. BenMAP User's Manual Appendices. Prepared for U.S. Environmental Protection
Agency Office of Air Quality Planning and Standards. Research "Wangle Park, NC. September. Available on the
Internet at .
55 Bell, M.L., et al. (2004). Ozone and short-term mortality in 95 U.S. urban communities, 1987-2000. JAMA, 2004.
292(19): p. 2372-8. EPA-HQ-OAR-2009-0472-1662.
56 Huang, Y.; Dominici, F.; Bell, M. L. (2005) Bayesian hierarchical distributed lag models for summer ozone
exposure and cardio-respiratory mortality. Environmetrics. 16: 547-562. EPA-HQ-OAR-2009-0472-0233.
57 Schwartz, J. (2005) How sensitive is the association between ozone and daily deaths to control for temperature?
Am. J. Respir. Crit. Care Med. 171: 627-631. EPA-HQ-OAR-2009-0472-1678.
58 Bell, M.L., F. Dominici, and J.M. Samet. (2005). A meta-analysis of time-series studies of ozone and mortality
with comparison to the national morbidity, mortality, and air pollution study. Epidemiology. 16(4): p. 436-45. EPA-
HQ-OAR-2009-0472-0222.
59 Ito, K., S.F. De Leon, and M. Lippmann (2005). Associations between ozone and daily mortality: analysis and
meta-analysis. Epidemiology. 16(4): p. 446-57. EPA-HQ-OAR-2009-0472-0231.
60 Levy, J.I., S.M. Chemerynski, and J.A. Sarnat. (2005). Ozone exposure and mortality: an empiric bayes
metaregression analysis. Epidemiology. 16(4): p. 458-68. EPA-HQ-OAR-2009-0472-0236.
61 Pope, C.A., III, R.T. Burnett, M.J. Thun, E.E. Calle, D. Krewski, K. Ito, and G.D. Thurston. (2002). "Lung
Cancer, Cardiopulmonary Mortality, and Long-term Exposure to Fine Paniculate Air Pollution." Journal of the
American Medical Association 287:1132-1141. EPA-HQ-OAR-2009-0472-0263.
62 Laden, F., J. Schwartz, F.E. Speizer, and D.W. Dockery. (2006). Reduction in Fine Paniculate Air Pollution and
Mortality. American Journal of Respiratory and Critical Care Medicine. 173: 667-672. EPA-HQ-OAR-2009-
0472-1661.
63 Industrial Economics, Incorporated (ffic). (2006). Expanded Expert Judgment Assessment of the Concentration-
Response Relationship Between PM2.5 Exposure and Mortality. Peer Review Draft. Prepared for: Office of Air
Quality Planning and Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC. August. EPA-
HQ-OAR-2009-0472-0242.
64 Woodruff, T.J., J. Grille, and K.C. Schoendorf. (1997). The Relationship Between Selected Causes of
Postneonatal Infant Mortality and Particulate Air Pollution in the United States. Environmental Health
Perspectives. 105(6):608-612. EPA-HQ-OAR-2009-0472-0382.
65 Peters, A., D.W. Dockery, J.E. Muller, and M.A. Mittleman. (2001). Increased Particulate Air Pollution and the
Triggering of Myocardial Infarction. Circulation. 103:2810-2815. EPA-HQ-OAR-2009-0472-0239.
66 Pope, C. A., Ill, J. B. Muhlestein, et al. 2006. "Ischemic heart disease events triggered by short-term exposure to
fine paniculate air pollution." Circulation 114(23): 2443-8.
67 Sullivan, J., L. Sheppard, et al. 2005. "Relation between short-term fine-paniculate matter exposure and onset of
myocardial infarction." Epidemiology 16(1): 41-8.
68 Zanobetti, A., M. Franklin, et al. 2009. "Fine paniculate air pollution and its components in association with
cause-specific emergency admissions." Environmental Health 8: 58-60.Zeger S; Dominici F; McDermott A; Samet
J. 2008. "Mortality in the Medicare population and chronic exposure to fine paniculate air pollution in urban centers
(2000-2005)." Environ Health Perspect 116: 1614-1619.
8-55
-------
69 Zanobetti A. and Schwartz, J. 2006. "Air pollution and emergency admissions in Boston, MA." J Epidemiol
Community Health 60(10): 890-5.
70 Schwartz J. (1995). Short term fluctuations in air pollution and hospital admissions of the elderly for respiratory
disease. Thorax. 50(5):531-538.
71 Schwartz J. (1994a). PM(10) Ozone, and Hospital Admissions For the Elderly in Minneapolis St Paul,
Minnesota. Arch Environ Health. 49(5):366-374. EPA-HQ-OAR-2009-0472-1673.
72 Schwartz J. (1994b). Air Pollution and Hospital Admissions For the Elderly in Detroit, Michigan. AmJRespir
CritCareMed. 150(3):648-655. EPA-HQ-OAR-2009-0472-1674.
73 Moolgavkar SH, Luebeck EG, Anderson EL. (1997). Air pollution and hospital admissions for respiratory causes
in Minneapolis St. Paul and Birmingham. Epidemiology. 8(4):364-370. EPA-HQ-OAR-2009-0472-1673.
74 Burnett RT, Smith-Doiron M, Stieb D, Raizenne ME, Brook JR, Dales RE, et al. (2001). Association between
ozone and hospitalization for acute respiratory diseases in children less than 2 years of age. Am J Epidemiol.
153(5):444-452. EPA-HQ-OAR-2009-0472-0223.
75 Kloog, I., B.A. Coull, A. Zanobetti, P. Koutrakis, J.D. Schwartz. 2012. Acute and Chronic Effects of Particles on
Hospital Admissions in New-England. PLoS ONE. Vol 7 (4): 1-8.
76 Moolgavkar, S.H. 2000. "Air Pollution and Hospital Admissions for Diseases of the Circulatory System in Three
U.S. Metropolitan Areas." Journal of the Air and Waste Management Association 50:1199-1206.
77 Babin, S. M., H. S. Burkom, et al. 2007. "Pediatric patient asthma-related emergency department visits and
admissions in Washington, DC, from 2001-2004, and associations with air quality, socio-economic status and age
group." Environ Health 6: 9.
78 Sheppard, L. (2003). Ambient Air Pollution and Nonelderly Asthma Hospital Admissions in Seattle,
Washington, 1987-1994. In Revised Analyses of Time-Series Studies of Air Pollution and Health. Special Report.
Boston, MA: Health Effects Institute. EPA-HQ-OAR-2009-0472-0318.
79 Peng, R. D., M. L. Bell, et al. 2009. "Emergency admissions for cardiovascular and respiratory diseases and the
chemical composition of fine particle air pollution." Environ Health Perspect 117(6): 957-63.
80 Peng, R. D., H. H. Chang, et al. 2008. "Coarse paniculate matter air pollution and hospital admissions for
cardiovascular and respiratory diseases among Medicare patients." JAMA 299(18): 2172-9.
81 Bell, M. L., K. Ebisu, et al. 2008. "Seasonal and Regional Short-term Effects of Fine Particles on Hospital
Admissions in 202 US Counties, 1999-2005." American Journal of Epidemiology 168(11): 1301-1310.
82 Peel, J. L., P. E. Tolbert, M. Klein, et al. (2005). Ambient air pollution and respiratory emergency department
visits. Epidemiology. Vol. 16 (2): 164-74. EPA-HQ-OAR-2009-0472-1663.
83 Wilson, A. M., C. P. Wake, T. Kelly, et al. (2005). Air pollution, weather, and respiratory emergency room visits
in two northern New England cities: an ecological time-series study. Environ Res. Vol. 97 (3): 312-21. EPA-HQ-
OAR-2009-0472-0246.
84 Mar, T. F., J. Q. Koenig and J. Primomo. 2010. "Associations between asthma emergency visits and paniculate
matter sources, including diesel emissions from stationary generators in Tacoma, Washington." Inhal Toxicol. Vol.
22 (6): 445-8.Mar, T. F., T. V Larson, et al. 2004. "An analysis of the association between respiratory symptoms in
subjects with asthma and daily air pollution in Spokane, Washington." Inhal Toxicol 16(13): 809-15.
85 Slaughter, James C et al. 2005. "Association between paniculate matter and emergency room visits, hospital
admissions and mortality in Spokane, Washington." Journal of Exposure Analysis and Environmental Epidemiology
(2005) 15, 153-159. doi:10.1038/sj.jea.7500382.
86 Glad, J.A., L.L. Brink, E.G. Talbott, P.C. Lee, X. Xu, M. Saul, and J. Rager. 2012. The Relationship of Ambient
Ozone and PM2.5 Levels and Asthma Emergency Department Visits: Possible Influence of Gender and Ethnicity.
Archives of Environmental & Occupational Health. Vol 62 (2): 103-108.
8-56
-------
87 Dockery, D.W., J. Cunningham, A.I. Damokosh, L.M. Neas, J.D. Spengler, P. Koutrakis, J.H. Ware, M. Raizenne,
andF.E. Speizer. (1996). Health Effects of Acid Aerosols On North American Children-Respiratory Symptoms.
Environmental Health Perspectives 104(5):500-505. EPA-HQ-OAR-2009-0472-0225.
88 Pope, C.A., III, D.W. Dockery, J.D. Spengler, and M.E. Raizenne. (1991). Respiratory Health and PM10
Pollution: A Daily Time Series Analysis. American Review of Respiratory Diseases 144:668-674. EPA-HQ-OAR-
2009-0472-1672.
89 Schwartz, J., and L.M. Neas. (2000). Fine Particles are More Strongly Associated than Coarse Particles with
Acute Respiratory Health Effects in Schoolchildren. Epidemiology 11:6-10.
90 Ostro, B., M. Lipsett, J. Mann, H. Braxton-Owens, and M. White. (2001). Air Pollution and Exacerbation of
Asthma in African-American Children in Los Angeles. Epidemiology 12(2):200-208.
91 Ostro, B.D. (1987). Air Pollution and Morbidity Revisited: A Specification Test. Journal of Environmental
Economics Management 14:87-98. EPA-HQ-OAR-2009-0472-1670.
92 Gilliland FD, Berhane K, Rappaport EB, Thomas DC, Avol E, Gauderman WJ, et al. (2001). The effects of
ambient air pollution on school absenteeism due to respiratory illnesses. Epidemiology 12(l):43-54. EPA-HQ-
OAR-2009-0472-1675.
93 Chen L, Jennison BL, Yang W, Omaye ST. (2000). Elementary school absenteeism and air pollution. Inhal
Toxicol 12(11):997-1016. EPA-HQ-OAR-2009-0472-0224.
94 Ostro, B.D. and S. Rothschild. (1989). Air Pollution and Acute Respiratory Morbidity: An Observational Study
of Multiple Pollutants. Environmental Research 50:238-247. EPA-HQ-OAR-2009-0472-0364.
95 U.S. Environmental Protection Agency. (2012). Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Particulate Matter. Prepared by: Office of Air and Radiation, EPA-
452/R-12-005. Retrieved November 22, 2013 at http://www.epa.gov/ttn/ecas/ria.html.
96 U.S. Environmental Protection Agency. (2008). Final Ozone NAAQS Regulatory Impact Analysis. Preparedby:
Office of Air and Radiation, Office of Air Quality Planning and Standards. Retrieved November 22, 2013 at
http://www.epa.gov/ttn/ecas/ria.html. EPA-HQ-OAR-2009-0472-0238.
97 U.S. Environmental Protection Agency. (2010). Summary of the updated Regulatory Impact Analysis (RIA) for
the Reconsideration of the 2008 Ozone National Ambient Air Quality Standard (NAAQS). Prepared by: Office of
Air and Radiation, Office of Air Quality Plannin and Standards. Retrieved November 22, 2013 at
http://www.epa.gov/ttnecasl/regdata/RIAs/sl-supplemental_analysis_full.pdf.
98 Abt Associates, Inc. 2012. BenMAP User's Manual Appendices. Prepared for U.S. Environmental Protection
Agency Office of Air Quality Planning and Standards. Research Triangle Park, NC. September. Available on the
Internet at .
99 Abt Associates, Inc. 2012. BenMAP User's Manual Appendices. Prepared for U.S. Environmental Protection
Agency Office of Air Quality Planning and Standards. Research Triangle Park, NC. September. Available on the
Internet at .
100 U.S. Bureau of the Census. 2002. Modified Race Data Summary File 2000 Technical Documentation. Census of
Population and Housing: Washington DC. Available on the Internet at
.
101 U.S. Environmental Protection Agency (U.S. EPA). 2011. Regulatory Impact Analysis for the Federal
Implementation Plans to Reduce Interstate Transport of Fine Particulate Matter and Ozone in 27 States; Correction
of SIP Approvals for 22 States. June. Available on the Internet at
.
102 Centers for Disease Control: Wide-ranging OnLine Data for Epidemiologic Research (CDC Wonder) (data from
years 1996-1998), Centers for Disease Control and Prevention (CDC), U.S. Department of Health and Human
Services, Available on the Internet at .
103 Agency for Healthcare Research and Quality (AHRQ). 2007. HCUPnet, Healthcare Cost and Utilization Project.
Rockville, MD. Available on the Internet at .
8-57
-------
104 Rosamond, W., G. Broda, E. Kawalec, S. Rywik, A. Pajak, L. Cooper, and L. Chambless. 1999. "Comparison of
Medical Care and Survival of Hospitalized Patients with Acute Myocardial Infarction in Poland and the United
States." American Journal of Cardiology 83:1180-1185.
105 American Lung Association. 2002. Trends in Asthma Morbidity and Mortality. American Lung Association,
Best Practices and Program Services, Epidemiology and Statistics Unit. Available on the Internet at
.
106 Adams PF, Hendershot GE, Marano MA. 1999. Current Estimates from the National Health Interview Survey,
1996. Vital Health Stat 10(200): 1-212.
107 U.S. Bureau of Census. 2000. Population Projections of the United States by Age, Sex, Race, Hispanic Origin
and Nativity: 1999 to 2100. Population Projections Program, Population Division, U.S. Census Bureau, Washington,
DC. Available on the Internet at .
108 National Center for Education Statistics (NCHS). 1996. The Condition of Education 1996, Indicator 42: Student
Absenteeism and Tardiness. U.S. Department of Education. Washington, DC.
109 American Lung Association (ALA). 2010. Trends in Asthma Morbidity and Mortality. American Lung
Association Epidemiology and Statistics Unit Research and Program Services Division. February. Table 7. Based on
NHIS data (CDC, 2008). Available at http://www.lungusa.org/finding-cures/our-research/trend-reports/asthma-
trend-report.pdf.
110 Centers for Disease Control and Prevention (CDC). 2008. National Center for Health Statistics. National Health
Interview Survey, 1999-2008.
111 Freeman(III), AM. 1993. The Measurement of Environmental and Resource Values: Theory and Methods.
Washington, DC: Resources for the Future.
112 Harrington, W., and P.R. Portney. 1987. Valuing the Benefits of Health and Safety Regulation. Journal of Urban
Economics 22:101-112.
113 Berger, M.C., G.C. Blomquist, D. Kenkel, and G.S. Tolley. 1987. Valuing Changes in Health Risks: A
Comparison of Alternative Measures. The Southern Economic Journal 53:977-984.
U.S. Environmental Protection Agency. (2012). Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Particulate Matter. Prepared by: Office of Air and Radiation, EPA-
452/R-12-005. Retrieved November 22, 2013 at http://www.epa.gov/ttn/ecas/ria.html.
115 U.S. Environmental Protection Agency (U.S. EPA). 2008. Regulatory Impact Analysis, 2008 National Ambient
Air Quality Standards for Ground-level Ozone, Chapter 6. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. March. Available at .
116 U.S. Environmental Protection Agency. (2010). Summary of the updated Regulatory Impact Analysis (RIA) for
the Reconsideration of the 2008 Ozone National Ambient Air Quality Standard (NAAQS). Prepared by: Office of
Air and Radiation, Office of Air Quality Plannin and Standards. Retrieved November 22, 2013 at
http://www.epa.gov/ttnecasl/regdata/RIAs/sl-supplemental_analysis_full.pdf.
117 U.S. Environmental Protection Agency (U.S. EPA). 2010. Guidelines for Preparing Economic Analyses. EPA
240-R-10-001. National Center for Environmental Economics, Office of Policy Economics and Innovation.
Washington, DC. December. Available on the Internet at .
118 Cropper, M. L. and A. J. Krupnick. 1990. The Social Costs of Chronic Heart and Lung Disease. Resources for the
Future. Washington, DC. Discussion Paper QE 89-16-REV.
119 Russell, M.W., D.M. Huse, S. Drowns, B.C. Hamel, and S.C. Hartz. 1998. "Direct Medical Costs of Coronary
Artery Disease in the United States." American Journal of Cardiology 81(9):1110-1115.
120 Wittels, E.H., J.W. Hay, and A.M. Gotto, Jr. 1990. "Medical Costs of Coronary Artery Disease in the United
States." American Journal of Cardiology 65(7):432-440.
8-58
-------
121 Agency for Healthcare Research and Quality (AHRQ). 2007. HCUPnet, Healthcare Cost and Utilization Project.
Rockville, MD. Available on the Internet at .
122 Smith, D.H., D.C. Malone, K.A. Lawson, LJ. Okamoto, C. Battista, and W.B. Saunders. 1997. "A National
Estimate of the Economic Costs of Asthma." American Journal of Respiratory and Critical Care Medicine 156(3 Pt
l):787-793.
123 Stanford, R., T. McLaughlin and L. J. Okamoto. 1999. "The cost of asthma in the emergency department and
hospital." AmJRespirCritCareMed.160 (1): 211-5.
124 Neumann, J.E., M.T. Dickie, and R.E. Unsworth. 1994. "Linkage Between Health Effects Estimation and
Morbidity Valuation in the Section 812 Analysis—Draft Valuation Document." Industrial Economics Incorporated
(lEc) Memorandum to Jim DeMocker, U.S. Environmental Protection Agency, Office of Air and Radiation, Office
of Policy Analysis and Review. March 31. Available on the Internet at
.
125 Rowe, R.D., and L.G. Chestnut. 1986. "Oxidants and Asthmatics in Los Angeles: A Benefits Analysis—
Executive Summary." Prepared by Energy and Resource Consultants, Inc. Report to the U.S. Environmental
Protection Agency, Office of Policy Analysis. EPA- 230-09-86-018. Washington, DC.
126 GeoLytics Inc. 2002. GeoLytics CensusCD® 2000 Short Form Blocks. CD-ROM Release 1.0. GeoLytics, Inc.
East Brunswick, NJ. Available: http://www.geolytics.com/ [accessed 29 September 2004].
127 Tolley, G.S. et al. 1986. Valuation of Reductions in Human Health Symptoms and Risks. University of Chicago.
Final Report for the U.S. Environmental Protection Agency, Office of Policy Analysis. January. Available at
.
128 U.S. EPA, (2001), "Guidance for Demonstrating Attainment of Air Quality Goals for PM2.5 and Regional
Haze", http://www.epa.gov/ttn/scram/guidance_sip.htm, Modeling Guidance, DRAFT-PM.
129 National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations. Washington, DC: The National Academies Press.
130 National Research Council (NRC). 2008. Estimating Mortality Risk Reduction and Economic Benefits from
Controlling Ozone Air Pollution. National Academies Press. Washington, DC.
131 U.S. Environmental Protection Agency (U.S. EPA). 2010. Quantitative Health Risk Assessment for Particulate
Matter—Final Report. EPA-452/R-10-005. Office of Air Quality Planning and Standards, Research Triangle Park,
NC. September. Available on the Internet at
.
132 U.S. Environmental Protection Agency (U.S. EPA). 2011. The Benefits and Costs of the Clean Air Act 1990 to
2020: EPA Report to Congress. Office of Air and Radiation, Office of Policy, Washington, DC. March. Available
on the Internet at .
133 World Health Organization (WHO). 2008. Part 1: Guidance Document on Characterizing and Communicating
Uncertainty in Exposure Assessment, Harmonization Project Document No. 6. Published under joint sponsorship of
the World Health Organization, the International Labour Organization and the United Nations Environment
Programme. WHO Press: Geneva, Switzerland. Available on the Internet at
.
134 Roman, Henry A., Katherine D. Walker, Tyra L. Walsh, Lisa Conner, Harvey M. Richmond, Bryan J. Hubbell,
and Patrick L. Kinney. 2008. "Expert Judgment Assessment of the Mortality Impact of Changes in Ambient Fine
Particulate Matter in the U.S." Environ. Sci. Technol., 42(7):2268-2274.
135 Industrial Economics, Incorporated (IEc). 2006. Expanded Expert Judgment Assessment of the Concentration-
Response Relationship Between PM2.5 Exposure and Mortality. Prepared for: Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, Research Triangle Park, NC. September. Available on the
Internet at .
136 U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2008. Characterizing
Uncertainty in Particulate Matter Benefits Using Expert Elicitation. EPA-COUNCIL-08-002. July. Available on the
8-59
-------
Internet at .
137 Roman, Henry A., Katherine D. Walker, Tyra L. Walsh, Lisa Conner, Harvey M. Richmond, Bryan J. Hubbell,
and Patrick L. Kinney. 2008. "Expert Judgment Assessment of the Mortality Impact of Changes in Ambient Fine
Paniculate Matter in the U.S." Environ. Sci. Technol., 42(7):2268-2274.
138 U.S. Environmental Protection Agency (U.S. EPA). 2006. Regulatory Impact Analysis, 2006 National Ambient
Air Quality Standards for Paniculate Matter, Chapter 5. Office of Air Quality Planning and Standards, Research
Triangle Park, NC. October. Available on the Internet at
.
139 Mansfield, Carol; Paramita Sinha; MaxHenrion. 2009. Influence Analysis in Support of Characterizing
Uncertainty in Human Health Benefits Analysis: Final Report. Prepared for U.S. EPA, Office of Air Quality
Planning and Standards. November. Available on the internet at
.
140 U.S. Environmental Protection Agency (U.S. EPA). 2009. Integrated Science Assessment for Paniculate Matter
(Final Report). EPA-600-R-08-139F. National Center for Environmental Assessment - RTF Division. December.
Available on the Internet at .
141 U.S. Environmental Protection Agency - Science Advisory Board (U.S. EPA-SAB). 2009. Consultation on
EPA's Paniculate Matter National Ambient Air Quality Standards: Scope and Methods Plan for Health Risk and
Exposure Assessment. EPA-COUNCIL-09-009. May. Available on the Internet at
.
142 U.S. Environmental Protection Agency - Science Advisory Board (U.S. EPA-SAB). 2009. Review of EPA's
Integrated Science Assessment for Paniculate Matter (First External Review Draft, December 2008). EPA-
COUNCIL-09-008. May. Available on the Internet at
.
143 U.S. Environmental Protection Agency (U.S. EPA). 2010. Proposed Regulatory Impact Analysis (RIA) for the
Transport Rule. Office of Air Quality Planning and Standards, Research Triangle Park, NC. January. Available on
the Internet at .
144 U.S. Environmental Protection Agency (U.S. EPA). 2011. Regulatory Impact Analysis for the Federal
Implementation Plans to Reduce Interstate Transport of Fine Paniculate Matter and Ozone in 27 States; Correction
of SIP Approvals for 22 States. June. Available on the Internet at
.
145 U.S. Environmental Protection Agency (U.S. EPA). 2011. Regulatory Impact Analysis for the Final Mercury and
Air Toxics Standards. EPA-452/R-11-011. December. Available on the Internet at
.
146 U.S. Environmental Protection Agency (U.S. EPA). 2011. Policy Assessment for the Review of the Paniculate
Matter National Ambient Air Quality Standards. EPA-452/D-11-003. Aprtl. Available on the Internet at
.
147 U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010. CASAC Review of
Policy Assessment for the Review of the PM NAAQS—Second External Review Draft (June 2010). EPA-CASAC-
10-015. Available on the Internet at
.
148 U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010. CASAC Review of
Policy Assessment for the Review of the PM NAAQS—Second External Review Draft (June 2010). EPA-CASAC-
10-015. Available on the Internet at
.
8-60
-------
149 U.S. Environmental Protection Agency (U.S. EPA). 201 la. The Benefits and Costs of the Clean Air Act 1990 to
2020: EPA Report to Congress. Office of Air and Radiation, Office of Policy, Washington, DC. March. Available
on the Internet at .
150 U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010. Review of EPA's
DRAFT Health Benefits of the Second Section 812 Prospective Study of the Clean Air Act. EPA-COUNCIL-10-
001. June. Available on the Internet at
.
151 National Research Council (NRC). 2008. Estimating Mortality Risk Reduction and Economic Benefits from
Controlling Ozone Air Pollution. National Academies Press. Washington, DC.
152 Mansfield, Carol; Paramita Sinha; Max Henrion. 2009. Influence Analysis in Support of Characterizing
Uncertainty in Human Health Benefits Analysis: Final Report. Prepared for U.S. EPA, Office of Air Quality
Planning and Standards. November. Available on the internet at
.
153 U.S. Environmental Protection Agency Science Advisory Board (U.S. EPA-SAB). 2010. Review of EPA's
DRAFT Health Benefits of the Second Section 812 Prospective Study of the Clean Air Act. EPA-COUNCIL-10-
001. June. Available on the Internet at
.
154 National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed Air Pollution
Regulations. Washington, DC: The National Academies Press. Washington, DC.
155 U.S. Environmental Protection Agency (U.S. EPA). 2009. Integrated Science Assessment for Paniculate Matter
(Final Report). EPA-600-R-08-139F. National Center for Environmental Assessment—RTF Division. December.
Available on the Internet at .
156 U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2004. Advisory Council on
Clean Air Compliance Analysis Response to Agency Request on Cessation Lag. EPA-COUNCIL-LTR-05-001.
December. Available on the Internet at
.
157 World Health Organization (WHO). 2008. Part 1: Guidance Document on Characterizing and Communicating
Uncertainty in Exposure Assessment, Harmonization Project Document No. 6. Published under joint sponsorship of
the World Health Organization, the International Labour Organization and the United Nations Environment
Programme. WHO Press: Geneva, Switzerland. Available on the Internet at
.
158 U.S. Environmental Protection Agency (U.S. EPA). 2010. Quantitative Health Risk Assessment for Particulate
Matter—Final Report. EPA-452/R-10-005. Office of Air Quality Planning and Standards, Research Triangle Park,
NC. September. Available on the Internet at
.
159 U.S. Environmental Protection Agency (U.S. EPA). 2011. The Benefits and Costs of the Clean Air Act 1990 to
2020: EPA Report to Congress. Office of Air and Radiation, Office of Policy, Washington, DC. March. Available
on the Internet at .
160 U.S. Environmental Protection Agency. (2012). Regulatory Impact Analysis for the Final Revisions to the
National Ambient Air Quality Standards for Particulate Matter. Prepared by: Office of Air and Radiation, EPA-
452/R-12-005. Retrieved November 22, 2013 at http://www.epa.gov/ttn/ecas/ria.html.
8-61
-------
Chapter 9 Economic Impact Analysis
9.1 Introduction
The standards will affect two sectors directly: vehicle manufacturing and petroleum
refining. For these two regulated sectors, the economic impact analysis discusses the market
impacts from the standards: the changes in price and quantity sold. In addition, although
analysis of employment impacts is not part of a benefit-cost analysis (except to the extent that
labor costs contribute to costs), employment impacts of federal rules are of particular concern in
the current economic climate of sizeable unemployment. Executive Order 13563, "Improving
Regulation and Regulatory Review" (January 18, 2011), states, "Our regulatory system must
protect public health, welfare, safety, and our environment while promoting economic growth,
innovation, competitiveness, and job creation" (emphasis added). For this reason, we are
examining the effects of these standards on employment in the regulated sectors.
The employment effects of environmental regulation are difficult to disentangle from
other economic changes and business decisions that affect employment, over time and across
regions and industries. In light of these difficulties, economic theory provides a constructive
framework for approaching these assessments and for better understanding the inherent
complexities in such assessments. Neoclassical microeconomic theory describes how profit-
maximizing firms adjust their use of productive inputs in response to changes in their economic
conditions.A In this framework, labor demand impacts for regulated sectors can be decomposed
into output and substitution effects. For the output effect, by affecting the marginal cost of
production, regulation affects the profit-maximizing quantity of output. The substitution effect
describes how, holding output constant, regulation affects the labor-intensity of production.
Because the output and substitution effects may be both positive, both negative or some
combination, standard neoclassical theory alone does not point to a definitive net effect of
regulation on labor demand at regulated firms.
In the labor economics literature there is an extensive body of peer-reviewed empirical
work analyzing various aspects of labor demand, relying on the above theoretical framework.6
This work focuses primarily on the effects of employment policies, e.g. labor taxes, minimum
wage, etc.c In contrast, the peer-reviewed empirical literature specifically estimating
employment effects of environmental regulations is very limited. Several empirical studies,
including Berman and Bui (2001)1 and Morgenstern et al (2002),2 suggest that net employment
impacts may be zero or slightly positive but small even in the regulated sector. Other research
suggests that more highly regulated counties may generate fewer jobs than less regulated ones.3
However, since these latter studies compare more regulated to less regulated counties, they
A See Layard, P.R.G., and A. A. Walters (1978), Microeconomic Theory (McGraw-Hill, Inc.), Chapter 9 (Docket
EPA-HQ-OAR-2011-0135), a standard microeconomic theory textbook treatment, for a discussion.
B See Hamermesh (1993), Labor Demand (Princeton, NJ: Princeton University Press), Chapter 2 (Docket EPA-HQ-
OAR-2011-0135) for a detailed treatment.
c See Ehrenberg, Ronald G., and Robert S. Smith (2000), Modern Labor Economics: Theory and Public Policy
(Addison Wesley Longman, Inc.), Chapter 4 (Docket EPA-HQ-OAR-2011-0135), for a concise overview.
9-1
-------
overstate the net national impact of regulation to the extent that regulation causes plants to locate
in one area of the country rather than another. List et al. (2003)4 find some evidence that this type
of geographic relocation may be occurring. Overall, the peer-reviewed literature does not contain
evidence that environmental regulation has a large impact on net employment (either negative or
positive) in the long run across the whole economy.
Analytic challenges make it very difficult to accurately produce net employment
estimates for the whole economy that would appropriately capture the way in which costs,
compliance spending, and environmental benefits propagate through the macro-economy.
Quantitative estimates are further complicated by the fact that macroeconomic models often have
very little sectoral detail and usually assume that the economy is at full employment. The EPA is
currently in the process of seeking input from an independent expert panel on modeling
economy-wide impacts, including employment effects. For more information, see:
https://federalregister.gov/a/2014-02471.
9.2 Impacts on Vehicle Manufacturing Sector
9.2.1 Vehicle Sales Impacts
This rule takes effect from MY 2017-2025. In the intervening years, it is possible that the
assumptions underlying a quantitative analysis, as well as market conditions, might change. For
this reason, we present a qualitative discussion of the effects on vehicle sales of the standards at
the aggregate market level. Light-duty vehicle manufacturers are expected to comply with the
standards primarily through technological changes to vehicles. These changes to vehicle design
and manufacturing are expected to increase manufacturers' costs of vehicle production. The
calculation is performed for an average car, an average truck and an average Class 2b/3 vehicle
rather than for individual vehicles. The analysis conducted for this rule does not have the
precision to examine effects on individual manufacturers or different vehicle classes.
Section VILA estimates the increase in vehicle costs due to the standards. These costs
differ across years and range from $46 to $65 for cars, $73 to $88 for trucks and $33 to $75 for
Class 2b/3 vehicles (see Section VILA). These costs are small relative to the cost of a vehicle.
In a fully competitive industry, these costs would be entirely passed through to consumers.
However in an oligopolistic industry such as the automotive sector, these increases in cost may
not fully pass through to the purchase price, and the consumers may face an increase in price that
is less than the increased manufacturers' costs of vehicle production.0 We do not quantify the
expected level of cost pass-through or the ultimate vehicle price increase consumers are expected
to face, apart from noting that prices are expected to increase by an amount up to the increased
manufacturers' costs.
D See, for instance, Gron, Ann, and Deborah Swenson, 2000. "Cost Pass-Through in the U.S. Automobile Market,"
Review of Economics and Statistics 82: 316-324 (Docket EPA-HQ-OAR-2011-0135-0056), who found significantly
less than full-cost pass-through using data from 1984-1994. Using full-cost pass-through overstates costs and thus
contributes to lower vehicle sales than using a lower estimate. To the extent that the auto industry has become more
competitive over time, full-cost pass-through may be more appropriate than a result based on this older study.
9-2
-------
This increase in price is expected to lower the quantity of vehicles sold. Given that we
expect that vehicle prices will not change by more than the cost increase, we expect that the
decrease in vehicle sales will be negligible.
The effect of these standards on the use and scrappage of older vehicles will be related to
its effects on new vehicle prices and the total sales of new vehicles. The increase in price is
likely to cause the turnover of the vehicle fleet (i.e., the retirement of used vehicles and their
replacement by new models) to slow slightly, thus reducing the anticipated effect of the
standards on fleet-wide emissions. Because we do not estimate the effect of the standards on
new vehicle price changes nor do we have a good estimate of the effect of new vehicle price
changes on vehicle turnover, we have not attempted to estimate explicitly the effects of the
standards on scrappage of older vehicles and the turnover of the vehicle fleet.
9.2.2 Employment Impacts in the Auto Sector
This chapter describes changes in employment in the auto sector due to this rule. As with
the refinery sector, discussed below, we focus on the auto manufacturing sector because it is
directly regulated, and because it is likely to bear a substantial share of changes in employment
due to this rule. We include discussion of effects on the parts manufacturing sector, because the
auto manufacturing sector can either produce parts internally or buy them from an external
supplier, and we do not have estimates of the likely breakdown of effort between the two sectors.
When the economy is at full employment, an environmental regulation is unlikely to have
much impact on net overall U.S. employment; instead, labor would primarily be shifted from one
sector to another. These shifts in employment impose an opportunity cost on society,
approximated by the wages of the employees, as regulation diverts workers from other activities
in the economy. In this situation, any effects on net employment are likely to be transitory as
workers change jobs (e.g., some workers may need to be retrained or require time to search for
new jobs, while shortages in some sectors or regions could bid up wages to attract workers).
On the other hand, if a regulation comes into effect during a period of high
unemployment, a change in labor demand due to regulation may affect net overall U.S.
employment because the labor market is not in equilibrium. Schmalansee and Stavins point out
that net positive employment effects are possible in the near term when the economy is at less
than full employment due to the potential hiring of idle labor resources by the regulated sector to
meet new requirements (e.g., to install new equipment) and new economic activity in sectors
related to the regulated sector.5 In the longer run, the net effect on employment is more difficult
to predict and will depend on the way in which the related industries respond to the regulatory
requirements. As Schmalansee and Stavins note, it is possible that the magnitude of the effect on
employment could vary over time, region, and sector, and positive effects on employment in
some regions or sectors could be offset by negative effects in other regions or sectors. For this
reason, they urge caution in reporting partial employment effects since it can "paint an inaccurate
picture of net employment impacts if not placed in the broader economic context."
We follow the theoretical structure in a study by Berman and Bui 6 of the impacts of
regulation in employment in the regulated sectors. In Berman and Bui's (2001, p. 274-75)
9-3
-------
theoretical model, the change in a firm's labor demand arising from a change in regulation is
decomposed into two main components: output and substitution effects.E
• The output effect describes how, if labor-intensity of production is held constant, a
decrease in output generally leads to a decrease in labor demand. However, as noted by
Berman and Bui, although it is often assumed that regulation increases marginal cost, and
thereby reduces output, it need not be the case. A regulation could induce a firm to
upgrade to less polluting and more efficient equipment that lowers marginal production
costs, for example. In such a case, output could theoretically increase.
• The substitution effect describes how, holding output constant, regulation affects the
labor-intensity of production. Although increased environmental regulation generally
results in higher utilization of production factors such as pollution control equipment and
energy to operate that equipment, the resulting impact on labor demand is ambiguous.
For example, equipment inspection requirements, specialized waste handling, or pollution
technologies that are added to the production process may affect the number of workers
necessary to produce a unit of output. Berman and Bui (2001) model the substitution
effect as the effect of regulation on pollution control equipment and expenditures that are
required by the regulation and the corresponding change in labor-intensity of production.
In summary, as the output and substitution effects may be both positive, both negative or
some combination, standard neoclassical theory alone does not point to a definitive net effect of
regulation on labor demand at regulated firms.
Following the Berman and Bui framework for the impacts of regulation on employment
in the regulated sector, we consider two effects for the auto sector: the output effect and the
substitution effect.
9.2.2.1 The Output Effect
The output effect depends on the effects of this rule on vehicle sales. If vehicle sales
decrease, employment associated with these activities will decrease. As discussed in Chapter
9.2.1, we do not make a quantitative estimate on the effect of the rule on vehicle sales, but we
note that the decrease in vehicle sales is expected to be negligible. Thus we expect any decrease
in employment in the auto sector through the output effect to be small as well.
E The authors also discuss a third component, the impact of regulation on factor prices, but conclude that this effect
is unlikely to be important for large competitive factor markets, such as labor and capital. Morgenstern, Pizer and
Shih (2002) use a very similar model, but they break the employment effect into three parts: 1) the demand effect; 2)
the cost effect; and 3) the factor-shift effect. See Morgenstern, Richard D., William A. Pizer, and Jhih-Shyang Shih.
"Jobs Versus the Environment: An Industry-Level Perspective." Journal of Environmental Economics and
Management 43 (2002): 412-436 (Docket EPA-HQ-OAR-2011-0135-0057).
9-4
-------
9.2.2.2 The Substitution Effect
The output effect, above, measures the effect due to new vehicle sales only. The
substitution effect includes the impacts due to the changes in technologies needed for vehicles to
meet the standards, separate from the effect on output (that is, as though holding output
constant). This effect includes both changes in employment due to incorporation of abatement
technologies and overall changes in the labor intensity of manufacturing.
One way to estimate this effect, given the cost estimates for complying with the rule, is to
use the ratio of workers to each $1 million of expenditures in that sector. The use of these ratios
has both advantages and limitations. It is often possible to estimate these ratios for quite specific
sectors of the economy: for instance, it is possible to estimate the average number of workers in
the light-duty vehicle manufacturing sector per $1 million spent in the sector, rather than use the
ratio from another, more aggregated sector, such as motor vehicle manufacturing. As a result, it
is not necessary to extrapolate employment ratios from possibly unrelated sectors. On the other
hand, these estimates are averages for the sectors, covering all the activities in those sectors; they
may not be representative of the labor required when expenditures are required on specific
activities, or when manufacturing processes change sufficiently that labor intensity changes. For
instance, the ratio for the motor vehicle manufacturing sector represents the ratio for all vehicle
manufacturing, not just for emissions reductions associated with compliance activities. In
addition, these estimates do not include changes in sectors that supply these sectors, such as steel
or electronics producers. They thus may best be viewed as the effects on employment in the auto
sector due to the changes in expenditures in that sector, rather than as an assessment of all
employment changes due to these changes in expenditures. In addition, this approach estimates
the effects of increased expenditures while holding constant the labor intensity of manufacturing;
it does not take into account changes in labor intensity due to changes in the nature of
production. This latter effect could either increase or decrease the employment impacts
estimated here.F
Some of the costs of this rule will be spent directly in the auto manufacturing sector, but
it is also likely that some of the costs will be spent in the auto parts manufacturing sector. We
separately present the ratios for both the auto manufacturing sector and the auto parts
manufacturing sector.
There are several public sources for estimates of employment per $1 million
expenditures. The U.S. Bureau of Labor Statistics (BLS) provides its Employment
Requirements Matrix (ERM),7 which provides direct estimates of the employment per $1 million
in sales of goods in 202 sectors. The values considered here are for Motor Vehicle
Manufacturing (NAICS 3361) and Motor Vehicle Parts Manufacturing (NAICS 3363) for 2010.
The Census Bureau provides both the Annual Survey of Manufacturers8 (ASM) and the
Economic Census. The ASM is a subset of the Economic Census, based on a sample of
F As noted above, Morgenstern et al. (2002) separate the effect of holding output constant into two effects: the cost
effect, which holds labor intensity constant, and the factor shift effect, which estimates those changes in labor
intensity.
9-5
-------
establishments; though the Census itself is more complete, it is conducted only every 5 years,
while the ASM is annual. Both include more sectoral detail than the BLS ERM: for instance,
while the ERM includes the Motor Vehicle Manufacturing sector, the ASM and Economic
Census have detail at the 6-digit NAICS code level (e.g., light truck and utility vehicle
manufacturing). While the ERM provides direct estimates of employees/Si million in
expenditures, the ASM and Economic Census separately provide number of employees and
value of shipments; the direct employment estimates here are the ratio of those values. The
values reported are for Motor Vehicle Manufacturing (NAICS 3361), Automobile and Light
Duty Motor Vehicle Manufacturing (NAICS 33611), and Motor Vehicle Parts Manufacturing
(NAICS 3363), for 2011 for the ASM, and 2007 for the Economic Census.
The values used here are adjusted to remove the employment effects of imports through
use of a ratio of domestic production to domestic sales of 0.667.G
Table 9-1 provides the values, either given (BLS) or calculated (ASM, Economic Census)
for employment per $1 million of expenditures, all based on 2011 dollars, though the underlying
data come from different years (which may account for some of the differences). These values
have changed from the Draft RIA to use the most recent values for the ASM, and to put them all
in 2011$. The different data sources provide similar magnitudes for the estimates for the sectors.
Parts manufacturing appears to be more labor-intensive than vehicle manufacturing; light-duty
vehicle manufacturing appears to be slightly less labor-intensive than motor vehicle
manufacturing as a whole.
Table 9-1 Employment per $1 Million Expenditures (2011$) in the Motor Vehicle
Manufacturing Sector"
Source
BLS ERM
ASM
ASM
Economic Census
Economic Census
BLS ERM
ASM
Economic Census
Sector
Motor Vehicle Mfg
Motor Vehicle Mfg
Light Duty Vehicle Mfg
Motor Vehicle Mfg
Light Duty Vehicle Mfg
Motor Vehicle Parts Mfg
Motor Vehicle Parts Mfg
Motor Vehicle Parts Mfg
Ratio of workers
per $1 million
expenditures
0.754
0.633
0.583
0.651
0.590
2.558
2.190
2.656
Ratio of workers per $1
million expenditures,
adjusted for domestic vs.
foreign production
0.503
0.422
0.389
0.434
0.393
1.706
1.461
1.771
Note:
G To estimate the proportion of domestic production affected by the change in sales, we use data from Ward's
Automotive Group for total car and truck production in the U.S. compared to total car and truck sales in the U.S.
For the period 2001-2010, the proportion is 66.7 percent (Docket EPA-HQ-OAR-2011-0135).
9-6
-------
"BLS ERM refers to the U.S. Bureau of Labor Statistics' Employment Requirement Matrix. ASM refers to
the U.S. Census Bureau's Annual Survey of Manufactures. Economic Census refers to the U.S. Census
Bureau's Economic Census.
Over time, the amount of labor needed in the auto industry has changed: automation and
improved methods have led to significant productivity increases. The BLS ERM, for instance,
provided estimates that, in 1993, 1.64 workers in the Motor Vehicle Manufacturing sector were
needed per $1 million of 2005$, but only 0.86 workers by 2010 (in 2005$).9 Because the ERM
is available annually for 1993-2010, we used these data to estimate productivity improvements
over time. We regressed logged ERM values on year for both the Motor Vehicle Manufacturing
and Motor Vehicle Parts Manufacturing sectors. We used this approach because the coefficient
describing the relationship between time and productivity is a direct measure of the percent
change in productivity per year. The results suggest a 3.9 percent per year productivity
improvement in the Motor Vehicle Manufacturing Sector, and a 3.8 percent per year
improvement in the Motor Vehicle Parts Manufacturing Sector. We then used the equation
resulting from the regression to project the ERM through 2025. In the results presented below,
these projected values (adjusted to 2011$) were used directly for the BLS ERM estimates. For
the ASM, we used the ratio of the projected value in each future year to the projected value in
2011 (the base year for the ASM) to determine how many workers will be needed per $1 million
of 2011$; for the Economic Census estimates, we used the ratio of the projected value in the
future years to the projected value in 2007 (the base year for that estimate).
Section 2.7 of the RIA discusses the vehicle cost estimates developed for this rule. The
maximum value for employment impacts per $1 million (before adjustments for changes in
productivity, after accounting for the share of domestic production) is 1.771 in 2011$ if all the
additional costs are in the parts sector; the minimum value is 0.389 in 2011, if all the additional
costs are in the light-duty vehicle manufacturing sector. Increased costs of vehicles and parts
would, by itself, and holding labor intensity constant, be expected to increase employment
between 2017 and 2025 by some hundreds of jobs each year.
While we estimate employment impacts, measured in job-years, beginning with program
implementation, some of these employment gains may occur earlier as auto manufacturers and
parts suppliers hire staff in anticipation of compliance with the standard. A job-year is a way to
calculate the amount of work needed to complete a specific task. For example, a job-year is one
year of work for one person. The decline in maximum employment between 2024 and 2025 is
due to a combination of expected higher productivity and rounding, which makes an employment
decrease from 760 to 743 job-years appear larger than it is.
Table 9-2 Employment Effects due to Increased Costs of Vehicles and Parts, in job-years
Year
2016
2017
2018
2019
Costs (Millions
of 2011$)
$ 21
$ 297
$ 615
$ 653
Maximum Employment Due to
Substitution Effect (if all
expenditures are in the Parts
Sector)
0
400
800
800
Minimum Employment Due to
Substitution Effect (if all
expenditures are in the Light Duty
Vehicle Mfg Sector)
0
100
200
200
9-7
-------
2020
2021
2022
2023
2024
2025
$ 697
$ 725
$ 758
$ 751
$ 761
$ 773
800
800
800
800
800
700
200
200
200
200
200
200
9.2.2.3 Summary of Employment Effects in the Auto Sector
The overall effect of the rule on auto sector employment depends on the relative
magnitude of the output effect and the substitution effect. Because we do not have quantitative
estimates of the output effect, and only a partial estimate of the substitution effect, we cannot
reach a quantitative estimate of the overall employment effects of the rule on auto sector
employment or even whether the total effect will be positive or negative. However, given that
the expected increase in production costs to the auto manufacturers is relatively small, we expect
that the magnitudes of all effects combined will be small as well.
The standards are not expected to provide incentives for manufacturers to shift
employment between domestic and foreign production. This is because the standards will apply
to vehicles sold in the U.S. regardless of where they are produced. If foreign manufacturers
already have increased expertise in satisfying the requirements of the standards, there may be
some initial incentive for foreign production, but the opportunity for domestic manufacturers to
sell in other markets might increase. To the extent that the requirements of this rule might lead
to installation and use of technologies that other countries may seek now or in the future,
developing this capacity for domestic production now may provide some additional ability to
serve those markets. This potential benefit will not apply if other countries are not likely to have
similar standards.
9.3 Impacts on Petroleum Refinery Sector
9.3.1 Refinery Sales Impacts
The key change for refiners from the standards will be more stringent sulfur
requirements. This change to fuels is expected to increase manufacturers' costs of gasoline
production by about 0.7 cents per gallon (see Section VII.B of the Preamble).
In a perfectly competitive industry, this cost would be passed along completely to
consumers. In an imperfectly competitive industry, as noted above, full cost pass-through is not
necessary: firms may choose to reduce impacts on sales by not passing along full costs. In 2004,
the Federal Trade Commission reported that "concentration for most levels of the petroleum
industry has remained low to moderate."10 Thus the assumption of competitive markets has
some foundation in this industry. We estimate that the price increase that consumers are likely to
face should be positive and up to the increase in manufacturers' costs of gasoline production.
The effect of higher gasoline prices on gasoline sales is expected to be different over the
short and long term. In the long run, in response to the increase in fuel costs, consumers can
more easily change their driving habits, including where they live or what vehicles they use.
9-8
-------
Because of this, we expect that gasoline sales will decrease more in the long run compared to the
short run as a result of the price increase due to the rule. However, because manufacturers' costs
are expected to increase less than one cent per gallon, we expect that the decrease in gasoline
sales will be negligible over all time horizons.
9.3.2 Refinery Employment Impacts
The Berman and Bui framework of output and substitution effects can also be applied to
the impact of the rule on employment in the refinery sector.11 Here we use a fully qualitative
approach. A qualitative discussion allows for a wider incorporation of additional considerations,
such as timing of impacts and the effects of the rule on imports and exports. Because the
discussion is qualitative, we do not sum the net effects on employment.
The output effect on refining sector employment is expected to be negative. The
discussion in Chapter 9.3.1 above suggests that the standards will cause a small decrease in the
quantity of gasoline demanded due to higher production costs being passed through to
consumers. This slightly reduced level of sales will likely have a negative impact on
employment in the refining sector. This effect will persist as long as the increase in price is in
place. The higher long-run elasticity suggests that sales will be lower in the long run than in the
short run, leading to a greater reduction in employment due to the output effect over time. While
we do not quantify the level of job losses that are expected here, recall that the quantity of
gasoline sold as a result of the standards here is expected to decrease by only a very small
amount over any time horizon.
The substitution effect of the rule on employment in the refining sector can be either
positive or negative in the Berman and Bui framework; here, we expect a small, possibly positive
impact. In order to satisfy the requirements of the rule, firms in the refining industry are
expected to need to perform additional work that will require hiring more employees. This effect
may be larger in the short run, when initial investments for compliance need to be made; over
time, the increase in employment due to these investments may be reduced. Chapter 4.5.1
discusses the expected employment needed to reduce the sulfur content of fuels; as noted there,
to meet the Tier 3 sulfur standards, refiners are expected to invest $2 billion between 2012 and
2019 and utilize approximately 250 front-end design and engineering jobs and 15000
construction jobs. As the petroleum sector employed approximately 71,000 workers in 2011,
this increase in employment is small when compared to 2011 levels.
These standards are not expected to provide incentives to shift employment between
domestic and foreign production. First, the standards apply to gasoline sold in the U.S.
regardless of where it has been produced. U.S. gasoline demand is projected to continue to
decline for the foreseeable future in response to higher gasoline prices, more stringent vehicle
and engine greenhouse gas and fuel economy standards as well as increased use of renewable
fuels. As a result, this analysis of incentives to shift employment between domestic and foreign
production focuses on investments for existing capacity instead of expanding capacity.H In this
H While refinery capacity has been increasing around the world in recent years, it has been designed primarily to
supply foreign markets other than the U.S. (e.g., increasing demand in China and India).
9-9
-------
case, what is relevant is whether the necessary modifications to comply with Tier 3 would be
significantly cheaper overseas than in the U.S.
The main impacts on capital and operating costs to comply with Tier 3 associated with
adding hydrotreating capacity are likely to be similar overseas as in the U.S. This is particularly
true when analyzing likely sources of U.S. imports. The majority of gasoline imported to the
U.S. today comes into the East Coast and is sourced out of either Europe or refineries in Canada
or the Caribbean that exist almost solely to supply the U.S. market.
These Canadian and Caribbean refineries, by virtue of their focus on the U.S. market, are
very similar to U.S.-based refineries and are expected to have to incur similar capital and
operating costs as their U.S.-based competitors meeting the 10-ppm standard. Furthermore, the
European refineries are already producing gasoline to a 10-ppm sulfur cap for Europe. To the
extent they have refinery streams that are more difficult to hydrotreat, the U.S. market currently
serves as an outlet for their higher sulfur gasoline streams. As a result, they may incur capital and
operating costs on a per gallon basis at least as high as for their U.S.-based competitors for these
remaining higher sulfur gasoline streams. Alternatively, they may instead choose to find markets
outside the U.S., opening the way for increased U.S.-based refinery demand.
Finally, despite refining industry projections that previously imposed diesel rules would
lead to greater U.S. reliance on imports through major negative impacts on domestic refining, the
reverse has actually occurred. Over the last 8 years, imports of gasoline and diesel fuel have
continued to be the marginal supply, and have even dropped precipitously so that the U.S. is now
a net exporter of diesel fuel and is importing half the gasoline that it did at its peak in 2006.
With the projected decline in future gasoline demand in the U.S. as vehicle fuel efficiency
improves, gasoline imports are expected to continue to decline.
Thus it is expected that for the refining sector, the output effect will lower employment,
and the substitution effect may raise employment. As a whole then, it is not evident whether the
rule will increase or decrease employment in the refining sector. However, given the small
anticipated reduction in quantity sold, it appears that the standards will not have major
employment consequences for this sector.
The petroleum refining industry is one of the manufacturing industries studied by Berman
and Bui (2001)7 when they looked at the effect of environmental expenditures on employment.
They found that "Employment effects are very small, generally positive, but not statistically
different from zero" (p. 281) [Berman and Bui, Table 3]. Berman and Bui also state that the
estimates rule out large negative effects (p. 282). Because most of the abatement cost of the
regulations they analyze is incurred by refineries, in their sample, they report separate
employment effects for refineries and non-refineries "which are also all small." (p. 282). Berman
and Bui suggest some explanations for the zero or small estimates, particularly for oil refineries:
they are capital-intensive industries with relatively little employment when compared to other
manufacturing; they face relatively inelastic demand because they sell output in local markets
and/or because there are no unregulated refineries to compete with; and, finally, regulations may
have been associated with productivity gains in petroleum refineries. We note that the
regulations that these estimate are derived from are not directly comparable to the current rule; it
9-10
-------
is based on the costs of reductions in refinery air pollution emissions instead of changing fuel
properties, and therefore may not be applicable for the standards here.
Section VII.B.5 of the Preamble contains some historical discussion regarding the impact
on refineries and refining capacity of earlier rules which resulted in higher costs for refiners.
Over the period 2003-2011, when a number of rules were being implemented, EIA data show a
net of two refinery closures on its website. Meanwhile, over this same period the average size of
U.S. refineries increased from 113,000 barrels per day to 123,000 barrels per day, and total U.S.
refining capacity increased by six percent. Thus, historically during a time when rules with
much larger expected impacts were being implemented (the 2003 ultra-low sulfur nonroad diesel
proposal alone was expected to have a cost impact on refineries more than five times greater than
the current rule), U.S. refining capacity increased even as the number of U.S. refineries slightly
fell. While closing refineries has a negative effect on industry employment, it is likely that the
increased refining capacity at many of the remaining plants had a positive effect on industry
employment.
The standards are also likely to have a positive impact on employment among producers
of equipment that refiners will use to comply with the standards. Chapter 5 notes that some
refiners are expected to revamp their current treatment units, and others will need to add
additional treatment units. Producers of this equipment are expected to hire additional labor to
meet this increased demand. We also note that the employment effects may be different in the
immediate implementation phase than in the ongoing compliance phase. It is expected that the
employment increases through the substitution effect from revamping old equipment and
installing additional equipment should occur in the near term, when current unemployment levels
are high, and the opportunity cost of workers is relatively low. Meanwhile, the employment
decreases in the refining sector from the output effect will not start until 2017, when compliance
is required, and when unemployment is expected to be reduced; in a time of full employment,
any changes in employment levels in the regulated sector are mostly expected to be offset by
changes in employment in other sectors.
9-11
-------
References
German, E. and L. T. M. Bui (2001). "Environmental Regulation and Labor Demand: Evidence from the South
Coast Air Basin." Journal of Public Economics 79(2): 265-295 (DocketEPA-HQ-OAR-2011-0135).
2Morgenstern, Richard D., William A. Pizer, and Jhih-Shyang Shih. "Jobs Versus the Environment: An Industry-
Level Perspective." Journal of Environmental Economics and Management 43 (2002): 412-436 (Docket EPA-HQ-
OAR-2011-0135-0057).
3 Greenstone, M. (2002). "The Impacts of Environmental Regulations on Industrial Activity: Evidence from the
1970 and 1977 Clean Air Act Amendments and the Census of Manufactures," Journal of Political Economy 110(6):
1175-1219 (Docket EPA-HQ-OAR-2011-0135); Walker, Reed. (201 ^."Environmental Regulation and Labor
Reallocation." American Economic Review: Papers and Proceedings 101(3): 442-447 (DocketEPA-HQ-OAR-
2011-0135).
4List, J. A., D. L. Millimet, P. G. Fredriksson, and W. W. McHone (2003). "Effects of Environmental Regulations
on Manufacturing Plant Births: Evidence from a Propensity Score Matching Estimator." The Review of Economics
and Statistics 85(4): 944-952 (DocketEPA-HQ-OAR-2011-0135).
5 Schmalensee, Richard, and Robert N. Stavins. "A Guide to Economic and Policy Analysis of EPA's Transport
Rule." White paper commissioned by Excelon Corporation, March 2011 (Docket EPA-HQ-OAR-2011-0135-0054).
6Berman, E. and L. T. M. Bui (2001). "Environmental Regulation and Labor Demand: Evidence from the South
Coast Air Basin." Journal of Public Economics 79(2): 265-295 (DocketEPA-HQ-OAR-2011-0135).
7 http://www.bls.gov/emp/ep_data_emp_requirements.htm.
8 http://www.census.gov/manufacturing/asm/index.html.
9 http://www.bls.gov/emp/ep_data_emp_requirements.htm; this analysis used data for sectors 88 (Motor Vehicle
Manufacturing) and 90 (Motor Vehicle Parts Manufacturing) from "Chain-weighted (2000 dollars) real domestic
employment requirements table. . . adjusted to remove imports."
10 Federal Trade Commission, Bureau of Economics. "The Petroleum Industry: Mergers, Structural Change, and
Antitrust Enforcement." http://www.ftc.gov/os/2004/08/040813mergersinpetrolberpt.pdf, accessed 8/16/11 (Docket
EPA-HQ-OAR-2011-0135-0055).
11 Berman, E. and L. T. M. Bui (2001). "Environmental Regulation and Labor Demand: Evidence from the South
Coast Air Basin." Journal of Public Economics 79(2): 265-295 (DocketEPA-HQ-OAR-2011-0135).
9-12
-------
Chapter 10 Final Regulatory Flexibility Analysis
10.1 Introduction
This chapter discusses our Final Regulatory Flexibility Analysis (FRFA) which evaluates
the potential impacts of the proposed standards on small entities. The Regulatory Flexibility Act,
as amended by the Small Business Regulatory Enforcement Fairness Act of 1996 (SBREFA),
generally requires an agency to prepare a regulatory flexibility analysis of any rule subject to
notice and comment rulemaking requirements under the Administrative Procedure Act or any
other statute unless the agency certifies that the rule will not have a significant economic impact
on a substantial number of small entities. Prior to issuing a proposal for this rulemaking, we
analyzed the potential impacts of these regulations on small entities. As a part of this analysis,
we convened a Small Business Advocacy Review Panel (SBAR Panel, or 'the Panel'). During
the Panel process, we gathered information and recommendations from Small Entity
Representatives (SERs) on how to reduce the impact of the rule on small entities, and those
comments are detailed in the Final Panel Report which is located in the public record for this
rulemaking (Docket ID Number EPA-HQ-OAR-201 l-0135-0423).Pursuant to this requirement,
we have prepared an IRF A for the proposed rule. Throughout the process of developing the
IRFA, we conducted outreach and held meetings with representatives from the various small
entities that could be affected by the rulemaking to gain feedback, including recommendations,
on how to reduce the impact of the rule on these entities. The small business recommendations
stated here reflect the comments of the small entity representatives (SERs) and members of the
Small Business Advocacy Review Panel (SBAR Panel, or 'the Panel').
10.2 Overview of the Regulatory Flexibility Act
In accordance with section 609(b) of the Regulatory Flexibility Act, we convened an
SBAR Panel before conducting the FRF A. A summary of the Panel's recommendations is
presented in the preamble to the proposed rule. Further, a detailed discussion of the Panel's
advice and recommendations (as well as comments from the Small Entity Representatives) can
be found in the Final Panel Report contained in the docket for this proposed rulemaking.l The
regulatory alternatives that are being adopted in the final rule are described below.
Section 609(b) of the Regulatory Flexibility Act further directs the Panel to report on the
comments of small entity representatives and make findings on issues related to identified
elements of the Regulatory Flexibility Analysis under section 603 of the Regulatory Flexibility
Act. Key elements of a Regulatory Flexibility Analysis are:
• A description of and, where feasible, an estimate of the number of small entities to
which the rule will apply;
• Projected reporting, recordkeeping, and other compliance requirements of the rule,
including an estimate of the classes of small entities which will be subject to the
requirements and the type of professional skills necessary for preparation of the report
or record;
10-1
-------
• An identification to the extent practicable, of all other relevant Federal rules which
may duplicate, overlap, or conflict with the rule;
• Any significant alternatives to the rule which accomplish the stated objectives of
applicable statutes and which minimize any significant economic impact of the rule
on small entities.
The Regulatory Flexibility Act was amended by SBREFA to ensure that concerns
regarding small entities are adequately considered during the development of new regulations
that affect those entities. Although we are not required by the Clean Air Act to provide special
treatment to small businesses, the Regulatory Flexibility Act requires us to carefully consider the
economic impacts that our rules will have on small entities. The recommendations made by the
Panel may serve to help lessen these economic impacts on small entities when consistent with
Clean Air Act requirements.
10.3 Need for the Rulemaking and Rulemaking Objectives
A detailed discussion on the need for, and objectives of, this rule are located in the
preamble to the final rule. As presented in Chapter 9 of this RIA, controlling exhaust and
evaporative emissions from light-duty vehicles and trucks and complete heavy-duty vehicles and
reducing sulfur levels in gasoline have important public health and welfare benefits. Further, as
discussed in Section II of the preamble to the final rule, section 202 of the Clean Air Act
(specifically, sections 202(a) and (k)) authorizes EPA to establish emissions standards for motor
vehicles to address air pollution that may reasonably be anticipated to endanger public health or
welfare. EPA also has authority to establish fuel controls to address such air pollution under
section 21 l(c) of the Clean Air Act. Emissions from motor vehicles and their fuels contribute to
pollutants for which EPA has established health-based NAAQS, and motor vehicles also emit air
toxics and contribute to near-road air pollution.
EPA's current Tier 2 Vehicle and Gasoline Sulfur Program, which was finalized in
February 2000, took a systems-based approach to motor vehicle pollution by setting standards
for both passenger vehicles and their fuel (gasoline). The Tier 2 program set stricter tailpipe and
evaporative emissions standards for criteria pollutants from vehicles beginning with model year
(MY) 2004 and phasing in through 2009. The program also lowered the sulfur content of
gasoline, to a 30 parts per million (ppm) annual refinery average, 80 ppm per-gallon cap, and 95
ppm downstream cap; beginning in 2004 and phasing in through 2011.
The Tier 3 rule is a comprehensive, systems-based approach to address the impact of
motor vehicles on air quality and health, similar to the Tier 2 rule. The Tier 3 program
establishes new standards for light-duty vehicles and trucks and complete heavy-duty vehicles
and new fuel standards for gasoline. Such standards were assumed in the 2008 NAAQS as part
of the strategy for reaching attainment with the NAAQS. Subsequently, a May 21, 2010
Presidential Memorandum directed EPA to "review for adequacy" the current non-greenhouse
gas (GHG) emissions regulations for new motor vehicles and fuels, including tailpipe emissions
standards for NOx and air toxics, and sulfur standards for gasoline. The memo further directed
EPA to "promulgate such regulations as part of a comprehensive approach toward regulating
motor vehicles" if EPA determines new regulations are required. Based on our review, we have
10-2
-------
concluded that improved vehicle technology, combined with lower sulfur gasoline, make it
feasible and cost-effective to reduce emissions well below the current Tier 2 levels. These
emission reductions are necessary to reduce air pollution that is (and projected to continue to be)
at levels that endanger public health and welfare.
In the absence of additional controls such as Tier 3 standards, areas would have to adopt
other measures to reduce emissions from other sources under their state or local authority. Few
other measures exist for providing multi-pollutant reductions of the same magnitude and cost-
effectiveness as those expected from the Tier 3 standards.
10.4 Definition and Description of Small Entities
The Regulatory Flexibility Act defines small entities as including small businesses, small
organizations, and small governmental jurisdictions. For the purposes of assessing the impacts
of a rule on small entities, a small entity is defined as: (1) a small business that meets the
definition for business based on the Small Business Administration's (SBA) size standardsA; (2)
a small governmental jurisdiction that is a government of a city, county, town, school district or
special district with a population of less than 50,000; and (3) a small organization that is any not-
for-profit enterprise which is independently owned and operated and is not dominant in its field.
This rulemaking is expected to affect a variety of small businesses, but will not affect any small
governmental jurisdictions or small organizations as described above. Table 10-1 below
provides an overview of the primary SB A small business categories potentially affected by this
regulation.
Table 10-1 Industry Sectors Potentially Affected by the Rule
Industry Sector
Gasoline fuel refiners and
importers
Ethanol producers
Gasoline additive
manufacturers
Transmix processors
Petroleum bulk stations &
terminals
Other warehousing and
storage-bulk petroleum storage
Light-duty vehicle and light-
duty truck manufacturers
On-highway heavy-duty engine
& vehicle (>8,500 pounds
NAICS Code
324110
325193
325199
325998
424690
Varied
424710
493190
336111,336112
333618,336120,336211
336312
SBA Size Standard for Small
Business (less than or equal to):
1,500 employees
1,000 employees
1,000 employees
500 employees
100 employees
1,500 employees
100 employees
$25.5 million (annual receipts)
1,000 employees
1,000 employees
750 employees
A The SBA definitions of small business by size standards using the North American Industry Classification System
(NAICS) can be found at 13 CFR 121.201.
10-3
-------
GVWR) manufacturers
Independent commercial
importers
Alternative fuel converters
811111,811112,811198
335312
336312,336322,336399
811198
$7 million (annual receipts)
1,000 employees
750 employees
$7 million (annual receipts)
EPA used a variety of sources to identify which entities are appropriately considered
"small" using the criteria for small entities developed by the Small Business Administration as a
guide. Information about these entities came from sources including the Energy Information
Administration (EIA) within the U.S. Department of Energy, oil industry literature, EPA's motor
vehicle certification data, and previous vehicle and fuel rulemakings that have affected these
industries. EPA then found employment information or annual revenue information for these
companies using the business information database Hoover's Online (a subsidiary of Dun and
Bradstreet).
10.5 Summary of Small Entities to Which the Rulemaking Will Apply
10.5.1 Fuels-related Industries
Small entities that may be subject to the Tier 3 gasoline sulfur standards include:
domestic gasoline refiners, importers of gasoline into the U.S., ethanol producers, gasoline
additive manufacturers, transmix processors, and terminal operators. Based on current data, EPA
believes that there are about 50 gasoline refiners, of these, we believe that there are currently 13
refiners producing gasoline that meet SBA's small business definition. Of the seven transmix
processors that we have identified, we believe that five would be considered small entities.
Transmix processors do not have a specific NAICS code, and thus do not have a corresponding
SBA definition, so these parties were estimated to be small entities by using the respective size
standard for the industry these entities had listed as their "primary" business (refining- 1,500
employees or less). For fuel terminals, we believe that there are 1,100 companies; of the 980
companies that we were able to find employee count and/or revenue information for, we believe
that 900 of these companies would be considered small entities. There are approximately 204
ethanol producers; for those companies for which we were able to find employment data, we
believe that the majority of this sector (all but 16 ethanol producers) would be considered small
businesses.
It should be noted that because of the dynamics in the fuels industry (i.e., mergers and
acquisitions), the actual number of refiners that ultimately qualify as a small business under this
program could be different from this initial estimate.
10.5.2 Vehicle-related Industries
The motor vehicle manufacturing industry is made up primarily of large manufacturers
including General Motors, Ford, Toyota, and Honda. Based on EPA certification records, we
have identified a total of 27 car and truck manufacturers which have certified vehicles for sale in
the U.S. Of these companies, EPA has identified 4 motor vehicle manufacturers that qualify as a
small business under SBA definitions. Two of these small entities produce either gasoline-
10-4
-------
fueled, natural-gas fueled, or hybrid gasoline-electric vehicles. Two additional small
manufacturers exclusively produce all-electric vehicles.
Companies that convert motor vehicles to run on alternative fuels will be subject to the
proposed regulations. Based on EPA certification records, we have identified 13 companies
which convert vehicles to run on alternative fuels. Of these companies, EPA has identified 9
alternative fuel converters that qualify as a small business under the SBA definitions.
Another industry sector that will be subject to the proposed regulations consists of
companies that import specialized cars and trucks into the United States (U.S.), referred to as
Independent Commercial Importers (ICIs). ICIs work with customers to bring in cars from
overseas either because the owners are moving to the U.S., or because the vehicle is not
otherwise available in the U.S. (e.g., high-performance sports cars and right-hand drive postal
vehicles). We have identified 8 ICIs that are currently importing cars and trucks into the U.S.
All of these companies qualify as a small business under the SBA definitions.
10.6 Related Federal Rules
The primary federal rules that are related to this final rule are: the Tier 2
Vehicle/Gasoline Sulfur rulemaking (65 FR 6698, February 10, 2000), the 2017 Light-duty
Greenhouse Gas (LD GHG) rule (77 FR 62623), and the Greenhouse Gas Emissions Standards
and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles (HD GHG)
rule (76 FR 57106).
The LD GHG and HD GHG rules are coordinated efforts by EPA and the National
Highway Traffic Safety Administration (NHTSA) taking steps to reduce GHG emissions and
improve fuel efficiency from on-road vehicles and engines.
10.7 Reporting, Recordkeeping, and Other Compliance Requirements
For any fuel control program, EPA must have assurance that fuel produced, distributed,
sold and used meets the applicable standard. The recordkeeping, reporting, and compliance
provisions of the Tier 3 fuels program will be consistent with those currently in place for the Tier
2 gasoline program. Further, to the program will use existing registration and reporting systems
that parties in the fuel production and distribution industry are already familiar with. A complete
discussion of the fuel-related compliance provisions can be found in Section V.F of the preamble
to the final rule.
For any motor vehicle emissions control program, EPA must have assurances that the
regulated products will meet the standards. The final program for manufacturers subject to this
rule will include testing, reporting, and record keeping requirements for manufacturers of
vehicles covered by the Tier 3 regulations. Testing requirements for these manufacturers will
include certification emissions (including deterioration factor) testing and in-use testing.
Reporting requirements will include emissions test data and technical data on the vehicles.
Manufacturers must keep records of this information.
10-5
-------
10.8 Regulatory Alternatives
As a part of the SBREFA process, we conducted outreach to small entities and convened
a SBREFA Panel to gain feedback and advice from these entities. Prior to convening the Panel,
we held outreach meetings with the SERs to learn the needs of small entities and potential
challenges that these entities may face. The feedback that we received from SERs as a result of
these meetings was used during the Panel process to develop a wide range of regulatory
alternatives to mitigate the impacts of the rulemaking on small businesses. It was agreed that
EPA should consider the issues raised by the SERs (and issues raised in the course of the Panel)
and that EPA should consider the comments on flexibility alternatives that would help to
mitigate any negative impacts on small businesses.
The Panel consisted of members from EPA, the Office of Management and Budget
(OMB), and SBA's Office of Advocacy. Following the Panel convening, a Final Panel Report
detailing all of the alternatives that were recommended by the Panel was issued. A full
discussion of the regulatory alternatives discussed and recommended by the Panel, all written
comments received from SERs during the SBREFA process, and summaries of the outreach
meetings held with the SERs can be found in the SBREFA Final Panel Report.2 In the proposal,
we either proposed or requested comment on the various recommendations put forth by the
Panel. Below we discuss those flexibility options recommended in the Panel Report, our
proposed regulatory alternatives, and those provisions which are being finalized. All of the
flexibilities that are being finalized for small businesses, as well as those for all entities that may
be affected by the rulemaking, are described in Sections IV and V of the preamble to the final
rule.
10.8.1 Fuel-related Alternatives
10.8.1.1 Delayed Standards for Small Refiners
Panel Recommendations
The Panel recommended that EPA propose a delay option, similar to previous fuels
rulemakings, in the Tier 3 proposed rule. The Panel recommended that EPA allow small refiners
to postpone their compliance with the Tier 3 program for up to three years. Small refiners
choosing this flexibility option would have from January 1, 2017 through December 31, 2019 to
continue production of gasoline with an average sulfur level of 30 ppm (per the Tier 2 gasoline
sulfur program), and compliance with the 10 ppm sulfur standard would begin on January 1,
2020. Any small refiner choosing this proposed option would be allowed to continue use of their
Tier 2 gasoline sulfur credits through December 31, 2019 to meet the refiner average 30 ppm
sulfur standard.
The Panel also recommended that EPA request comment on case-by-case hardship
provisions that would provide additional relief for any refiner experiencing extreme difficulty in
compliance with the Tier 3 requirements, as discussed below in Section 10.8.1.1.4.
10-6
-------
What We Proposed and Public Comments Received on the NPRM
We proposed to allow small refiners to postpone their compliance with the Tier 3
program for up to three years—small refiners choosing this flexibility option would have from
January 1, 2017 through December 31, 2019 to continue production of gasoline with an average
sulfur level of 30 ppm (per the Tier 2 gasoline sulfur program). Compliance with the 10 ppm
sulfur standard would begin on January 1, 2020. We also proposed that small refiners would be
allowed to continue use of their Tier 2 gasoline sulfur credits through December 31, 2019 to
meet their refinery average 30 ppm sulfur standard, but these credits could not be used for
compliance with the proposed Tier 310 ppm average sulfur standard. With regard to early credit
generation, we proposed that all refiners and importers (including small refiners) could generate
early credits relative to the 30 ppm sulfur standard from January 1, 2014 through December 31,
2016, and relative to the 10 ppm sulfur standard from January 1, 2017 through December 31,
2019. We also proposed to extend the small refiner provisions to small volume refineries
(refineries with a crude oil throughput of less than or equal to 75,000 barrels per calendar day
(bpcd)).
Comments received on the proposed rule were generally in support of allowing an
additional three years for small refiners (and small volume refineries) to comply with the Tier 3
program. However, some commenters did not agree with our proposed provisions for early
credits. Some small refiners commented that they believe small refiners should also have their
own three-year early credit generation period, from January 1, 2017 through December 31, 2019,
relative to the 30 ppm sulfur standard.
What We're Finalizing
As described in Section V.E. of the preamble to the final rule, we are finalizing a three-
year delay for approved small refiners, until January 1, 2020. With regard to the ABT program,
as discussed in preamble Section V.D.5, we are finalizing that small refiners may generate early
credits relative to the 30 ppm average sulfur standard from January 1, 2014 through December
31, 2019, and these credits could be used for either Tier 2 or Tier 3 compliance. (However,
credits generated by a small refiner from January 1, 2017 through December 31, 2019 could only
be traded and used by other small refiners.) Further, from January 1, 2017 through December
31, 2019, if a small refiner's annual average sulfur level is below 10 ppm, they may elect to split
the generation of credits between both the 10 ppm and 30 ppm standards (without double-
counting). For example, during this time, a small refiner with an annual gasoline sulfur average
of 8 ppm could generate 20 Tier 2 credits (30 ppm-10 ppm) and 2 Tier 3 credits (10 ppm-8 ppm).
All of the small refiner provisions are also applicable to approved small volume
refineries.
10.8.1.2 Provisions for Additive Manufacturers
Panel Recommendations
During the SBREFA Panel process, different requirements than those proposed (and
being finalized today) were discussed for additive manufacturers. Thus, the provisions
10-7
-------
recommended by the Panel were not applicable to the provisions proposed and now finalized for
these parties. More information on the Panel's recommendations for gasoline additive
manufacturers can be found in the Final Panel Report, located in the rulemaking docket.
What We Proposed and Public Comments Received on the NPRM
We proposed that parties introducing additives to gasoline greater than 1.0 volume
percent would be required to satisfy all of the obligations of a fuel manufacturer, including
demonstration that the finished blend meets the applicable sulfur specification. We also
proposed a maximum sulfur contribution of 3 ppm from the use of a gasoline additive added
downstream of the refinery at less than 1.0 volume percent (when added at the maximum
recommended treatment rate). Lastly, we proposed that additive manufacturers would need to
maintain records of their additive production quality control activities for five years.
What We're Finalizing
As discussed further in Section V.C., manufacturers of gasoline additives that are used
downstream of the refinery at less than 1.0 volume percent will be required to limit the sulfur
contribution to the finished gasoline from the use of the additive to less than 3 ppm when the
additive is used at the maximum recommended treatment rate. For each batch of additive
produced, the manufacturer must retain sulfur test records for 5 years, and must make these
records available to EPA upon request. Parties that introduce additives to gasoline at over 1.0
volume percent will be required to satisfy all of the obligations of a fuel manufacturer, including
demonstration that the finished blend meets the applicable sulfur specification.
10.8.1.3 Refinery Gate and Downstream Caps
Panel Recommendations
The Panel recommended that EPA assess and request comment on retaining the current
Tier 2 refinery gate and downstream caps of 80 and 95 ppm, respectively, to help provide
maximum flexibility and avoid system upsets for the entire refining and distribution system.
Further, the SBA and OMB Panel members recommended that EPA propose retaining the 80
ppm and 95 ppm caps.
With regard to a 20 ppm refinery gate cap, the Panel had concerns that such a standard
could cause operational problems for small refiners during a refinery turnaround or an upset,
because a cap of this level could result in a refiner not being able to produce gasoline (as noted in
their comments in Section 8 of the Panel Report). The Panel likewise had concerns that a
downstream cap of 25 ppm could cause problems for small downstream entities, such as
transmix processors, because they may not be able to reprocess finished gasoline down to this
level (also noted in their comments in Section 8 of the Panel Report). Thus, the Panel
recommended that EPA request comment on additional refinery gate and downstream caps above
20/25 ppm, but below 80/95 ppm. Additionally, the Panel recommended that EPA allow the
current Tier 2 80 ppm sulfur refinery gate cap and 95 ppm sulfur downstream cap in Alaska to
remain at these levels indefinitely.
10-8
-------
What We Proposed and Public Comments Received on the NPRM
We proposed two options for the per-gallon sulfur caps—maintaining the Tier 2 80 ppm
refinery gate sulfur and 95 ppm downstream sulfur caps and, beginning January 1, 2020,
lowering to 50 ppm refinery gate and 65 ppm downstream caps. In addition, we requested
comment on lowering to a 20 ppm refinery gate cap and a 25 ppm downstream cap. Due to
technical and economic concerns for refiners and distributors in the state of Alaska, we also
proposed that if we were to finalize 50/65 ppm caps, Alaska would be allowed to retain the 80/95
ppm caps.
We received comments on both of the proposed per-gallon cap options of 80/95 ppm and
50/65 ppm, as well as comments on finalizing lower caps of 20/25 ppm and a 20 ppm overall
cap. Comments in support of lower refinery gate and downstream caps noted potential
environmental benefits, greater certainty that vehicles would see lower and more uniform
gasoline sulfur levels, and the ability to enable new vehicle technologies that require very low
sulfur levels. Comments in support of maintaining the 80/95 ppm caps cited concerns over cost,
flexibility during turnarounds/unplanned shut downs (due to refinery fire, natural disaster, etc.),
and potential impacts on gasoline supply and pricing.
What We're Finalizing
As discussed in greater detail in Section V.C of the preamble, we are retaining the Tier 2
per-gallon sulfur caps of 80 ppm at the refinery gate and 95 ppm downstream. We are also
committing to monitor and further evaluate in-use sulfur levels and their impact on vehicle
emissions, which will include: analyses of in-use fuel surveys, refinery batch data, and the sulfur
credit market; evaluation of any implementation issues; and an ongoing evaluation of vehicle
manufacturers' data on Tier 3 vehicles' in-use performance. If it is warranted, we will reassess
the sulfur cap level and the need for potential future regulatory action.
10.8.1.4 Hardship Provisions
Panel Recommendations
During the Panel process, EPA stated its intent to propose hardship provisions (for all
gasoline refiners and importers) similar to those in prior EPA fuels programs: a) the extreme
unforeseen circumstances hardship provision, and b) the extreme hardship provision. A hardship
based on extreme unforeseen circumstances is intended to provide short term relief due to
unanticipated circumstances beyond the control of the refiner, such as a natural disaster or a
refinery fire. An extreme hardship is intended to provide short-term relief based on extreme
circumstances (e.g., extreme financial problems, extreme operational or technical problems, etc.)
that impose extreme hardship and thus significantly affect a refiner's ability to comply with the
program requirements by the applicable dates. In the context of the proposal, the Panel agrees
that such relief could consider long-term relief on the sulfur cap (similar to that for Alaska) if the
circumstances both warrant it and can be structured in a way to allow for it. The Panel agrees
with the proposal of such provisions and recommended that EPA include them in the Tier 3
proposed rulemaking.
10-9
-------
What We Proposed and Public Comments Received on the NPRM
Similar to previous EPA fuels programs, we proposed the extreme hardship and extreme
unforeseen circumstances hardship provisions both to accommodate a refiner's inability to
comply with at the start of the Tier 3 program and to deal with unforeseen circumstances that
may occur at any point during the program. We proposed that any refiner could apply for a
hardship waiver, but relief would be granted on a case-by-case basis following a showing of
certain requirements; primarily that compliance through the use of credits is not feasible.
Further, we proposed that any hardship waiver granted would likely consist of short-term relief,
and would be based on the nature and degree of the hardship and EPA's assessment of the credit
market at that time.
In general, comments received agreed with the inclusion of hardship provisions in the
Tier 3 program. While not directly related to the hardship provisions, we received comments on
the one-year deficit carryforward provision (an individual refinery that does not meet the average
sulfur standard in a given year may carry a credit deficit forward for 1 year). We received
comments expressing concern that it will be more challenging for refineries to make up their
credit deficit in one year with a 10 ppm sulfur standard, thus the commenters requested that the
deficit carryforward allowance be extended to two or three years. One commenter suggested that
an extension of the deficit carryforward provision could also be used as a form of hardship relief.
What We're Finalizing
We are finalizing both the extreme hardship and extreme unforeseen circumstances
hardship provisions. We continue to believe that providing short-term relief to those refiners that
need additional time due to hardship circumstances will help to facilitate the adoption of the
overall Tier 3 program for the majority of the industry. The provisions themselves, and the
conditions under which a refiner would receive hardship relief, are similar to those in previous
fuels regulations, and are necessary and appropriate to ensure that any waivers granted would be
limited in scope. Further, we expect to impose appropriate conditions to ensure that the refiner is
making best efforts to achieve compliance offsetting any loss of emission control from the
program. A complete discussion of the hardship provisions is located in Section V.E.2 of the
preamble to the final rule.
As discussed more fully in preamble Section V.D.7, while we acknowledge the increased
hurdle to make up a deficit in the Tier 3 program, we have concerns with the enforceability
allowing for deficit carryforward greater than one year. Further, refiners also have the
opportunity to purchase credits from others. However, if for some reason credits are unavailable
or are prohibitively expensive such that the refiner could not make up the deficit in one year, we
would consider this in an evaluation of a hardship application. As such, we are finalizing that
hardship waivers could grant relief in the form of additional deficit carryforward of up to three
years, depending on the level of hardship and the status of the credit market.
10-10
-------
10.8.2 Vehicle-Related Alternatives
10.8.2.1 Lead Time for Exhaust and Evaporative Emission Standards
Panel Recommendations
In the types of businesses subject to the potential Tier 3 standards, small businesses have
limited resources available for developing new designs to comply with new emission standards.
In addition, it is often necessary for these businesses to rely on vendor companies for technology.
Moreover, percentage phase-in requirements pose a dilemma for a small manufacturer that has a
limited product line (e.g., the manufacturer certifies vehicles in only one or two test groups).
Thus, similar to the flexibility provisions implemented in previous vehicle rules, the Panel
recommended that EPA allow small businesses the following flexibility options for meeting the
potential Tier 3 exhaust emissions standards.
The Panel recommended that small businesses be given additional leadtime to comply
with the potential Tier 3 exhaust standards and allow small businesses to comply with the
standards with 100 percent of their vehicles starting in model year 2022. (This is similar to the
Tier 2 rule where EPA allowed small manufacturers to wait until the end of the phase-in to
comply with the Tier 2 standards.) During the Panel process, the proposed Tier 3 rule was
expected to have several different phase-in schedules; with the final dates varying from model
year 2021 for the new exhaust PM standards and use of the new El 5 certification fuel, to model
year 2022 for the new evaporative emission standards, to model year 2025 for the new exhaust
gaseous pollutant standards. The Panel noted that requiring all small businesses to comply with
the full slate of Tier 3 requirements in model year 2022 should provide sufficient lead time for
manufacturers to plan for and implement the technology changes needed to comply with the Tier
3 standards.
One of the SERs recommended that EPA adopt relaxed exhaust standards for small
manufacturers. The SER noted that the exhaust emission averaging program being proposed by
EPA would allow large manufacturers that have many engine families to certify their small,
niche products at levels numerically higher than the standards. Small manufacturers that
typically do not have more than one or two emission families generally cannot use averaging to
the same extent because of their limited product offerings. The SER's concern was that the high-
performance vehicles produced by large manufacturers which they compete against would be
able to certify at numerically higher levels at less cost than the SER would incur. While EPA
was planning to propose the same standards for all manufacturers, the Panel recommended that
EPA request comment on allowing small manufacturers to meet relaxed exhaust emission
standards. This could also be included as part of the hardship provision discussed below. The
Panel recommended that EPA request comment on the relaxed standards recommended by the
SER. The SER-recommended relaxed NMOG+NOx standards over the Federal Test Procedure
(FTP) are 0.125 grams/mile in model year 2020 and 0.070 grams/mile in model year 2025. In
addition, the Supplemental FTP standards would be the standards for the corresponding bins
which the manufacturer selected for complying with the FTP standards. For example, if the
manufacturer certified to the proposed Tier 3 Bin 125 standards over the FTP, the manufacturer
would have to comply with the corresponding Tier 3 Bin 125 standards for the Supplemental
FTP.
10-11
-------
The Panel recommended that small businesses comply with the Tier 3 evaporative
emission standards, including the leak standard, with 100 percent of their vehicles starting in
model year 2022. For evaporative emissions, where the Tier 3 standards could begin as early as
2017 and phase-in through 2022, this provision would allow small businesses and SVMs to wait
until the last year of the Tier 3 phase-in period for evaporative emission standards for all of their
vehicles. This start date is consistent with the start date described above for the Tier 3 exhaust
emission requirements being recommended by the Panel for small businesses.
What We Proposed
We proposed that small businesses be given additional lead time to comply with the
potential Tier 3 exhaust and evaporative emission standards and that small businesses be allowed
to comply with the standards with 100 percent of their vehicles starting in model year 2022.
(This is similar to the Tier 2 rule where EPA allowed small manufacturers to wait until the end of
the phase-in to comply with the Tier 2 standards.) The proposed Tier 3 rule has several different
phase-in schedules; with the final dates varying from model year 2021 for the new light-duty
exhaust PM standards to model year 2025 for the new light-duty exhaust gaseous pollutant
standards. Our assessment was that requiring all small businesses to comply with the full slate of
Tier 3 requirements in model year 2022 would provide sufficient lead time for manufacturers to
plan for and implement the technology changes needed to comply with the Tier 3 standards.
Although we proposed the same standards for all manufacturers starting with model year
2022, we also requested comment on allowing small manufacturers to meet relaxed exhaust
emission standards. The relaxed standards could be written directly into the regulations, or
potentially we could allow manufacturers to request relaxed exhaust standards as part of the
hardship provision discussed below, or we could allow manufacturers to request alternative
standards based on a comparison of vehicles with similar attributes that are certified by larger
manufacturers. With regard to the relaxed standard, we are requested comment on the standards
recommended by the SER. The SER-recommended relaxed NMOG+NOx standards over the
Federal Test Procedure (FTP) are 0.125 grams/mile in model year 2020 and 0.070 grams/mile in
model year 2025. In addition, the Supplemental FTP standards would be the standards for the
corresponding bins which the manufacturer selected for complying with the FTP standards. For
example, if the manufacturer certified to the proposed Tier 3 Bin 125 standards over the FTP, the
manufacturer would have to comply with the corresponding Tier 3 Bin 125 standards for the
Supplemental FTP.
Public Comments Received on the NPRM and What We Are Finalizing
We did not receive comments from non-SVM small businesses subject to the Tier 3
vehicle standards about our proposed small entity phase-in provisions. However, we received
comments from SVMs, as well as from the Alliance of Automobile Manufacturers and the
Association of Global Automakers, arguing that the proposed phase-in did not provide adequate
lead time relief for SVMs, and that the long-term Tier 3 standards for light-duty vehicles are not
technologically feasible for SVMs. They especially highlighted the ability of large
manufacturers to offset high emissions from high-performance, luxury models by averaging with
their low-emitting models, while competing SVM products must be designed to actually achieve
low emissions while still meeting customers' performance expectations. Their limited
10-12
-------
production can also result in emission control technology suppliers placing a lower priority on
SVM orders than on those of larger, high-volume manufacturers.
Because of these factors, SVMs suggested that their companies meet a slightly more
stringent NMOG+NOx standard than what we proposed for SVMs in the early years of the
program (125 mg/mi) and a permanently relaxed standard of 51 mg/mi beginning in MY 2022.
Ferrari (which also raised the issue of "operational independence" as described below) suggested
a compliance schedule for SVMs similar to the California LEV III program, with either a
permanently relaxed standard (matching the California LEV III 70 mg/mi long-term standard) or
a delay until MY 2030 to meet the primary 30 mg/mi Tier 3 standard (when they suggest that
SVMs could potentially comply). CARB comments supported Tier 3 adoption of its LEV III
provisions for SVMs, including the long-term 70 mg/mi standard beginning in MY 2025.
After considering the comments, we agree with SVMs that their unique logistical and
technological challenges, especially in the later years of the primary phase-in schedule, warrant a
significant period of relaxed standards for these manufacturers. However, we see no
fundamental reasons why, given sufficient lead time, all companies, regardless of company size
and vehicle characteristics, will not be able to meet the Tier 3 standards. Thus, we are finalizing
a program for SVMs, available to non-SVM small businesses as well, under which they can
choose an alternative NMOG+NOx fleet average phase-in schedule: Meeting a standard of 125
mg/mi for model years 2017 through 2021, meeting a standard of 51 mg/mi for model years
2022 through 2027, and then meeting the final Tier 3 standard of 30 mg/mi thereafter.
Because companies choosing this 3-stage compliance option are certifying to Tier 3 bin
standards in MY 2017, we expect that other exhaust emissions standards, including SFTP and
PM standards, would apply for their vehicles as well, to the same degree and on the same
schedule as for other manufacturers. Application of evaporative emissions and onboard
diagnostics (OBD) standards, on the other hand, is not affected by choice of the 3-stage
compliance option, and small companies may separately choose to delay compliance with
evaporative emissions and OBD standards (except as noted in Section IV.G.3) until MY 2022, as
proposed. In addition, small companies choosing the 3-stage compliance option may also delay
the longer useful life and new test fuel requirements for exhaust emissions standards until MY
2022 to align these changes with the 3-stage schedule. This option would not preclude use of
other applicable small entity flexibility provisions discussed in this subsection.
Although we are adopting this revised implementation schedule for SVMs and small
businesses, we believe the proposed approach of allowing postponement of Tier 3 compliance
until MY 2022 may be useful for small companies needing more lead time to begin certifying
Tier 3 vehicles. Therefore we are finalizing the proposed approach as an additional but separate
option for such companies, including SVMs, ICIs, and alternative fuel vehicle converters.
Furthermore, because the optional 3-stage SVM implementation schedule, and the record of
comments that prompted it, are specific to the light-duty sector, we are not extending it to heavy-
duty vehicles and instead are finalizing only the proposed approach of allowing postponement of
Tier 3 compliance until MY 2022 for any SVMs and small businesses in the heavy-duty sector.
Companies that take advantage of one of the SVM and small business implementation
schedule provisions in either the light-duty or heavy-duty sector are not allowed to generate or
10-13
-------
use exhaust emissions Tier 3 credits in that sector while doing so. That is, they cannot earn or
use exhaust emissions Tier 3 credits before MY 2027 while using the light-duty SVM revised
implementation schedule, and they also cannot do so before MY 2022 while using the postponed
compliance schedule that we proposed.
10.8.2.2 Assigned Deterioration Factors
Panel Recommendations
Under EPA's regulations, manufacturers must demonstrate that their vehicles comply
with the emission standards throughout the "useful life" period. This is generally done by testing
vehicles at low-mileage and then applying a deterioration factor to these emission levels. The
deterioration factors are determined by aging new emission control systems and then testing the
aged systems again to determine how much deterioration in emissions has occurred. In order to
reduce the testing burden on small manufacturers, EPA suggested that small manufacturers could
use deterioration factor values assigned by EPA instead of performing the extended testing. A
manufacturer would apply the assigned deterioration factors to its low-mileage emission level to
demonstrate whether it complied with the Tier 3 emission standards. EPA currently allows this
flexibility for small manufacturers. The Panel recommended that EPA propose that small
businesses be allowed the option to use EPA-developed assigned deterioration factors in
demonstrating compliance with the Tier 3 exhaust and evaporative emission standards. In the
past, EPA has relied on deterioration factor data from large manufacturers to develop the
assigned DFs for small manufacturers. EPA would expect to follow a similar procedure to
determine the assigned DFs for the Tier 3 standards once large manufacturers start certifying
their Tier 3 designs. Given that larger manufacturers would begin phasing in to the Tier 3
standards in model year 2017, EPA should have a significant set of emissions deterioration data
upon which to base the assigned DFs for small businesses within the first few years of the Tier 3
program. EPA recognizes that assigned DFs need to be determined well in advance of model
year 2022 in order to provide sufficient time for small businesses to decide whether or not to use
the assigned DFs for certification purposes.
What We Proposed
We proposed that small businesses be allowed the option to use EPA-developed assigned
deterioration factors in demonstrating compliance with the Tier 3 exhaust and evaporative
emission standards. In the past, EPA has relied on deterioration factor data from large
manufacturers to develop the assigned deterioration factors for small manufacturers. EPA would
expect to follow a similar procedure to determine the assigned deterioration factors for the Tier 3
standards once large manufacturers start certifying their Tier 3 designs. Given that larger
manufacturers would begin phasing in to the Tier 3 standards in model year 2017, we expected
that we would have a significant set of emissions deterioration data upon which to base the
assigned deterioration factors for small businesses within the first few years of the Tier 3
program, we recognized in the proposal that assigned deterioration factors need to be
determined well in advance of model year 2022 in order to provide sufficient time for small
businesses to decide whether or not to use the assigned deterioration factors for certification
purposes.
10-14
-------
Public Comments Received on the NPRM and What We Are Finalizing
We are adopting the assigned deterioration factor provisions for small businesses and
SVMs (as well as for small volume test groups), as proposed. Commenters expressed support,
but asked that the Agency commit itself to keeping these factors up to date as durability data
accumulates. We are committed to periodically updating and publishing these assigned
deterioration factors. Given that SVMs will now be allowed to use the revised implementation
schedule, starting in MY 2017, it becomes necessary to consider assigned deterioration factors in
stages. Because there may not be a sufficient base of accumulated durability data on Tier 3
vehicles by MY 2017, we expect that the current set of assigned factors based on Tier 2 vehicles
may continue in place for some time, noting that the MY 2017-2021 SVM fleet average of 125
mg/mi is not too much different from the average of today's Tier 2 vehicle emissions. By MY
2022, when the SVM NMOG+NOx fleet average standard drops to 51 mg/mile, we expect to
have new assigned factors available. We note that small businesses and SVMs may also, with
advance EPA approval, use deterioration factors developed by another manufacturer (40CFR
86.1826-0 l(b)).
10.8.2.3 Reduced Testing Burden and OBD Requirements
Panel Recommendations
Under EPA's regulations, manufacturers must perform in-use testing on their vehicles
and demonstrate their in-use vehicles comply with the emission standards. The current in-use
testing regulations provide for reduced levels of testing for small manufacturers, including no
testing in some cases. EPA suggested that these provisions should continue for small
manufacturers with the Tier 3 program. The Panel recommended that EPA propose that small
businesses be allowed to have reduced burden under the in-use testing program for Tier 3
vehicles.
One SER requested that EPA eliminate some of the evaporative emission testing
requirements for small businesses based on its belief that some of the tests may be duplicative.
While EPA noted (during the Panel process) that it understood the reasons behind the
manufacturer's suggestion, EPA believed it may be premature to consider such an option in the
Tier 3 rule given the impact of the CC>2 emission standards on engine and fuel system
development. Currently, it is generally understood that the 2-day diurnal test drives the purge
characteristics of evaporative control systems, while the refueling test, and to a lesser degree the
3-day test, drive the capacity requirement of evaporative canisters. Prospectively, due to
expected changes in engine and fuel system designs in response to upcoming CC>2 emission
standard requirements, this may not be the case. Therefore, at the time of the Panel process,
EPA noted its belief that it is appropriate to retain all of the evaporative test procedures. It can
be noted that under current regulations, EPA does allow manufacturers to waive 2-day diurnal
testing for certification purposes (see 40 CFR 86.1829-0l(b)(2)(iii)) and perform only the 2-day
diurnal test as part of the in-use testing program (see 40 CFR 86.1845-04(c)(5)(ii)). These
provisions would continue in the Tier 3 program. In general, EPA noted that it is open to
changes that reduce test burden while maintaining the environmental effectiveness of its
programs and could consider changes like those suggested by the SER in the future as the
impacts of the future regulations on engine and vehicle design become clearer. EPA also stated
10-15
-------
that it intends to request comment in the Tier 3 proposal on streamlining the current test
procedures for small businesses in ways that would still maintain the overall stringency of the
tests.
What We Proposed
Under EPA's regulations, manufacturers must perform in-use testing on their vehicles
and demonstrate their in-use vehicles comply with the emission standards. The current in-use
testing regulations provide for reduced levels of testing for small manufacturers, including no
testing in some cases. We proposed that small businesses be allowed to have reduced burden
under the in-use testing program for Tier 3 vehicles. Small manufacturers that sell less than
5,000 units per year would not be required to do any in-use testing. Small manufacturers that
sell between 5,001 and 15,000 units per year would be required to test 2 vehicles per test group,
but only under the high-mileage conditions specified in the program.
Under current regulations, manufacturers may waive testing for PM emissions for light-
duty vehicles and trucks, except for diesel-fueled vehicles. Manufacturers are still subject to the
standards and must make a statement of compliance with the PM standards. With the Tier 3
proposal, we proposed new PM standards and further proposed to require manufacturers to test
for PM emissions for all fuels. Because PM testing requires additional test equipment and
facilities, the costs incurred for PM testing can be substantial, especially for a company selling
small numbers of vehicles. Therefore, we proposed to continue the waiver for PM testing in the
Tier 3 timeframe for small businesses. Small businesses would not be required to measure PM
emissions when they certify to the Tier 3 emission standards. In lieu of testing, small businesses
would be required to make a statement of compliance with the Tier 3 PM standards. We would
retain the ability to determine the PM emissions results in confirmatory or in-use testing.
We proposed new OBD requirements for vehicles certifying to the Tier 3 standards. The
proposed OBD requirements were the same as CARS's existing OBD requirements. The
proposed OBD provisions require additional amounts of testing and information that can add
significant cost to manufacturers if they are not already meeting the CARB OBD requirements.
Small business vehicle manufacturers tend to comply with the CARB OBD requirements
because they want to sell in the California market. On the other hand, alternative-fuel converters
do not generally certify with CARB because of the significant cost burden of complying with the
CARB OBD requirements. We therefore proposed that small business alternative-fuel
converters may continue to comply with EPA's existing OBD requirements (see 40 CFR
86.1806-05) when the Tier 3 standards become effective.
Alternative-fueled vehicles, MDPVs, and FFVs do not have SFTP emissions
requirements under the existing regulations. We proposed to apply the Tier 3 SFTP standards to
all vehicles, including alternative-fueled vehicles, MPDVs, and FFVs. Because SFTP testing
includes emission measurement over the SC03 test cycle, which requires additional test facilities
beyond those needed to run the FTP, the costs incurred for SC03 testing can be substantial,
especially for companies like alternative fuel converters that sell very low numbers of converted
vehicles. We proposed that the categories of vehicles newly subject to the SFTP standards,
including alternative-fuel conversions, have the option to substitute the FTP emissions levels for
the SC03 emissions results for purposes of compliance when calculating the SFTP emissions.
10-16
-------
However, we would retain the ability to determine the composite emissions using SC03 test
results in confirmatory or in-use testing. Because the vehicles being converted to an alternative
fuel will likely have been tested for SFTP compliance, we expressed the view that the SFTP
emissions would be similarly low, and therefore the added SC03 testing burden is unnecessary.
Public Comments Received on the NPRM and What We Are Finalizing
We received no adverse comments on our proposal to continue providing for reduced
levels of in-use testing and for waiving of PM testing for SVMs and small businesses, and are
retaining these existing provisions in Tier 3. In lieu of PM testing, these companies will be
required to make a statement of compliance with the Tier 3 PM standards. We may however
measure PM emissions in EPA confirmatory or in-use testing.
VNG, a natural gas fuel network provider, objected that the proposed OBD exception
disadvantages larger manufacturers and should be made equally available to all vehicle
manufacturers' small volume test groups. We expect that larger manufacturers (as well as
SVMs) wishing to produce alternative fuel vehicles will be well familiar with CARS's OBD
requirements and will be well-positioned to implement these requirements in Tier 3. We note
that larger OEMs themselves did not make an argument for extending this provision beyond
converters. We are finalizing the exception to the Tier 3 OBD requirements as proposed. We
further note that the optional delay in Tier 3 implementation until MY 2022 that is available to
small businesses, discussed above, includes a delay in the Tier 3 OBD requirement to MY 2022,
as proposed, except that vehicles already meeting this requirement in MY 2017 must continue to
do so in subsequent years. We are also adopting this Tier 3 OBD delay to MY 2022 for small
companies taking advantage of the revised light-duty 3-stage implementation schedule discussed
above, even though it involves other Tier 3 requirements starting in MY 2017, in order to avoid
overburdening these manufacturers with multiple sets of new OBD design constraints.
10.8.2.4 Hardship Relief Provisions
Panel Recommendations
The Panel recommended that hardship provisions be provided to small businesses for the
Tier 3 exhaust and evaporative emission standards. Under the hardship provisions, small
businesses would be allowed to apply for additional time to meet the 100 percent phase-in
requirements for exhaust and evaporative emissions. All hardship requests would be subject to
EPA review and approval. Appeals for such hardship relief would be required to be made in
writing and submitted well before the earliest date of noncompliance. The request should
identify how much time is being requested. It must also include evidence that the
noncompliance would occur despite the manufacturer's best efforts to comply, and must contain
evidence that severe economic hardship would be faced by the company if the relief is not
granted. The above provision should effectively provide the opportunity for small businesses to
obtain more time to comply with the new Tier 3 standards. (The existing hardship provisions
limit the extra time that can be requested to 1 year, but such a limit may or may not be included
in the proposed Tier 3 hardship provisions.)
10-17
-------
What We Proposed
We proposed hardship relief provisions for small businesses subject to the Tier 3 exhaust
and evaporative emission standards. Under the proposed hardship provisions, small businesses
would be allowed to apply for additional time to meet the 100 percent phase-in requirements for
exhaust and evaporative emissions. All hardship relief requests would be subject to EPA review
and approval. Appeals for such hardship relief would need to be made in writing and must be
submitted well before the earliest date of potential noncompliance. The request would need to
identify how much time is being requested. It must also include evidence that the
noncompliance would occur despite the manufacturer's best efforts to comply, and must contain
evidence that severe economic hardship would be faced by the company if the relief is not
granted. The proposed hardship relief provision would effectively provide the opportunity for
small businesses to obtain more time to comply with the new Tier 3 standards. The existing
hardship relief provisions limit the extra time that can be requested to 1 year, but we proposed
that such a limit is not needed as part of the Tier 3 hardship relief provisions.
Public Comments Received on the NPRM and What We Are Finalizing
Commenters supported these proposed provisions, within the context of a revised
approach to SVM lead time, discussed above. We are finalizing the provisions as proposed.
10.8.2.5 Applicability of Flexibilities
Panel Recommendations
Under EPA's current Tier 2 regulations, EPA provides a number of flexibilities for small
volume manufacturers. The criteria for determining if a company is a small volume
manufacturer is based on the annual production level of vehicles and is based on whether the
company produces less than 15,000 vehicles per year. Unlike EPA's small volume manufacturer
criteria noted above, SBA defines which manufacturers are small businesses (and therefore
should be considered under the SB AR Panel process) based on the number of employees for
vehicle manufacturers and annual revenues for Ids and alternative fuel converters. For
example, SBA defines small business vehicle manufacturers as those who have less than 1,000
employees. Similarly, SBA defines small business ICIs as those who have annual revenues of
less than $7 million per year.
The Panel recommended that EPA propose to allow all small businesses that meet the
SBA criteria be eligible for the flexibilities described above. In addition, in the Panel Report,
EPA stated that it is expecting to propose that manufacturers that meet a specified sales-based
criterion to be eligible for the flexibilities described above. It is relatively easy for a
manufacturer to project and ultimately determine sales. Determining the annual revenues or
number of employees is less straightforward. In the recent rule setting the first light-duty vehicle
and truck CC>2 emission standards, EPA adopted provisions for small manufacturers based on a
sales cutoff of 5,000 vehicles per year as opposed to the 15,000 level noted earlier that is used in
the Tier 2 program. In the Panel Report, EPA noted that it expects to propose a small volume
manufacturer definition based on the 5,000 vehicle per year level for the Tier 3 program. EPA
believes the 5,000 unit cut-off for small volume manufacturers would include all of the small
10-18
-------
business vehicle manufacturers, ICIs, and alternative fuel converters that meet the applicable
SBA definition as well as some additional companies that have similar concerns to small
businesses. Lastly, EPA noted in the Panel Report that it expects to propose the flexibilities
described above to be available to any manufacturer that meets either the SBA small business
criteria or the sales-based criteria.
What We Proposed
Under EPA's Tier 2 regulations, EPA provides a number of flexibilities for small volume
manufacturers. The criteria for determining if a company is a "small volume manufacturer" is
based on the annual production level of vehicles and is based on whether the company produces
less than 15,000 vehicles per year. Unlike EPA's existing small volume manufacturer criteria,
the Small Business Administration (SBA) defines which manufacturers are small businesses
based on the number of employees for vehicle manufacturers and annual revenues for ICIs and
alternative fuel converters. For example, SBA defines a small business vehicle manufacturer as
those who have less than 1,000 employees.
We proposed that all small businesses that are subject to the Tier 3 standards and that
meet the SBA criteria be eligible for the flexibilities described above. Unless otherwise noted,
the proposed flexibilities would be available to all small business vehicle manufacturers, ICIs,
and alternative fuel converters subject to the Tier 3 standards. In addition, we proposed that
manufacturers subject to the Tier 3 standards which meet a specified sales-based criteria be
eligible for the flexibilities described above. It is relatively easy for a manufacturer to project
and ultimately determine sales. Determining the annual revenues or number of employees is less
straightforward. In the recent rule setting the first light-duty vehicle and truck CC>2 emission
standards, EPA adopted provisions for small manufacturers based on a sales cutoff of 5,000
vehicles per year as opposed to the 15,000 level noted earlier that is used in the Tier 2 program.6
We proposed that the small volume manufacturer definition be based on the 5,000 vehicle per
year level for the Tier 3 program. For purposes of the Tier 3 rule, the 5,000 limit would be based
on a running three-year average of the number of light-duty vehicles, light-duty trucks, medium-
duty passenger vehicles, and complete heavy-duty trucks below 14,000 pounds GVWR. We
expressed the belief that the 5,000 unit cut-off for small volume manufacturers would include all
of the small entity vehicle manufacturers, ICIs, and alternative fuel converters that currently
meet the applicable SBA definition, as well as a few additional companies that have similar
concerns to small businesses.
We requested comment on extending eligibility for the Tier 3 SVM provisions to small
manufacturers that are owned by large manufacturers but are able to demonstrate that they are
operationally independent. We established such a provision in the light-duty greenhouse gas
(GHG) program, and CARB did so in LEV III as well.
See 75 FR 25324, May 7, 2010.
10-19
-------
Public Comments Received on the NPRM and What We Are Finalizing
As proposed, we are using the federal Small Business Administration (SBA) criteria to
define small businesses eligible for the special provisions. SBA defines small business vehicle
manufacturers as those with less than 1,000 employees, and small business ICIs and alternative
fuel vehicle converters are evaluated using SBA criteria based on annual revenues. See Section
IV.H.3 for a discussion of additional provisions that apply specifically to ICIs. Also, as
proposed, we are defining SVMs in 40 CFR 86.1838-01 for purposes of Tier 3 as companies
with nationwide annual U.S. sales volumes at or below 5,000 vehicles, though the 15,000 vehicle
threshold used in Tier 2 continues to apply in a few regulatory provisions that Tier 3 changes are
not impacting. Eligibility will be evaluated using an average of 2012-2014 MY sales. For
companies with no 2012 MY sales, projected sales may be used, but their eligibility will be re-
evaluated thereafter using a three-year running average.
VNG commented that the proposed 5,000 vehicle threshold could potentially limit the
ability (or willingness) of natural gas SVMs to scale up production by forcing a tradeoff between
sales and regulatory burden, pointing also to the fact that 15,000 vehicles is only 0.1% of annual
light-duty vehicle sales. We do not believe that the SVM relief provisions are so advantageous
as to cause self-limiting of sales, except perhaps for a company very near the threshold. Even if
this were to happen, we do not see how moving the threshold to 15,000 would prevent the same
dynamic from happening at that sales level. Furthermore, our use of a three-year rolling average
of sales for determining SVM eligibility protects the SVMs from being penalized for having an
especially good year not reflective of a its long-term growth trend. See the 2017 and later light-
duty GHG final rule for a discussion of our basis for adopting the 5,000 vehicle threshold (77 FR
62793, October 15, 2012).
Comments from CARB and Ferrari supported the extension of SVM eligibility to
operationally independent small manufacturers. No commenters opposed it; however, Advanced
Biofuels USA recommended caution to avoid advantaging SVMs capable of leveraging parent
company resources to drastically increase U.S. market share within 2-3 years. Given the
precedent established in our GHG program, and the value of this extension for harmonization
with LEV III, we are adopting this change into Tier 3 using the same eligibility criteria as in our
GHG program, described in 40 CFR 86.1838-01 (d). We believe these criteria are sufficiently
strict and objective to address the concerns expressed by Advanced Biofuels USA.
10.9 Economic Effects
The following section summarizes the economic impact on small businesses of the Tier 3
exhaust and evaporative emission standards and the fuel requirements. As noted earlier, the
types of companies that will be affected by the Tier 3 exhaust and evaporative emissions include
vehicle manufacturers, ICIs, and alternative fuel converters. Similarly, the types of companies
that will be affected by the fuel requirements include gasoline refiners and importers, ethanol
producers, gasoline additive manufacturers, transmix processors, and terminal operators.
To gauge the impact of the Tier 3 standards on small businesses, EPA employed a cost-
to-sales ratio test to estimate the number of small businesses that would be impacted by less than
one percent, between one and three percent, and above three percent. The costs used in this
10-20
-------
analysis for the in-use gasoline requirements are based on the cost estimates developed in
Chapter 5 of this RIA. The costs used in this analysis for the Tier 3 exhaust and evaporative
emission standards are based on the cost estimates developed in Chapter 2 of this RIA,
supplemented with additional information for alternative fuel converters. A description of the
inputs used for the vehicle-related sectors and the methodology used to develop the estimated
impact on small businesses in each vehicle-related sector is presented in the docket for this
rulemaking.3
During discussions with SERs during the SBREFA Panel process, ethanol SERs
mentioned that ethanol producers are largely already meeting a 10 ppm sulfur standard and
terminal operator SERs indicated that effects on their industry would largely be based on
decisions made by refiners upstream. The final Tier 3 program retains the 80 ppm refinery gate
cap and 95 ppm downstream cap, which means that downstream parties such as transmix
processors, gasoline additive manufacturers, and terminal operators should not incur additional
costs as a result of this program.
While many gasoline refiners and importers are currently meeting, or are close to, a 10
ppm sulfur standard, there are some refiners that could incur increased costs as a result of the
Tier 3 program. Of the 12 refiners that EPA believes would be considered small refiners for the
Tier 3 program, it is projected that the majority of these refiners will experience costs less than
one percent of their sales, three refiners will experience costs between one and three percent of
their sales, and two refiners would incur costs of three percent or more of sales, as noted in Table
10-2 below.
For vehicle manufacturers, EPA identified four small businesses. One of the small
businesses manufactures adaptive vehicles (vehicles with adaptive equipment for persons with
disabilities) in both gasoline-powered and CNG-powered versions. One of small businesses
manufactures a hybrid gasoline-electric vehicle. Both of these manufacturers purchase engines
from large manufacturers and then use the engines in their vehicles. Both of these manufacturers
rely on third-party testing facilities to perform emissions testing. Both of these manufacturers
are projected to incur compliance costs of less than one percent. Finally, two additional small
businesses identified by EPA manufacture all-electric vehicles. For the two small businesses
manufacturing all-electric vehicles, the estimated costs for meeting the Tier 3 vehicle standards
are zero.
For alternative fuel converters, EPA identified nine small businesses. Of the small
business alternative fuel converters, four are projected to incur compliance costs above three
percent and three are projected to incur compliance costs between one and three percent. Two
small business alternative fuel converters will be impacted by less than one percent.
For ICIs, EPA identified eight small businesses. All eight of the small business ICIs are
projected to incur compliance costs between one and three percent.
Table 10-2 summarizes the impacts of the regulations on small businesses impacted by
the Tier 3 fuel requirements and exhaust and evaporative emission standards.
10-21
-------
Table 10-2 Summary of Impacts on Small Businesses
Industry Sector
Gasoline refiners and importers
Vehicle manufacturers
Alternative fuel converters
Independent commercial importers
Totals
0-1 Percent
7
4
2
0
13
1-3 Percent
3
0
3
8
14
>3 Percent
2
0
4
0
6
For a complete discussion of the economic impacts of the final Tier 3 rulemaking, see
Chapter 9, the Economic Impact Analysis chapter, of this Regulatory Impact Analysis.
10-22
-------
References
1 Final Report of the Small Business Advocacy Review Panel on EPA's Planned Proposed Rule Control of Air
Pollution from New Motor Vehicles: Tier 3 Emission and Fuel Standards, October 3, 2011.
2 Final Report of the Small Business Advocacy Review Panel on EPA's Planned Proposed Rule Control of Air
Pollution from New Motor Vehicles: Tier 3 Emission and Fuel Standards, Octobers, 2011.
3 "Small Business Impact Memo, Proposed Tier 3 Motor Vehicle Emission Standards and Related Provisions," EPA
memorandum from Phil Carlson to EPA Docket EPA-HQ-OAR-2011-0135, December 1, 2011.
10-23
------- |