Control of Air Pollution from
New Motor Vehicles: Heavy-Duty Engine
and Vehicle Standards
Regulatory Impact Analysis
SEPA
United States
Environmental Protection
Agency
-------
Control of Air Pollution from
New Motor Vehicles: Heavy-Duty Engine
and Vehicle Standards
This technical report does not necessarily represent final EPA decisions
or positions. It is intended to present technical analysis of issues using
data that are currently available. The purpose in the release of such
reports is to facilitate the exchange of technical information and to
inform the public of technical developments.
Regulatory Impact Analysis
Assessment and Standards Division
Office of Transportation and Air Quality
U.S. Environmental Protection Agency
NOTICE
4>EPA
United States
Environmental Protection
Agency
EPA-420-R-22-035
December 2022
-------
Table of Contents
Table of Contents i
List of Acronyms v
List of Tables xiv
List of Figures xxvi
Executive Summary 1
Chapter 1 Technology to Control Emissions from Heavy-Duty Engines 6
1.1 Compression-Ignition Engine Technologies 6
1.1.1 Current Catalyst Technologies 6
1.1.2 Catalyst Durability 12
1.1.3 Improving SCR NOx Reduction at Low Exhaust Temperatures 15
1.1.4 Crankcases 19
1.1.5 Opposed-Piston Diesel Engines 22
1.2 Spark-Ignition Engine Technologies 23
1.2.1 Technology Description for NMHC, CO, and NOx Control 23
1.2.2 Technology Description for PM Control 32
1.2.3 Technologies to Address Evaporative and Refueling Emissions 33
1.3 Fuels Considerations 37
1.3.1 Natural gas 37
1.3.2 Biodiesel 38
Chapter 2 Compliance Provisions 58
2.1 Compression-Ignition Engine Dynamometer Test Procedures 58
2.1.1 Current CI Test procedures 58
2.1.2 Final updates to CI Test procedures 60
2.2 Manufacturer-Run Off-Cycle Field Testing Program for Compression-Ignition Engines..67
2.2.1 Current Field Testing Program and Off-Cycle Standards 67
2.2.2 Information evaluated for final updates 70
2.2.3 Final Updates to CI Engine Off-Cycle Test Program and Off-Cycle Standards 87
2.3 Spark-Ignition Test Procedures and Standards 92
2.3.1 Current SI Test procedures 92
2.3.2 Summary of Updates Considered for SI Test Procedures and Standards 94
2.4 Useful Life 99
2.4.1 History of Regulatory Useful Life 99
2.4.2 Identifying Appropriate Useful Life Periods 101
Chapter 3 Feasibility Analysis for the Final Standards 108
3.1 Compression-Ignition Technology Feasibility 108
3.1.1 Diesel Technology Demonstration Programs 108
3.1.2 Baseline Technology Effectiveness 131
3.1.3 GHG Impacts 133
3.1.4 Estimated Direct Manufacturing Costs for Technology Packages Evaluated 134
3.2 Spark-Ignition Technology Feasibility 136
3.2.1 Baseline Technology Effectiveness 136
3.2.2 Projected Technology Effectiveness 146
3.2.3 Estimated Direct Manufacturing Costs for Technology Packages Evaluated 157
Chapter 4 Health and Environmental Impacts 168
l
-------
4.1 Health Effects Associated with Exposure to Pollutants 168
4.1.1 Ozone 168
4.1.2 Particulate Matter 170
4.1.3 Nitrogen Oxides 175
4.1.4 Carbon Monoxide 176
4.1.5 Diesel Exhaust 178
4.1.6 Air Toxics 180
4.1.7 Exposure and Health Effects Associated with Traffic 183
4.2 Environmental Effects Associated with Exposure to Pollutants 186
4.2.1 Visibility Degradation 186
4.2.2 Plant and Ecosystem Effects of Ozone 189
4.2.3 Deposition 190
4.2.4 Environmental Effects of Air Toxics 197
4.3 Environmental Justice 197
Chapter 5 Emissions Inventory 220
5.1 Introduction 220
5.2 Model and Data Updates 220
5.2.1 Methodology Overview 221
5.2.2 MOVES Emission Rates for Control Scenarios 222
5.3 National Emissions Inventory Results 258
5.3.1 Final Standards 259
5.3.2 Proposed Option 2 259
5.3.3 Impacts of Heavy-Duty Gasoline Refueling Controls 260
5.4 Emissions Inventories for Air Quality Modeling 260
5.4.1 Control Scenario Evaluated for the Air Quality Modeling Analysis 262
5.4.2 Estimated Differences in the Emission Reductions between the Final Control Scenario
and the Control Scenario Analyzed for Air Quality Modeling 265
5.5 Chapter 5 Appendix 266
5.5.1 Zero-Mile Emission Rates for the Control Scenarios 266
5.5.2 Details of the Emission Impacts of the Final Standards 267
5.5.3 Onroad Heavy-Duty NOx Emissions by Engine Operational Process for the Baseline,
Final and Proposed Option 2 Standards 271
5.5.4 Year-Over-Year Criteria Pollutant Emissions for Calendar Years Between 2027 and
2045 272
5.5.5 Sensitivity Analysis of Emissions Impacts of 2026 Service Class Pull Ahead Credits
Pathway 277
Chapter 6 Air Quality Impacts 284
6.1 Current Air Quality 284
6.1.1 Ozone 284
6.1.2 PM2 5 286
6.1.3 N02 287
6.1.4 CO 288
6.1.5 Air Toxics 288
6.1.6 Visibility 288
6.1.7 Deposition 288
6.2 Air Quality Impacts of the Final Rule 289
ii
-------
6.3 Air Quality Modeling Methodology for Proposal Analysis 2
6.3,1 Air Quality Model 2
63,2 Model Domain and Configuration 2
6.3.3 Model Inputs 2
6.3.4 CMAQ Evaluation 2
6.3.5 Model Simulation Scenarios 2
6.4 Air Quality Modeling Results of the Proposed Rule 2
6.4.1 Ozone Design Value Impacts of Proposed Rulemaking 2
6.4.2 Annual PM2.5 Design Value Impacts of Proposed Rulemaking 2
6.4.3 24-hour PM2.5 Design Value Impacts of Proposed Rulemaking 2
6.4.4 Nitrogen Dioxide Concentration Impacts of Proposed Rulemaking 3
6.4.5 Carbon Monoxide Concentration Impacts of Proposed Rulemaking 3
6.4.6 Air Toxics Impacts of Proposed Rulemaking 3
6.4.7 Visibility Impacts of Proposed Rulemaking 3
6.4.8 Deposition Impacts of Proposed Rulemaking 3
6.4.9 Demographic Analysis of Air Quality 3
Chapter 7 Program Costs 3:
7.1 Technology Package Costs 3
7.1.1 Direct Manufacturing Costs 3
7.1.2 Indirect Costs 3
7.1.3 Technology Costs per Vehicle 3
7.2 Operating Costs 3
7.2.1 Costs Associated with Increased Diesel Exhaust Fluid (DEF) Consumption in Diesel
Engines 3'
7.2.2 Costs Associated with Changes in Fuel Consumption on Gasoline Engines 3
7.2.3 Emission-Related Repair Cost Impacts Associated with the Final Program 3
7.3 Program Costs 3
7.3.1 Total Technology Costs 3
7.3.2 Total Operating Costs 3
7.3.3 Total Program Costs 3
Chapter 8 Estimated Benefits 3
8.1 Overview 3
8.2 Health Impact Assessment for PM2.5 and Ozone 3
8.2.1 Preparing Air Quality Modeling Data for Health Impacts Analysis 3
8.2.2 Selecting Air Pollution Health Endpoints to Quantify 3
8.2.3 Calculating Counts of Air Pollution Effects Using the Health Impact Function 3
8.2.4 Quantifying Ozone-Attributable Premature Mortality 3
8.2.5 Quantifying PM2.5-Attributable Premature Mortality 3
8.3 Economic Valuation Methodology for Health Benefits 3
8.4 Characterizing Uncertainty in the Estimated Benefits 3
8.5 Estimated Number and Economic Value of Health Benefits 3
8.6 Present Value of Total Benefits 3
8.7 Unquantified Benefits 3
Chapter 9 Comparison of Benefits and Costs 41
9.1 Methods 41
9.2 Results 41
111
-------
Chapter 10 Economic Impact Analysis 405
10.1 Impact on Sales, Fleet Turnover and Mode Shift 405
10.1.1 Sales 406
10.1.2 EPA's Research to Estimate Sales Effects 408
10.1.3 Fleet Turnover and Emissions Impacts 415
10.1.4 Potential for Mode Shift 417
10.1.5 Effects on Domestic and International Shares of Production 418
10.1.6 Summary of Sales, Turnover, and Mode Shift Impacts 419
10.2 Employment Impacts 419
10.2.1 Economic Framework for Employment Impact Assessment 419
10.2.2 Employment Impacts in the Motor Vehicle and Parts Manufacturing Sectors 421
10.2.3 Employment Impacts on Related Sectors 428
10.2.4 Summary of Employment Impacts 429
Chapter 11 Small Business Analysis 432
11.1 Definition and Description of Small Businesses 432
11.2 Overview of the Heavy-Duty Program and Type of Entities Covered 432
11.3 Impacts on Small Entities: Heavy-Duty Secondary Vehicle Manufacturers 433
11.4 Impacts on Small Entities: Heavy-Duty Alternative Fuel Engine Converters 437
11.5 Summary Table of Impacts on Small Businesses Subject to the Rule 438
iv
-------
List of Acronyms
Acronym
Definition
°C
Degrees Celsius
Microgram
um
Micrometers
20xx$
U.S. Dollars in calendar year 20xx
A/C
Air Conditioning
ABT
Averaging, Banking and Trading
AC
Alternating Current
ACEA
European Automobile Manufacturers Association
ACES
Advanced Collaborative Emission Study
AECD
Auxiliary Emissions Control Device
AEO
Annual Energy Outlook
AES
Automatic Engine Shutdown
AFDC
Alternative Fuels Data Center
AHS
American Housing Survey
A1
Aluminum
A12Ti05
Aluminum Titanate
AMOC
Atlantic Meridional Overturning Circulation
AMT
Automated Manual Transmission
ANL
Argonne National Laboratory
ANPR(M)
Advanced Notice of Proposed Rulemaking
APU
Auxiliary Power Unit
AO
Air Quality
AQCD
Air Quality Criteria Document
AR4
Fourth Assessment Report
ARB
California Air Resources Board
ARB HHDDT
California Air Resources Board Heavy Heavy-Duty Diesel Test
ASC
Ammonia Slip Catalyst
ASL
Aggressive Shift Logic
ASM
Annual Survey of Manufacturers
ASTM
ASTM International, formerly American Society for Testing and Materials
ASTM
ASTM International, formerly known as American Society for Testing and Materials
AT
Automatic Transmissions
ATA
American Trucking Association
ATIS
Automated Tire Inflation System
ATRI
Alliance for Transportation Research Institute
ATSDR
Agency for Toxic Substances and Disease Registry
ATUS
American Time Use Survey
Avg
Average
B100
1 methyl-ester biodiesel fuel
B20
0.2 biodiesel blended with 0.8 petroleum distilate diesel fuel
BenMAP
Benefits Mapping and Analysis Program
BETP
Bleed Emissions Test Procedure
BEV
Battery Electric Vehicle
bhp
Brake Horsepower
bhp-hr
Brake Horsepower Hour
BLS
Bureau of Labor Statistics
BMEP
Brake Mean Effective Pressure
BSFC
Brake Specific Fuel Consumption
BTS
Bureau of Transportation Statistics
BTS
Bureau of Labor Statistics
V
-------
BTU
British Thermal Unit
Ca
Calcium
CAA
Clean Air Act
CAAA
Clean Air Act Amendments
CaC03
Calcium Carbonate
CAD/CAE
Computer Aided Design And Engineering
CAE
Computer Aided Engineering
CAFE
Corporate Average Fuel Economy
CAN
Controller Area Network
CARB
California Air Resources Board
CaS04
Calcium Sulfate
CBI
Confidential Business Information
CCP
Coupled Cam Phasing
CCSP
Climate Change Science Program
CDA
Cylinder Deactivation
CDC
Centers for Disease Control
CDPF
Catalyzed Diesel Particulate Filter
CFD
Computational Fluid Dynamics
CFR
Code of Federal Regulations
ch4
Methane
CI
Compression-ignition
CILCC
Combined International Local and Commuter Cycle
CIPM
International Committee for Weights and Measures (Bureau International des Poids et
Mesures)
CITT
Chemical Industry Institute of Toxicology
CMAQ
Community Multiscale Air Quality
CNG
Compressed Natural Gas
CO
Carbon Monoxide
C02
Carbon Dioxide
C02eq
C02 Equivalent
COFC
Container-on-Flatcar
COI
Cost of Illness
COPD
Chronic Obstructive Pulmonary Disease
CoV
Coefficient of Variation
CPS
Cam Profile Switching
CPSI
Cells per Square Inch
CRC
Coordinating Research Council
CRGNSA
Columbia River Gorge National Scenic Area
CRR
Rolling Resistance Coefficient
CS
Climate Sensitivity
CSI
Cambridge Systematics Inc.
CSS
Coastal Sage Scrub
CSV
Comma-separated Values
Cu
Copper
CuO
Copper(II) oxide or cupric oxide
CVD
Cardiovascular Disease
CVT
Continuously-Variable Transmission
CW
Curb Weight
D/UAF
Downward and Upward Adjustment Factor
DAAAC
Diesel Aftertreatment Accelerated Aging Cycle
DARAP
Diesel Aftertreatment Rapid Aging Protocol
DCP
Dual Cam Phasing
DCT
Dual Clutch Transmission
vi
-------
DE
Diesel Exhaust
DEAC
Cylinder Deactivation
DEER
Diesel Engine-Efficiency and Emissions Research
DEF
Diesel Exhaust Fluid
deSOx
Removal of sulfur oxide compounds
DF
Deterioration Factor
DHHS
U.S. Department of Health and Human Services
Diesel HAD
Diesel Health Assessment Document
DMC
Direct Manufacturing Costs
DO
Dissolved Oxygen
DOC
Diesel Oxidation Catalyst
DOD
Department of Defense
DOE
Department of Energy
DOHC
Dual Overhead Camshaft Engines
DOT
Department of Transportation
DPF
Diesel Particulate Filter
DPM
Diesel Particulate Matter
DR
Discount Rate
RIA
Draft Regulatory Impact Analysis
DWL
Discrete Variable Valve Lift
EAS
Exhaust Aftertreatment System
EC
European Commission
EC
Elemental Carbon
EC
Economic Census
ECM
Electronic Control Module
ED
Emergency Department
EERA
Energy and Environmental Research Associates
EFR
Engine Friction Reduction
EGR
Exhaust Gas Recirculation
EHPS
Electrohydraulic Power Steering
EIA
Energy Information Administration (part of the U.S. Department of Energy)
EISA
Energy Independence and Security Act
EIVC
Early Intake Valve Closing
EMS-HAP
Emissions Modeling System for Hazardous Air Pollution
EO
Executive Order
EPA
Environmental Protection Agency
EPMA
electron probe microanalysis
EPS
Electric Power Steering
ERG
Eastern Research Group
ERM
Employment Requirements Matrix
ESC
Electronic Stability Control
ETC
Electronic Throttle Control
ETW
Estimated Test Weight
EV
Electric Vehicle
F
Frequency
FCEV
Fuel Cell Electric Vehicle
Fe
Iron
FEL
Family Emission Limit
FET
Federal Excise Tax
FEV1
Functional Expiratory Volume
FHWA
Federal Highway Administration
FIA
Forest Inventory and Analysis
FMCSA
Federal Motor Carrier Safety Administration
vii
-------
FOH
Fuel Operated Heater
FR
Federal Register
FRM
Final Rulemaking
FTP
Federal Test Procedure
FVC
Forced Vital Capacity
s
Gram
g/s
Gram-per-second
g/ton-mile
Grams emitted to move one ton (2000 pounds) of freight over one mile
gal
Gallon
gal/1000 ton-
mile
Gallons of fuel used to move one ton of payload (2,000 pounds) over 1000 miles
GCAM
Global Change Assessment Model
GCW
Gross Combined Weight
GCWR
Rated Gross-combined Weight (vehicle + trailer)
GDI
Gasoline Direct Injection
GDI
Gasoline Direct Injection
GDP
Gross Domestic Product
GEM
Greenhouse gas Emissions Model
GEOS
Goddard Earth Observing System
GHG
Greenhouse Gas
GIFT
Geospatial Intermodal Freight Transportation
GPF
Gasoline Particulate Filter
GREET
Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation
GSF1
Generic Speed Form one
GUI
Graphical User Interface
GVWR
Gross Vehicle Weight Rating
GVWR
Rated Gross Vehicle Weight
GWP
Global Warming Potential
H20
Water
HABs
Harmful Algal Blooms
HAD
Diesel Health Assessment Document
HC
Hydrocarbon
HD
Heavy-Duty
HDDEFTP
Heavy-Duty Diesel Engine Federal Test Procedure
HDE
Heavy-Duty Engine
HDOE FTP
Heavy-Duty Otto-Cycle Engine Federal Test Procedure
HDT
Heavy-Duty Truck
HDUDDS
Heavy Duty Urban Dynamometer Driving Cycle
HDV
Heavy-Duty Vehicle
HEG
High Efficiency Gearbox
HEI
Health Effects Institute
HES
Health Effects Subcommittee
HEV
Hybrid Electric Vehicle
HFC
Hydrofluorocarbon
HFET
Highway Fuel Economy Dynamometer Procedure
HHD
Heavy Heavy-Duty
HHDDT
Highway Heavy-Duty Diesel Transient
HNCO
Iso-cyanic Acid
hp
Horsepower
hrs
Hours
HRV
Heart Rate Variability
HSC
High Speed Cruise Duty Cycle
HTUF
Hybrid Truck User Forum
viii
-------
HWFE
Highway Fuel Economy Drive Cycle
hz
Hertz
IARC
International Agency for Research on Cancer
IATC
Improved Automatic Transmission Control
IC
Indirect Costs
ICCT
International Council on Clean Transport
ICD
International Classification of Diseases
ICF
ICF International
ICM
Indirect Cost Multiplier
ICP
Intake Cam Phasing
ICP-MS
nductively coupled plasma mass spectrometry
IMAC
Improved Mobile Air Conditioning
IMPROVE
Interagency Monitoring of Protected Visual Environments
IPCC
Intergovernmental Panel on Climate Change
IRAF
infrequent regeneration factor
IRFA
Initial Regulatory Flexibility Analysis
IRIS
Integrated Risk Information System
ISA
Integrated Science Assessment
JAMA
Journal of the American Medical Association
k
Thousand
K
Potassium
kg
Kilogram
KI
kinetic intensity
km
Kilometer
km/h
Kilometers per Hour
kW
Kilowatt
L
Liter
LA92
Inventory development dynamometer driving schedule
lb
Pound
LD
Light-Duty
LDT
Light-Duty Truck
LHD
Light Heavy-Duty
LIVC
Late Intake Valve Closing
LLC
Low Load Cycle
LLNL
Lawrence Livermore National Laboratory's
LPG
Liquified Petroleum Gas
LRR
Lower Rolling Resistance
LSC
Low Speed Cruise Duty Cycle
LT
Light Trucks
LTCCS
Large Truck Crash Causation Study
LUB
Low Friction Lubes
LUC
Land Use Change
m2
Square Meters
m3
Cubic Meters
MAGICC
Model for the Assessment of Greenhouse-gas Induced Climate Change
MCF
Mixed Conifer Forest
MD
Medium-Duty
MDPV
Medium-Duty Passenger Vehicle
MECA
Manufacturers of Emissions Control Association
mg
Milligram
Mg
Magnesium
Mg(OH)2
Magnesium Hydroxide
mg/hp-hr
Milligrams per horsepower-hour
ix
-------
MHD
Medium Heavy-Duty
MHEV
Mild Hybrid
mi
mile
min
Minute
MM
Million
MMBD
Million Barrels per Day
MMT
Million Metric Tons
Mn
Manganese
MOVES
MOtor Vehicle Emissions Simulator
MP-AES
Microwave Plasma Atomic Emission Spectroscopy
mpg
Miles per Gallon
mph
Miles per Hour
MRL
Minimal Risk Level
MSAT
Mobile Source Air Toxic
MT
Manual Transmission
MTS
Maximum Test Speed
MW
Megawatt
MY
Model Year
N2
Molecular Nitrogen
N20
Nitrous Oxide
Na
Sodium
NA
Not Applicable
NAAQS
National Ambient Air Quality Standards
NACFE
North American Council for Clean Freight Efficiency
NAFA
National Association of Fleet Administrators
NAICS
North American Industry Classification System
NAS
National Academy of Sciences
NASTC
National Association of Small Trucking Companies
NATA
National Air Toxic Assessment
NCAR
National Center for Atmospheric Research
NCI
National Cancer Institute
NCLAN
National Crop Loss Assessment Network
NEC
Net Energy Change Tolerance
NEI
National Emissions Inventory
NEMS
National Energy Modeling System
NEPA
National Environmental Policy Act
NESCAUM
Northeastern States for Coordinated Air Use Management
NESCCAF
Northeast States Center for a Clean Air Future
NESHAP
National Emissions Standards for Hazardous Air Pollutants
NH3
Ammonia
NHS
National Highway System
NHTSA
National Highway Traffic Safety Administration
NiMH
Nickel Metal-Hydride
NIOSH
National Institute of Occupational Safety and Health
NIST
National Institute for Standards and Technology
Nm
Newton-meters
NMHC
Nonmethane Hydrocarbons
NMMAPS
National Morbidity, Mortality, and Air Pollution Study
NO
Nitric Oxide
no2
Nitrogen Dioxide
NOAA
National Oceanic and Atmospheric Administration
NOx
Oxides of Nitrogen
NPRM
Notice of Proposed Rulemaking
X
-------
NPV
Net Present Value
NRC
National Research Council
NRC-CAN
National Research Council of Canada
NREL
National Renewable Energy Laboratory
NTE
Not-to-exceed
NTEA
National Truck and Equipment Association
NTP
National Toxicology Program
NVH
Noise Vibration and Harshness
O&M
Operating and maintenance
03
Ozone
OAQPS
Office of Air Quality Planning and Standards
OBD
Onboard diagnostics
OC
Organic Carbon
OE
Original Equipment
OEHHA
Office of Environmental Health Hazard Assessment
OEM
Original Equipment Manufacturer
OHV
Overhead Valve
OMB
Office of Management and Budget
OOIDA
Owner-Operator Independent Drivers Association
OPEC
Organization of Petroleum Exporting Countries
ORD
EPA's Office of Research and Development
ORNL
Oak Ridge National Laboratory
ORVR
On-Board Refueling Vapor Recovery
ORVR
Onboard refueling vapor recovery
OTAQ
Office of Transportation and Air Quality
P
Phosphorus
Pa
Pascal
PAH
Polycyclic Aromatic Hydrocarbons
PCV
Positive Crankcase Ventilation
PEF
Peak Expiratory Flow
PEMFC
Proton-Exchange Membrane Fuel Cell
PEMS
Portable Emissions Monitoring System
PFI
Port Fuel Injection
PFI
Port Fuel Injection
PGM
Platinum Group Metal
PHEV
Plug-in Hybrid Electric Vehicles
PLT
Production-line testing
PM
Particulate Matter
PM10
Coarse Particulate Matter (diameter of 10 um or less)
pm25
Fine Particulate Matter (diameter of 2.5 um or less)
POM
Polycyclic Organic Matter
Ppb
Parts per Billion
Ppm
Parts per Million
Psi
Pounds per Square Inch
PTO
Power Take Off
R&D
Research and Development
RBM
Resisting Bending Moment
REL
Reference Exposure Level
RESS
Rechargeable Energy Storage System
RFA
Regulatory Flexibility Act
RfC
Reference Concentration
RFS2
Renewable Fuel Standard 2
RIA
Regulatory Impact Analysis
xi
-------
RMC-SET
Ramped Modal Cycle Supplementary Emissions Test
RPE
Retail Price Equivalent
RPM
Revolutions per Minute
RSWT
Reduced-Scale Wind Tunnel
S
Second
S
Sulfur
SAB
Science Advisory Board
SAB-HES
Science Advisory Board - Health Effects Subcommittee
SAE
SAE International, formerly Society of Automotive Engineers
SAR
Second Assessment Report
SAV
Submerged Aquatic Vegetation
SBA
Small Business Administration
SBREFA
Small Business Regulatory Enforcement Fairness Act
SCR
Selective Catalyst Reduction
SEA
Selective enforcement audit
SER
Small Entity Representation
SET
Supplemental Emission Test
SGDI
Stoichiometric Gasoline Direct Injection
SHED
Sealed Housing Evaporative Determination
SHEV
Strong Hybrid Vehicles
SI
Spark-Ignition
SiC
Silicon Carbide
SIDI
Spark Ignition Direct Injection
S02
Sulfur Dioxide
SOA
Secondary Organic Aerosol
SOC
State of Charge
SOFC
Solid Oxide Fuel Cells
SOHC
Single Overhead Cam
SOx
Sulfur Oxides
SOx
Oxides of Sulfur
SPR
Strategic Petroleum Reserve
SSZ-13
A Chabazite-type Aluminosilicate ABC-6 Zeolite
STB
Surface Transportation Board
Std.
Standard
STP
Scaled Tractive Power
SUV
Sport Utility Vehicle
svoc
Semi-Volatile Organic Compound
SwRI
Southwest Research Institute
TAR
Technical Assessment Report
TC
Total Costs
TCO
Total Cost of Ownership
TCp
Total Cost package
TDS
Turbocharging And Downsizing
THC
Total Hydrocarbon
TIAX
TIAX LLC
TMC
Technology & Maintenance Council
TOFC
Trailer-on-Flatcar
Ton-mile
One ton (2000 pounds) of payload over one mile
TRBDS
Turbocharging and Downsizing
TRU
Trailer Refrigeration Unit
TSD
Technical Support Document
TSS
Thermal Storage
TW
Test Weight
xii
-------
TWC
Three-Way Catalyst
U.S.
United States
U/DAF
Upward and Downward Adjustment Factor
UBE
Useable battery energy
UCT
Urban Creep and Transient Duty Cycle
UFP
Ultra Fine Particles
ULSD
Ultra-low sulfur diesel
URE
Unit Risk Estimate
USD A
United States Department of Agriculture
USGCRP
United States Global Change Research Program
uv
Ultraviolet
UV-b
Ultraviolet-b
VGT
Variable-geometry Turbine
VIN
Vehicle Identification Number
VIUS
Vehicle Inventory Use Survey
VMT
Vehicle Miles Traveled
voc
Volatile Organic Compound
VSL
Vehicle Speed Limiter
VTEC-E
Variable Valve Timing & Lift Electronic Control-Economy
VTRIS
Vehicle Travel Information System
WL
Variable Valve Lift
WT
Variable Valve Timing
WACAP
Western Airborne Contaminants Assessment Project
WHR
Waste Heat Recovery
WHTC
World Harmonized Transient Cycle
WHVC
World Harmonized Vehicle Cycle
WRF
Weather Research Forecasting
WTP
Willingness-to-Pay
WTVC
World Wide Transient Vehicle Cycle
WVU
West Virginia University
Zn
Zinc
ZSM-5
Zeolite Socony Mobil-5, an Aluminosilicate Pentasil Zeolite within the family of zeolites
Xlll
-------
List of Tables
Table ES-1: Projected Heavy-Duty Emission Reductions in 2045 from the Proposed Options 1
and 2 Standards 2
Table ES-2: Total Program Costs: Undiscounted Annual Costs in 2045 and Annualized Costs
through 2045 at 3% and 7% Discount Rates (Billions of 2017 dollars) 3
Table ES-3: 2045 Annual Value, Present Value and Equivalent Annualized Value of Benefits of
the Proposed Options 1 and 2 (billions, 2017$)a'b 3
Table ES-4: 2045 Annual Value, Present Value and Equivalent Annualized Value of Costs,
Benefits and Net Benefits of the Proposed Options 1 and 2 (billions, 2017$)a'b 4
Table 1-1: Engine-displacement specific catalyst substrate volume for MY2019 MHDDE and
HHDDE 11
Table 1-2: Average Crankcase Emission Rates (gram/hour) 20
Table 1 -3: ORVR results with modeled values for test procedure limitations 33
Table 1-4: NREL Fuel Samples off Specification for ASTM D6751 (or equivalent B20) limit for
Na + K and Ca + Mg 44
Table 1-5: Test procedure LODs for NREL studies 44
Table 1-6: NREL 2016 Metals results for UOP-389 and ICP-MS70 46
Table 1-7: EPA 2017 Metals results for ICP-AES analysis of Na, K, Ca, and Mg performed by
CDF A 49
Table 1-8: EPA 2017 Metals results for ICP-AES analysis of Mo, P, B, Ba, Cu, Mn, Si, Ti, V,
and Zn performed by CDF A 50
Table 1-9: BEV & FCEV Sales Percentages Error! Bookmark not defined.
Table 1-10: BEV Truck Offerings or Planned Offerings in the US (as April 2022) Error!
Bookmark not defined.
Table 1-11: BEV Bus Offerings or Planned Offerings in the US (as April 2022) Error!
Bookmark not defined.
Table 2-1: SET Mode Weighting Factors for the 2010 NOx and Phase 1 GHG Standards 59
Table 2-2: New SET Mode Weighting Factors in Phase 2 61
Table 2-3: Breakdown of vehicles from combined NREL Fleet DNA and CARB datasets 63
Table 2-4: Representative Summary of GEM Generated Profiles for the Engine Duty-cycle 66
Table 2-5: NOx Emission Levels for a 2010 Compliant Engine on Three Candidate LLCs 67
Table 2-6: Engine Standards and In-use Measurement Allowance 68
Table 2-7: Percentage of Tests Meeting or Exceeding 0.90 Pass Ratio Threshold 70
Table 2-8: Average Pass Ratios for Data Submitted Since 2005 70
Table 2-9: Number of Diesel Vehicles with MY 2010+ Engines by NOx FEL Group from the
heavy-duty in-use testing Program 72
Table 3-1: Major engine specifications for the MY2017 Cummins X15 engine used for the
CARB Low NOx Stage 3 Research Program 108
Table 3-2: Summary of catalyst specifications for developmental EAS with light-off SCR 111
Table 3-3: Baseline (degreened) emissions results for the OE Cummins EAS 112
xiv
-------
Table 3-4: 0-hour (degreened) emissions results for the developmental EAS system with light-off
SCR 112
Table 3-5: Emissions results for the developmental EAS system with light-off SCR after 334
hours of accelerated thermal and chemical aging using the DAAAC (equivalent to
approximately 145,000 miles of operation). The SET (2021) results represent updated 40 CFR
§1036.505 SET procedures 112
Table 3-6: Emissions results for the developmental EAS system with light-off SCR after 667
hours of accelerated thermal and chemical aging using the DAAAC (equivalent to
approximately 290,000 miles of operation). The SET (2021) results represent updated 40 CFR
§1036.505 SET procedures 113
Table 3-7: Emissions results for the developmental EAS system with light-off SCR after 1000
hours of accelerated thermal and chemical aging using the DAAAC (equivalent to
approximately 435,000 miles of operation). The SET (2021) results represent updated 40 CFR
§1036.505 SET procedures 113
Table 3-8: Emissions results for the developmental EPA Stage 3 EAS system with light-off SCR
and separate DOC and DPF after 1000 hours of accelerated thermal and chemical aging using
the DAAAC (equivalent to approximately 435,000 miles of operation). The SET (2021)
results represent updated 40 CFR §1036.505 SET procedures 116
Table 3-9: Emissions results for the developmental EPA Stage 3 EAS system with light-off SCR
and separate DOC and DPF after 1379 hours of accelerated thermal and chemical aging using
the DAAAC (equivalent to approximately 600,000 miles of operation). The SET (2021)
results represent updated 40 CFR §1036.505 SET procedures 117
Table 3-10: Off-cycle NOx emissions results for the developmental EAS system with light-off
SCR and separate DOC and DPF after 1000 hours of accelerated thermal and chemical aging
using the DAAAC (equivalent to approximately 435,000 miles of operation) 122
Table 3-11: Off-cycle NMHC emissions results for the developmental EAS system with light-off
SCR and separate DOC and DPF after 1000 hours of accelerated thermal and chemical aging
using the DAAAC (equivalent to approximately 435,000 miles of operation) 122
Table 3-12: Off-cycle CO emissions results for the developmental EAS system with light-off
SCR and separate DOC and DPF after 1000 hours of accelerated thermal and chemical aging
using the DAAAC (equivalent to approximately 435,000 miles of operation) 122
Table 3-13: Summary of catalyst specifications for developmental EAS with close-coupled light-
off SCR 127
Table 3-14: Summary of catalyst specifications for developmental EAS with light-off SCR
mounted under-cab Error! Bookmark not defined.
Table 3-15: Major engine specifications for the MY2018 Cummins XI5 engine used for EAS
and CDA development by EPA Error! Bookmark not defined.
Table 3-16: Summary of certification data for FTP cycle 132
Table 3-17: Summary of certification data for SET cycle 132
Table 3-18: 2018 DetroitDD15 engine emissions in g/hp-hr 133
Table 3-19: 2018 Cummins B6.7 engine emissions in g/hp-hr 133
Table 3-20: 2018 Navistar A26 engine emissions in g/hp-hr 133
xv
-------
Table 3-21: MY2019 average engine displacement and diesel EAS specifications for Light HDE,
Medium HDE, and Heavy HDE applications based on projected-sales-weighted-averages
from EPA heavy-duty diesel engine certification data Error! Bookmark not defined.
Table 3-22: MY2019 DOC System Costs for Light HDE, Medium HDE, & Heavy HDE
Applications Error! Bookmark not defined.
Table 3-23: MY2019 CDPF System Costs for Engine-dynamometer Certified Light HDE,
Medium HDE, Heavy HDE, and Urban Bus Applications Error! Bookmark not defined.
Table 3-24: MY2019 SCR System Costs for Engine-dynamometer Certified Light HDE,
Medium HDE, Heavy HDE, and Urban Bus Applications Error! Bookmark not defined.
Table 3-25: Summary of MY2019 EAS Costs for Engine-dynamometer Certified Light HDE,
Medium HDE, Heavy HDE, and Urban Bus Applications Error! Bookmark not defined.
Table 3-26: Proposed Option 1 and 2 SCR System Costs for Engine-dynamometer Certified
Light HDE, Medium HDE, Heavy HDE, and Urban Bus Applications. Error! Bookmark not
defined.
Table 3-27: Summary of proposed Option 1 and 2 EAS Costs* for Engine-dynamometer
Certified Light HDE, Medium HDE, Heavy HDE, and Urban Bus Applications Error!
Bookmark not defined.
Table 3-28: Summary of CDA Costs from Teardown Study 135
Table 3-29: Summary of Costs from the FEV Exhaust Aftertreatment System Study 136
Table 3-30: Heavy-Duty Gasoline Vehicle Emissions Investigation Vehicle Specifications 137
Table 3-31: Average distance from exhaust manifold to catalyst 138
Table 3-32: Description of Real-world PEMS Testing Routes 139
Table 3-33: Test weights and dynamometer coefficients used for NVFEL HD gasoline testing.
139
Table 3-34: Family Emission Limits Reported for the Six Certified HD SI Engines in MY 2019;
NOx and NMHC values are converted from g/hp-hr to mg/hp-hr to match the units of our
proposed standards 147
Table 3-35: Average emission performance for Certified HD SI Engines in MY 2019 147
Table 3-36: Comparison of Simulated 6.8L V10 SET Composite Emissions to Proposed
Standards 150
Table 3-37: Proposed Option 1 Spark-Ignition Exhaust Emission Standards for SET Duty Cycle
Error! Bookmark not defined.
Table 3-38: SET Operation Mode Power Comparison Error! Bookmark not defined.
Table 3-39: Major engine specifications of the MY2019 HD SI gasoline engine used for the EPA
demonstration program Error! Bookmark not defined.
Table 3-40: Spark-Ignition Demonstration Program Preliminary FTP Results .Error! Bookmark
not defined.
Table 3-41: Spark-Ignition Demonstration Program Preliminary SET Results .Error! Bookmark
not defined.
Table 3-42: 2019 MY Sales-Weighted Baseline SI Engine Technology Costs (2019$) 158
xvi
-------
Table 3-43: 2019 MY HHD and Urban Bus Gaseous-fueled Technology Baseline Costs (2019$)
158
Table 3-44: Projected Gaseous Fueled Engine Technology Cost based on Proposed Standards
(2019$) 159
Table 3-45: Projected Liquid Fueled SI Engine Technology Cost to meet Proposed Standards
($2019$) 160
Table 3-46: Summary of HD SI Engine Technology Cost Comparison 160
Table 3-47: Assumptions for gasoline-fueled heavy-duty spark-ignition vehicles for conventional
carbon requirements to meet the proposed ORVR 163
Table 3-48: Estimated Costs for ORVR Over Tier 3 as Baseline 164
Table 5-1: Updates to MOVES CTINPRM from MOVES2014b Error! Bookmark not defined.
Table 5-2: MOVES Heavy-duty Regulatory Classes and Relevant MOVES Fuel Types Error!
Bookmark not defined.
Table 5-3: MOVES Running Operating Mode Definitions Error! Bookmark not defined.
Table 5-4: Heavy-duty Compression Ignition Duty-Cycle Cycle NOx Standards for the Proposed
Options and Alternative Scenarios Error! Bookmark not defined.
Table 5-5: Proposed Option 1 Weighted Average Heavy Heavy-duty Compression Ignition
Duty-Cycle Test NOx Standards Error! Bookmark not defined.
Table 5-6: Rjuty Ratios Calculated for Each Scenario Error! Bookmark not defined.
Table 5-7: Calculation of Rin use by MOVES Operation Mode Error! Bookmark not defined.
Table 5-8. Calculated Off-Cycle NOx Standards used for the Control Scenarios Error!
Bookmark not defined.
Table 5-9: Calculation of Voluntary Idle NOx/C02 Standard (g/kg) Error! Bookmark not
defined.
Table 5-10: Calculation of the Off-cycle NOx Standard Compliant Emission Rate for HHD
Diesel Vehicles for the Proposed Option 1 Error! Bookmark not defined.
Table 5-11: Useful Life and Warranty Periods for Heavy-duty DieselA Engines and
Aftertreatment Systems in the Control Scenarios Error! Bookmark not defined.
Table 5-12: Estimated Vehicle Age at the End of the Warranty Period and the Useful Life for
Each Heavy-duty Diesel Regulatory Class for the Proposed Option 1 Model Year 2027
Scenario Error! Bookmark not defined.
Table 5-13: Estimated Vehicle Age at the End of the Warranty Period and the Useful Life for
Each Heavy-duty Diesel Regulatory Class in the Baseline and Control Scenarios Error!
Bookmark not defined.
Table 5-14: NOx Tampering & Mal-maintenance (T&M) Emission Effects Error! Bookmark
not defined.
Table 5-15: Calculation of sa Error! Bookmark not defined.
Table 5-16: MOVES ageGroupID Which Are Used to Define Running and Start Emission Rates
Error! Bookmark not defined.
Table 5-17: Calculation of NOx 12-hour Cold Starts from the CARB Stage 1 HHD Engine from
the Cold and Hot FTP Cycle Error! Bookmark not defined.
xvii
-------
Table 5-18: HHD Cold Start Emissions for Proposed and Alternative Scenarios Error!
Bookmark not defined.
Table 5-19: Calculation of HHD and MHD Extended Idle NOx g/hr Emission Rates Error!
Bookmark not defined.
Table 5-20: Running Emission Rate Reductions From Heavy-duty Gasoline Vehicles Due to
Proposed and Alternative Standards, Rgasoiine, Across All Heavy-duty Gasoline Regulatory
Classes and Operating Modes Error! Bookmark not defined.
Table 5-21: Phase-In of Onboard Refueling Vapor Recovery (ORVR) for Heavy-duty Trucks
Error! Bookmark not defined.
Table 5-22: National Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,
2040, and 2045 — Proposed Option 1 Program Emissions Relative to Heavy-Duty Vehicle
Emissions Baseline Error! Bookmark not defined.
Table 5-23: National Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,
2040, and 2045 — Proposed Option 2 Program Emissions Relative to Heavy-Duty Vehicle
Emissions Baseline Error! Bookmark not defined.
Table 5-24: National Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,
2040, and 2045 — Alternative Program Emissions Relative to Heavy-Duty Vehicle Emissions
Baseline Error! Bookmark not defined.
Table 5-25: Emission Reductions Due to Adoption of ORVR for Heavy-duty Vehicles Relative
to Heavy-Duty Vehicle Emissions Baseline Error! Bookmark not defined.
Table 5-26. Summary of Differences between Emissions in Proposed Option 1 and the Control
Scenario Analyzed for Air Quality Modeling Error! Bookmark not defined.
Table 5-27: Duty-Cycle NOx Standards for Proposed Option 1 and the Control Scenario
Analyzed for Air Quality ModelingA Error! Bookmark not defined.
Table 5-28. Rjuty Ratios Calculated for the Control Scenario Analyzed for Air Quality Modeling
Error! Bookmark not defined.
Table 5-29: Warranty Mileages and Years in Option 1 and the Control Scenario Analyzed for Air
Quality Modeling Error! Bookmark not defined.
Table 5-30:Useful Life Mileages and Years in Option 1 and the Control Scenario Analyzed for
Air Quality Modeling Error! Bookmark not defined.
Table 5-31: Onroad Vehicle Emission Reductions from the Air Quality Modeling Control
Scenario Using SMOKE-MOVES Inventories and National Inventories Error! Bookmark
not defined.
Table 5-32 Comparison of the Onroad Vehicle Emission Reductions from the Air Quality
Modeling Control Scenario vs. Option 1 Error! Bookmark not defined.
Table 5-33: Sales Volumes of Model Year 2027 LHD2b3 Diesel-Fueled Vehicles Estimated by
MOVES CTINPRM Error! Bookmark not defined.
Table 5-34: National Heavy-duty Vehicle NOx Emissions (Annual US Tons) For Calendar Years
Between 2027 and 2045 Error! Bookmark not defined.
Table 5-35: National Heavy-Duty Vehicle VOC Emissions (Annual US Tons) For Calendar
Years Between 2027 and 2045 Error! Bookmark not defined.
xviii
-------
Table 5-36: National Heavy-duty Vehicle PM2.5 (Exhaust Only) Emissions (Annual US Tons)
For Calendar Years Between 2027 and 2045 Error! Bookmark not defined.
Table 5-37: National Heavy-Duty Vehicle CO Emissions (Annual US Tons) For Calendar Years
Between 2027 and 2045 Error! Bookmark not defined.
Table 5-38: Fractions of the BEVs and FCEVs Based on AEO2018 Compared with the Fractions
of the BEVs Based on NREL Study in Each of the Heavy-Duty Vehicle Categories by Model
Year Error! Bookmark not defined.
Table 5-39: Percent Reduction from the Baseline for Proposed Option 1 and Sensitivity Cases
Based on AEO2018 and NREL in Calendar Years 2030, 2040, and 2045 ....Error! Bookmark
not defined.
Table 5-40: Percent Reduction from the Baseline for Proposed Option 2 and Sensitivity Cases
Based on AEO2018 and NREL in Calendar Years 2030, 2040, and 2045 ....Error! Bookmark
not defined.
Table 5-41: Percent Reduction from the Baseline for the Alternative and Sensitivity Cases Based
on AEO2018 and NREL in Calendar Years 2030, 2040, and 2045 Error! Bookmark not
defined.
Table 5-42: Duty-Cycle NOx Standards for the CARB Omnibus and EPA Proposed Option 1A
Error! Bookmark not defined.
Table 5-43: Weighted Average Heavy Heavy-duty Compression Ignition Duty-Cycle Test NOx
Standards used to estimate NOx emission rates for the Omnibus Nationwide Scenarios Error!
Bookmark not defined.
Table 5-44: Calculated Average Off-Cycle Standards for the Omnibus from the Average Idling
and Engine Cycles Standards and Off-Cycle Scaling Factors..Error! Bookmark not defined.
Table 5-45: Warranty Mileages and Years in Omnibus and EPA Proposed Option 1 Error!
Bookmark not defined.
Table 5-46: Useful Life Mileages and Years in Omnibus Program and EPA Proposed Option 1
Error! Bookmark not defined.
Table 5-47: Extended Idle NOx emission rates for the Omnibus Program . Error! Bookmark not
defined.
Table 5-48: Running Emission Rate Reductions From Heavy-duty Gasoline Vehicles For the
EPA Proposed Option 1 and Omnibus Nationwide Scenarios. Error! Bookmark not defined.
Table 5-49 National Heavy-duty Vehicle NOx Emission Reductions Relative to the Baseline
Case For Omnibus Nationwide Scenarios (Reductions Relative to EPA Proposed Option 1
Shown for Comparison) Error! Bookmark not defined.
Table 6-1: Average Change in Projected 8-hour Ozone Design Values in 2045 due to Proposed
Rule 295
Table 6-2: Change in 8-hour Ozone Design Values for Counties Projected to be Above the Level
of the 2015 8-hour Ozone NAAQS in 2045 296
Table 6-3: Average Change in Projected Annual PM2.5 Design Values in 2045 due to Proposed
Rule """.'"".298
Table 6-4: Change in Annual PM2.5 Design Values for Counties Projected to be Above the Level
of the Annual PM2.5 NAAQS in 2045 299
xix
-------
Table 6-5: Average Change in Projected 24-hour PM2.5 Design Values in 2045 due to Proposed
Rule """..".".".301
Table 6-6: Change in 24-hour PM2.5 Design Values for Counties Projected to be Above the 24-
hour PM2.5 NAAQS in 2045 302
Table 6-7: Demographic Analysis of Projected 2045 Ozone Reductions (ppb) from the Proposed
Rule, by Race/Ethnicity 315
Table 6-8: Demographic Analysis of Projected 2045 PM2.5 Reductions (|ig/m3) from the
Proposed Rule, by Race/Ethnicity 316
Table 6-9 Demographic Analysis of Projected 2045 Ozone Reductions (ppb) from the Proposed
Rule, by Poverty Status 317
Table 6-10 Demographic Analysis of Projected 2045 PM2.5 Reductions (|ig/m3) from the HD
2027 Proposed Rule, by Poverty Status 318
Table 7-1: ICCT Cost Estimates of 12L Diesel Engine-Related Emission Control Costs
Associated with Past US Emission Standards (2015 dollars) 328
Table 7-2: Diesel Engine-Related Emission Control System Costs in the "No Action" Baseline*
328
Table 7-3: Gasoline Engine-Related Emission Control System Costs in the "No Action"
Baseline* 329
Table 7-4: CNG Engine-Related Emission Control System Costs in the "No Action" Baseline*
329
Table 7-5 Diesel Technology and Package Direct Manufacturing Costs per Engine by Regulatory
Class, 2017 dollars* 330
Table 7-6 Gasoline Technology and Package Direct Manufacturing Costs per Engine by
Regulatory Class, 2017 dollars* 331
Table 7-7: CNG Technology and Package Direct Manufacturing Costs per Engine by Regulatory
Class, 2017 dollars* Error! Bookmark not defined.
Table 7-8: GDP Price Deflators* Used to Adjust Costs to 2017 Dollars 331
Table 7-9: Seed volume factors used in this analysis for each Option 334
Table 7-10: Retail Price Equivalent Factors in the Heavy-Duty and Light-Duty Industries 337
Table 7-11: Scaling Factors Applied to RPE Contribution Factors to Reflect Changes in their
Contributions, Diesel & CNG Regulatory Classes* 343
Table 7-12: Scaling Factors Applied to RPE Contribution Factors to Reflect Changes in their
Contributions, Gasoline Regulatory Classes 343
Table 7-13: Simplified Example of Indirect Warranty Costs Calculated on an Incremental vs.
Absolute Technology Package Cost (values are not from the analysis and are for presentation
only) 344
Table 7-14: MY2027 Technology Costs for LHD2b3 Diesel, Average per Vehicle, 2017 Dollars
Error! Bookmark not defined.
Table 7-15: MY2031 Technology Costs for LHD2b3 Diesel, Average per Vehicle, 2017 Dollars
Error! Bookmark not defined.
xx
-------
Table 7-16: MY2027 Technology Costs forLHD45 Diesel, Average per Vehicle, 2017 Dollars
Error! Bookmark not defined.
Table 7-17: MY2031 Technology Costs forLHD45 Diesel, Average per Vehicle, 2017 Dollars
Error! Bookmark not defined.
Table 7-18: MY2027 Technology Costs for MHD67 Diesel, Average per Vehicle, 2017 Dollars*
Error! Bookmark not defined.
Table 7-19: MY2031 Technology Costs for MHD67 Diesel, Average per Vehicle, 2017 Dollars*
Error! Bookmark not defined.
Table 7-20 MY2027 Technology Costs for HHD8 Diesel, Average per Vehicle, 2017 Dollars
Error! Bookmark not defined.
Table 7-21: MY2031 Technology Costs for HHD8 Diesel, Average per Vehicle, 2017 Dollars
Error! Bookmark not defined.
Table 7-22: MY2027 Technology Costs for Urban bus Diesel, Average per Vehicle, 2017
Dollars Error! Bookmark not defined.
Table 7-23: MY2031 Technology Costs for Urban bus Diesel, Average per Vehicle, 2017
Dollars Error! Bookmark not defined.
Table 7-24: MY2027 Technology Costs for LHD45, MHD67 & HHD8 Gasoline, Average per
Vehicle, 2017 Dollars Error! Bookmark not defined.
Table 7-25: MY2031 Technology Costs for LHD45, MHD67 & HHD8 Gasoline, Average per
Vehicle, 2017 Dollars Error! Bookmark not defined.
Table 7-26: MY2027 Technology Costs forHHD8 CNG, Average per Vehicle, 2017 Dollars
Error! Bookmark not defined.
Table 7-27: MY2031 Technology Costs forHHD8 CNG, Average per Vehicle, 2017 Dollars
Error! Bookmark not defined.
Table 7-28: MY2027 Technology Costs for Urban bus CNG, Average per Vehicle, 2017 Dollars
Error! Bookmark not defined.
Table 7-29: MY2031 Technology Costs for Urban bus CNG Average per Vehicle, 2017 Dollars
Error! Bookmark not defined.
Table 7-30: Diesel Exhaust Fluid Consumption Rates for Diesel Vehicles in the Baseline Case
350
Table 7-31: Derivation of DEF Consumption per Ton of NOx Reduced 350
Table 7-32: Diesel Exhaust Fluid Price per Gallon (2017 dollars) 351
Table 7-33: Diesel Exhaust Fluid Average Cost per Mile during the first 10 Years of each MY
Lifetime by MOVES Sourcetypes equipped with LHD Engines {cents/mile in 2017 dollars,
Undiscounted) * Error! Bookmark not defined.
Table 7-34: Diesel Exhaust Fluid Average Cost per Mile during the first 10 Years of each MY
Lifetime by MOVES Sourcetypes equipped with LHD45 Engines {cents/mile in 2017 dollars,
Undiscounted) Error! Bookmark not defined.
Table 7-35: Diesel Exhaust Fluid Average Cost per Mile during the first 10 Years of each MY
Lifetime by MOVES Sourcetypes equipped with MHD67 Engines {cents/mile in 2017 dollars,
Undiscounted) Error! Bookmark not defined.
xxi
-------
Table 7-36: Diesel Exhaust Fluid Average Cost per Mile during the first 10 Years of each MY
Lifetime by MOVES Sourcetypes equipped with HHD8 Engines {cents/mile in 2017 dollars,
Undiscounted) Error! Bookmark not defined.
Table 7-37: Diesel Exhaust Fluid Average Cost per Mile during the first 10 Years of each MY
Lifetime by MOVES Sourcetypes equipped with Urban Bus Engines {cents/mile in 2017
dollars, Undiscounted) Error! Bookmark not defined.
Table 7-38: Gasoline Vehicle Average Retail Fuel Cost per Mile during the first 10 Years of
each Model Year Lifetime by MOVES Sourcetype equipped with LHD45 Engines*
{cents/mile in 2017 dollars, Undiscounted) Error! Bookmark not defined.
Table 7-39: Gasoline Vehicle Average Retail Fuel Cost per Mile during the first 10 Years of
each Model Year Lifetime by MOVES Sourcetype equipped with MHD67 Engines*
{cents/mile in 2017 dollars, Undiscounted) Error! Bookmark not defined.
Table 7-40: Gasoline Vehicle Average Retail Fuel Cost per Mile during the first 10 Years of
each Model Year Lifetime by MOVES Sourcetype equipped with HHD8 Engines* {cents/mile
in 2017 dollars, Undiscounted) Error! Bookmark not defined.
Table 7-41: Class-8 Diesel Repair & Maintenance Costs per Mile (100,000 miles per year, 2018
dollars*) Error! Bookmark not defined.
Table 7-42: Estimated Vehicle Age at the End of the Warranty Period and the Useful Life for
each Heavy Heavy-duty Diesel Combination Truck for the Baseline Case ..Error! Bookmark
not defined.
Table 7-43: Class-8 Diesel Short-haul Combination Truck Repair & Maintenance Costs per Mile
(see Table 7-41) with MOVES HHD8 Mileage Accumulation Error! Bookmark not defined.
Table 7-44: Class-8 Diesel Long-haul Combination Truck Repair & Maintenance Costs per Mile
(see Table 7-41) with MOVES HHD8 Mileage Accumulation Error! Bookmark not defined.
Table 7-45: Percentage of Total Repair & Maintenance Costs Attributable to Different Vehicle
Systems Error! Bookmark not defined.
Table 7-46: Diesel Emission Repair Costs, Average Cost per Mile during the first 10 Years of
each MY Lifetime by MOVES Sourcetypes equipped with LHD2b3 Engines {cents/mile in
2017 dollars, Undiscounted)* Error! Bookmark not defined.
Table 7-47: Diesel Emission Repair Costs, Average Cost per Mile during the first 10 Years of
each MY Lifetime by MOVES Sourcetypes equipped with LHD45 Engines {cents/mile in
2017 dollars, Undiscounted) * Error! Bookmark not defined.
Table 7-48: Diesel Emission Repair Costs, Average Cost per Mile during the first 10 Years of
each MY Lifetime by MOVES Sourcetypes equipped with MHD67 Engines {cents/mile in
2017 dollars, Undiscounted) * Error! Bookmark not defined.
Table 7-49: Diesel Emission Repair Costs, Average Cost per Mile during the first 10 Years of
each MY Lifetime by MOVES Sourcetypes equipped with HHD8 Engines {cents/mile in 2017
dollars, Undiscounted) * Error! Bookmark not defined.
Table 7-50: Diesel Emission Repair Costs, Average Cost per Mile during the first 10 Years of
each MY Lifetime by MOVES Sourcetypes equipped with Urban Bus Engines {cents/mile in
2017 dollars, Undiscounted) * Error! Bookmark not defined.
XXll
-------
Table 7-51: Gasoline Emission Repair Costs, Average Cost per Mile during the first 10 Years of
each MY Lifetime by MOVES Sourcetypes equipped with LHD45 Engines {cents/mile in
2017 dollars, Undiscounted)* Error! Bookmark not defined.
Table 7-52: Gasoline Emission Repair Costs, Average Cost per Mile during the first 10 Years of
each MY Lifetime by MOVES Sourcetypes equipped with MHD67 Engines {cents/mile in
2017 dollars, Undiscounted)* Error! Bookmark not defined.
Table 7-53: Gasoline Emission Repair Costs, Average Cost per Mile during the first 10 Years of
each MY Lifetime by MOVES Sourcetypes equipped with HHD8 Engines {cents/mile in 2017
dollars, Undiscounted)* Error! Bookmark not defined.
Table 7-54: CNG Emission Repair Costs, Average Cost per Mile during the first 10 Years of
each MY Lifetime by MOVES Sourcetypes equipped with HHD8 Engines {cents/mile in 2017
dollars, Undiscounted)* Error! Bookmark not defined.
Table 7-55: CNG Emission Repair Costs, Average Cost per Mile during the first 10 Years of
each MY Lifetime by MOVES Sourcetypes equipped with Urban Bus Engines {cents/mile in
2017 dollars, Undiscounted) * Error! Bookmark not defined.
Table 7-56: Technology Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Diesel, Millions of 2017 dollars * 365
Table 7-57: Technology Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Diesel, Millions of 2017 dollars * Error! Bookmark not defined.
Table 7-58: Technology Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Diesel by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
Table 7-59: Technology Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Diesel by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
Table 7-60: Technology Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Diesel by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
Table 7-61: Technology Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Diesel by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
Table 7-62: Technology Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Gasoline, Millions of 2017 dollars * Error! Bookmark not defined.
Table 7-63: Technology Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Gasoline, Millions of 2017 dollars * Error! Bookmark not defined.
Table 7-64: Technology Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Gasoline by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
Table 7-65: Technology Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Gasoline by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
XXlll
-------
Table 7-66: Technology Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Gasoline by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
Table 7-67: Technology Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Gasoline by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
Table 7-68: Technology Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
CNG, Millions of 2017 dollars * Error! Bookmark not defined.
Table 7-69: Technology Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
CNG, Millions of 2017 dollars * Error! Bookmark not defined.
Table 7-70: Technology Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
CNG by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
Table 7-71: Technology Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
CNG by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
Table 7-72: Technology Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
CNG by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
Table 7-73: Technology Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
CNG by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
Table 7-74: Technology Cost Impacts Relative to the Baseline Case, All Fuels by Proposed
Option, Millions of 2017 dollars, Present Values at 3% Discounting * . Error! Bookmark not
defined.
Table 7-75: Technology Cost Impacts Relative to the Baseline Case, All Fuels by Proposed
Option, Millions of 2017 dollars, Present Values at 7% Discounting * . Error! Bookmark not
defined.
Table 7-76: Operating Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Diesel, Millions of 2017 dollars * 367
Table 7-77: Operating Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Diesel, Millions of 2017 dollars * Error! Bookmark not defined.
Table 7-78: Operating Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Diesel by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
Table 7-79: Operating Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Diesel by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
Table 7-80: Operating Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Diesel by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
xxiv
-------
Table 7-81: Operating Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Diesel by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
Table 7-82: Operating Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Gasoline, Millions of 2017 dollars Error! Bookmark not defined.
Table 7-83: Operating Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Gasoline, Millions of 2017 dollars Error! Bookmark not defined.
Table 7-84: Operating Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Gasoline by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
Table 7-85: Operating Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Gasoline by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
Table 7-86: Operating Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
Gasoline by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
Table 7-87: Operating Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
Gasoline by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
Table 7-88: Operating Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
CNG, Millions of 2017 dollars Error! Bookmark not defined.
Table 7-89: Operating Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
CNG, Millions of 2017 dollars Error! Bookmark not defined.
Table 7-90: Operating Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
CNG by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
Table 7-91: Operating Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
CNG by Regulatory Class, Millions of 2017 dollars, Present Values at 3% Discounting *
Error! Bookmark not defined.
Table 7-92: Operating Cost Impacts of the Proposed Option 1 Relative to the Baseline Case,
CNG by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
Table 7-93: Operating Cost Impacts of the Proposed Option 2 Relative to the Baseline Case,
CNG by Regulatory Class, Millions of 2017 dollars, Present Values at 7% Discounting *
Error! Bookmark not defined.
Table 7-94: Operating Cost Impacts Relative to the Baseline Case, All Fuels by Proposed
Option, Millions of 2017 dollars, Present Values at 3% Discounting * . Error! Bookmark not
defined.
Table 7-95: Operating Cost Impacts Relative to the Baseline Case, All Fuels by Proposed
Option, Millions of 2017 dollars, Present Values at 7% Discounting * . Error! Bookmark not
defined.
Table 7-96: Fuel Cost and Transfer Impacts of the Proposed Option 1 Relative to the Baseline
Case, Gasoline, Millions of 2017 dollars 369
xxv
-------
Table 7-97: Fuel Cost and Transfer Impacts of Proposed Option 2 Relative to the Baseline Case,
Gasoline, Millions of 2017 dollars Error! Bookmark not defined.
Table 7-98: Total Technology & Operating Cost Impacts of the Proposed Option 1 Relative to
the Baseline Case, All Regulatory Classes and All Fuels, Millions of 2017 dollars 370
Table 7-99: Total Technology & Operating Cost Impacts of Proposed Option 2 Relative to the
Baseline Case, All Regulatory Classes and All Fuels, Millions of 2017 dollars 371
Table 8-1: Summary of CMAQ-Derived Population-Weighted Ozone and PM2.5 Air Quality
Metrics for Health Benefits Endpoints Associated with Proposed Option 1 377
Table 8-2: Health Effects of Ambient Ozone and PM2.5 379
Table 8-3: Estimated Avoided PM2.5 Mortality and Illnesses in 2045 for Proposed Option 1 (95%
Confidence Interval) 3,13 388
Table 8-4: Estimated Avoided Ozone Mortality and Illnesses in 2045 for the Proposed Option 1
(95% Confidence Interval)21 389
Table 8-5: Estimated Economic Value of PM2.5- and Ozone-Attributable Premature Mortality
and Illnesses in 2045 for Proposed Option 1 (95% Confidence Interval; millions of 2017$)a
390
Table 8-6: Total Ozone and PM2.5-Attributable Benefits in 2045 for Proposed Option 1 (95%
Confidence Interval; billions of 2017$)a'b 391
Table 8-7: Benefits Inputs that Change Over Time used to Calculate Year-over-Year Estimates
392
Table 8-8: Undiscounted Stream and Present Value of Human Health Benefits from 2027
through 2045: Monetized Benefits Quantified as Sum of Short-Term Ozone Respiratory
Mortality Ages 0-99, and Long-Term PM2.5 All-Cause Mortality Ages 30+ (Discounted at 3%
and 7%; billions of 2017$)a b 393
Table 8-9: Undiscounted Stream and Present Value of Human Health Benefits from 2027
through 2045: Monetized Benefits Quantified as Sum of Long-Term Ozone Respiratory
Mortality Ages 30+, and Long-Term PM2.5 All-Cause Mortality Ages 65+ (Discounted at 3%
and 7%; billions of 2017$)a'b 394
Table 8-10: Unquantified Criteria Pollutant Health and Welfare Benefits Categories 396
Table 9-1: 2045 Annual Value, Present Value and Equivalent Annualized Value of Costs,
Benefits and Net Benefits of the Proposed Options 1 and 2 (billions, 2017$)a'b 404
Table 10-1: Pre and Low-Buy Sales Effects Coefficients 410
Table 10-2: Regulatory Costs and HD Vehicle Prices Used to Estimate Elasticities 411
Table 10-3: Elasticity Estimates 412
Table 10-4: Illustrative Pre-Buy Results from the 2031 Implementation Date 414
Table 10-5: Illustrative Low-Buy Results from the 2031 Implementation Date 415
Table 10-6: Sectors Used in this Analysis Error! Bookmark not defined.
Table 10-7: Employment per $1 Million Expenditures (2017$) in the Motor Vehicle
Manufacturing Sectora Error! Bookmark not defined.
Table 10-8: Estimated Employment Effects Due to Increased Costs of Vehicles and Parts (Cost
Effect), in Job-Years Error! Bookmark not defined.
xxvi
-------
Table 11-1: Summary of Impacts on Small Businesses Subject to the Rule
xxvii
-------
List of Figures
Figure 1-1: Functional schematic showing relative positioning of exhaust emission control
components arranged within an in-line exhaust system (top) and integrated into a box-style
system (bottom) 7
Figure 1-2: Integrated series heavy-duty truck exhaust emission control systems from Cummins
Emission Solutions (top) and box-style system from Eberspacher (bottom), with cut-away
showing some of the internal components (bottom right) 8
Figure 1-3: NOx reduction efficiency of an early, developmental Cu-zeolite SCR formulation
relative to DOC and SCR inlet temperatures and SCR space velocity (adapted from
McDonald et al. 2011) 12
Figure 1-4: Potential layout of a 2027+ dual-SCR system in an in-line configuration (top) and
comparable components integrated to improve passive thermal management (bottom) 19
Figure 1-5: Tailpipe Exhaust and Crankcase Emission Rates from Two Heavy-Duty Diesel
Trucks 21
Figure 1-6: Schematic of an ORVR system 35
Figure 1-7: SCR performance over the hot start HDDE FTP61' 41
Figure 1-8: SCR NOx conversion for the first inch of aged SCR catalysts.54 43
Figure 1-9: SCR NOx conversion for the seventh inch of aged SCR catalysts.54 43
Figure 2-1: Window size comparison and load distribution profile 64
Figure 2-2: Example of a LLC 7 Candidate Cycle 66
Figure 2-3: Sample of valid and invalid NTE events, separated by exclusion zones 69
Figure 2-4: NOx sensor time to on after engine cold-start 71
Figure 2-5: MOVES Operating Modes (OpModes) by scaled tractive power and vehicle speed. 73
Figure 2-6: MOVES OpMode Emission Rates from HHD Engine Broken Down by Engine
Family 74
Figure 2-7: NOx Emission Rates from 81 Vehicles with 0.20 g/bhp-hr FEL HHD Engines by
MOVES OpMode and Aftertreatment Temperature 75
Figure 2-8: NOx Emission Rates from 20 Vehicles with 0.20 g/bhp-hr FEL MHD Engines by
MOVES OpMode and Aftertreatment Temperature 75
Figure 2-9: NOx Emission Rates from 42 Vehicles with 0.20 g/bhp-hr FEL LHD Engines by
MOVES OpMode and Aftertreatment Temperature 76
Figure 2-10: Time Fraction from 81 Vehicles with 0.20 g/bhp-hr FEL HHD Engines by MOVES
OpMode and Aftertreatment Temperature 77
Figure 2-11: Time Fraction from 20 Vehicles with 0.20 g/bhp-hr FEL MHD Engines by MOVES
OpMode and Aftertreatment Temperature 77
Figure 2-12: Time Fraction from 42 Vehicles with 0.20 g/bhp-hr FEL LHD Engines by MOVES
OpMode and Aftertreatment Temperature 78
Figure 2-13: Total NOx Contribution from 81 vehicles with 0.20 g/bhp-hr FEL HHD Engines by
MOVES OpMode and Aftertreatment Temperature 79
xxviii
-------
Figure 2-14: Total NOx Contribution from 20 vehicles with 0.20 g/bhp-hr FEL MHD Engines by
MOVES OpMode and Aftertreatment Temperature 79
Figure 2-15: Total NOx Contribution from 42 vehicles with 0.20 g/bhp-hr FEL LHD Engines by
MOVES OpMode and Aftertreatment Temperature 80
Figure 2-16: Brake-specific NOx by Vehicle Speed Bins for 93 Vehicles with HHD Diesel
Engines and an FEL of 0.20 g/bhp-hr 81
Figure 2-17: Brake-specific NOx by Vehicle Speed Bins for 26 Vehicles with MHD Diesel
Engines and an FEL of 0.20 g/bhp-hr 82
Figure 2-18: Brake-specific NOx by Vehicle Speed Bins for 49 Vehicles with LHD Diesel
Engines and an FEL of 0.20 g/bhp-hr 82
Figure 2-19: Brake-specific NOx by Window Average Power Bins for 85 Vehicles with HHD
Diesel Engines and an FEL of 0.20 g/bhp-hr 83
Figure 2-20: Vehicle Speed Profile of HD Duty Cycles 84
Figure 2-21: MOVES OpMode Time Fraction for each Simulated HD Combination Long-Haul
Duty Cycle 85
Figure 2-22: Brake-specific NOx emissions by simulated cycle for HHD diesel engines with
NOx FEL of 0.20 g/bhp-hr 86
Figure 2-23: Distance-specific NOx emissions by simulated cycle for HHD diesel engines with
NOx FEL of 0.20 g/bhp-hr 86
Figure 2-24: Number of windows in each bin from 168 HDUIT shift days, sorted in order (the
first 30% are shown). The black lines indicate 2400 windows and 1200 windows. Note some
HDUIT tests have multiple bins containing fewer than 2400 windows 91
Figure 2-25: Range of NOx rates in windows in each bin, using data from 93 HDIUT results in
the HHD, 0.2 FEL category. The heavy line is the median value from all tests, with error bars
represent the 25th and 75th percentile of the data Error! Bookmark not defined.
Figure 2-26: Estimated Projected Ford E-450 ORVR Results based on Extrapolation 96
Figure 2-27: Estimated Projected Isuzu ORVR Results based on ExtrapolationError! Bookmark
not defined.
Figure 3-1: Developmental Cummins X15 Engine equipped with individual cylinder deactivation
undergoing engine dynamometer testing as part of the CARB Stage 3 research at SwRI 109
Figure 3-2: Schematic layout (not to scale) of the dual-SCR EAS tested as part of the CARB
Stage 3 research at SwRI 110
Figure 3-3: Developmental EAS with light-off SCR installed in engine dynamometer test cell at
SwRI (upper left, upper right) and details of the downstream, single "box" unit (lower left,
lower right) 110
Figure 3-4: The average pressure drop across the DPF on the SET for the degreened
aftertreatment and the equivalent of 290,000 miles of operation aftertreatment 115
Figure 3-5: Schematic layout (not to scale) of the dual-SCR EAS tested as part of the EPA Stage
3 research at SwRI 116
Figure 3-6: CARB Southern Route Cycle 119
Figure 3-7: Grocery Delivery Truck Cycle 120
XXIX
-------
Figure 3-8: Dray age Truck Cycle 120
Figure 3-9: Euro-VI ISC Cycle 121
Figure 3-10: ACES 4-hour Cycle 121
Figure 3-11: EAS with close-coupled light-off SCR 127
Figure 3-12: EAS with light-off SCR integrated into an under-cab mounting position. This
system is designed to be installed in a Navistar Day cab which is shown in the upper right. .128
Figure 3-13: EPA developmental MY2018 Cummins X15 Heavy HDE.... Error! Bookmark not
defined.
Figure 3-14: Variable reluctance sensor for valve position measurement. The final installation
will include valve position measurement at the valve bridges for each of the six cylinders.
Error! Bookmark not defined.
Figure 3-15: Example of engine-speed, engine load, and resulting SCR inlet temperature used
over the DAAAC 130
Figure 3-16: FTP composite and SET NOx emissions results including IRAF for the EPA Stage
3 developmental engine and emissions control system versus equivalent miles of operation.
Error! Bookmark not defined.
Figure 3-17: LLC NOx emissions results including IRAF for the EPA Stage 3 developmental
engine and emissions control system versus equivalent miles of operation. .Error! Bookmark
not defined.
Figure 3-18: FTP-75, Cold start three phase test cycle 140
Figure 3-19: EPA Highway Fuel Economy Cycle 141
Figure 3-20: EPA 4 phase LA92 test cycle 142
Figure 3-21: Super cycle, GEM greenhouse gas cycle 143
Figure 3-22: FTP-75 Cold start cumulative HC comparison 144
Figure 3-23: FTP-75 Cold start cumulative NOx comparison 144
Figure 3-24: FTP-75 Catalyst light-off time comparison 145
Figure 3-25: Extended idle catalyst cool down comparison 146
Figure 3-26: Comparison of emissions performance for MY 2019 certified HD SI engines;
Overall Average (grey dash) includes data from all six engines, Subset Average (yellow dash)
includes data from the three engines with the best NOx performance (Cert E4-E6) Error!
Bookmark not defined.
Figure 3-27: Comparison of operating modes in the fuel mapping-based SET 151
Figure 3-28: Engine RPM Down-Speeding FTP CO Comparison 155
Figure 3-29: Engine RPM Down-Speeding SET CO Comparison 156
Figure 4-1: Important Factors Involved in Seeing a Scenic Vista (Malm, 2016) 187
Figure 4-2: Mandatory Class I Federal Areas in the U.S 188
Figure 4-3: Nitrogen and Sulfur Cycling, and Interactions in the Environment 191
Figure 5-1: Percent change in in-use emission rates for 2010 standard (0.2 g/hp-hr) compliant
HHD vehicles, compared to the percent change in the 2010 duty-cycle standard across
MOVES operating modes Error! Bookmark not defined.
xxx
-------
Figure 5-2: Duty-cycle-based running NOx emissions, ERduty> use, for HHD diesel for the
control scenarios Error! Bookmark not defined.
Figure 5-3: Base NOx rates and off-cycle NOx standard compliant emission rates for HHD diesel
Error! Bookmark not defined.
Figure 5-4: Comparison of Running NOx emission rates for diesel-fueled HHD compliant with
the MY2030 proposed Option 1 duty-cycle and off-cycle standards Error! Bookmark not
defined.
Figure 5-5: Estimated zero-mile NOx emission rates for HHD diesel vehicles due to the proposed
Option 1, Option 2 and Alternative duty-cycle and off-cycle standards. Error! Bookmark not
defined.
Figure 5-6: Methodology to model the effects of tampering and mal-maintenance (T&M) on
emission rates according to warranty and useful life Error! Bookmark not defined.
Figure 5-7: NOx g/mile emission rates estimated from MOVES for HHD diesel long-haul
combination trucks for the model year 2027 fleet across vehicle age for the baseline and
control scenarios Error! Bookmark not defined.
Figure 5-8: NOx g/mile emission rates estimated from MOVES for HHD diesel long-haul
combination trucks for the model year 2031 fleet across vehicle age for the baseline and
control scenarios Error! Bookmark not defined.
Figure 5-9: Estimated relationship between the HHD NOx 12-hour cold-start and the composite
FTP NOx standards Error! Bookmark not defined.
Figure 5-10: Duty-cycle-based NOx start emissions for HHD Diesel for the baseline, proposed,
and alternative control scenarios Error! Bookmark not defined.
Figure 5-11: Duty-cycle-based running NOx emission rates for LHD gasoline for the control
scenarios Error! Bookmark not defined.
Figure 5-12: LHD gasoline Duty-cycle-based running THC emission rates for LHD gasoline for
the control scenarios Error! Bookmark not defined.
Figure 5-13: Modeling process of onroad emissions as part of the input for air quality modeling
Error! Bookmark not defined.
Figure 5-14: Estimated zero-mile emission rates for LHD45 diesel vehicles due to the proposed
and alternative duty-cycle and off-cycle standards Error! Bookmark not defined.
Figure 5-15: Estimated zero-mile emission rates for MHD diesel vehicles due to the proposed
and alternative duty-cycle and off-cycle standards Error! Bookmark not defined.
Figure 5-16: National NOx Emission Reductions from Heavy-duty Vehicles in Calendar Years
2030, 2040, and 2045 — Proposed Option 1 Program Reductions from the Baseline for Each
Fuel Type Category by Emission Process Error! Bookmark not defined.
Figure 5-17: National NOx Emission Reductions from Heavy-duty Vehicles in Calendar Years
2030, 2040, and 2045 — Proposed Option 1 Program Reductions from the Baseline for Each
Fuel Type Category by HD Regulatory Class Error! Bookmark not defined.
Figure 5-18: National VOC Emission Reductions from Heavy-duty Vehicles in Calendar Years
2030, 2040, and 2045 — Proposed Option 1 Program Reductions from the Baseline for Each
Fuel Type Category by Emission Process Error! Bookmark not defined.
XXXI
-------
Figure 5-19: National VOC Emission Reductions from Heavy-duty Vehicles in Calendar Years
2030, 2040, and 2045 — Proposed Option 1 Program Reductions from the Baseline for Each
Fuel Type Category by HD Regulatory Class Error! Bookmark not defined.
Figure 5-20: National Exhaust PM2.5 Emission Reductions from Heavy-duty Vehicles in
Calendar Years 2030, 2040, and 2045 — Proposed Option 1 Program Reductions from the
Baseline for Each Fuel Type Category by Emission Process... Error! Bookmark not defined.
Figure 5-21: National Exhaust PM2.5 Emission Reductions from Heavy-duty Vehicles in
Calendar Years 2030, 2040, and 2045 — Proposed Option 1 Program Reductions from the
Baseline for Each Fuel Type Category by HD Regulatory Class Error! Bookmark not
defined.
Figure 5-22: National CO Emission Reductions from Heavy-duty Vehicles in Calendar Years
2030, 2040, and 2045 — Proposed Option 1 Program Reductions from the Baseline for Each
Fuel Type Category by Emission Process Error! Bookmark not defined.
Figure 5-23: National CO Emission Reductions from Heavy-duty Vehicles in Calendar Years
2030, 2040, and 2045 — Proposed Option 1 Program Reductions from the Baseline for Each
Fuel Type Category by HD Regulatory Class Error! Bookmark not defined.
Figure 5-24: Comparison of Calendar Year 2045 Onroad Heavy-Duty NOx Emissions from
Different Engine Operational Process for the Baseline, Proposed Option 1, and Proposed
Option 2 Standards Error! Bookmark not defined.
Figure 5-25: National Heavy-duty Vehicle NOx Emissions (Annual US Tons) For Calendar
Years Between 2027 and 2045 Error! Bookmark not defined.
Figure 5-26: National Heavy-Duty Vehicle VOC Emissions (Annual US Tons) For Calendar
Years Between 2027 and 2045 Error! Bookmark not defined.
Figure 5-27: National Heavy-duty Vehicle PM2.5 (Exhaust Only) Emissions (Annual US Tons)
For Calendar Years Between 2027 and 2045 Error! Bookmark not defined.
Figure 5-28: National Heavy-Duty Vehicle CO Emissions (Annual US Tons) For Calendar Years
Between 2027 and 2045 Error! Bookmark not defined.
Figure 5-29: National Heavy-duty Vehicle NOx Emission Reductions (Annual US Tons)
Relative to the Baseline Case For Omnibus Nationwide Scenarios (Symbols) As Compared
with EPA Proposed Option 1 (Line) Error! Bookmark not defined.
Figure 6-1: 8-Hour Ozone Nonattainment Areas (2008 Standard) 285
Figure 6-2: 8-Hour Ozone Nonattainment Areas (2015 Standard) 285
Figure 6-3: Counties Designated Nonattainment for PM2.5 (1997, 2006, and/or 2012 standards)
287
Figure 6-4: Map of the CMAQ 12 km modeling domain (noted by the purple box) 291
Figure 6-5: Projected Change in 8-hour Ozone Design Values in 2045 due to Proposed Rule ..294
Figure 6-6: Projected Change in Annual PM2.5 Design Values in 2045 due to Proposed Rule ..297
Figure 6-7: Projected Change in 24-hour PM2.5 Design Values in 2045 due to Proposed Rule .300
Figure 6-8: Projected Absolute Change in Annual Ambient N02 Concentrations in 2045 304
Figure 6-9: Percent Change in Annual Ambient NO2 Concentrations in 2045 304
Figure 6-10: Absolute Change in Annual Ambient CO Concentrations in 2045 305
XXXll
-------
Figure 6-11: Percent Change in Annual Ambient CO Concentrations in 2045 306
Figure 6-12: Changes in Ambient Benzene Concentrations in 2045 due to Proposed Rule:
Absolute Changes in |ig/m3 (left) and Percent Changes (right) 307
Figure 6-13: Changes in Ambient Naphthalene Concentrations in 2045 due to Proposed Rule:
Absolute Changes in |ig/m3 (left) and Percent Changes (right) 307
Figure 6-14: Changes in Ambient Acetaldehyde Concentrations in 2045 due to Proposed Rule:
Absolute Changes in |ig/m3 (left) and Percent Changes (right) 308
Figure 6-15: Changes in Ambient Formaldehyde Concentrations in 2045 due to Proposed Rule:
Absolute Changes in |ig/m3 (left) and Percent Changes (right) 308
Figure 6-16: Absolute Change in Annual Deposition of Nitrogen in 2045 310
Figure 6-17: Percent Change in Annual Deposition of Nitrogen in 2045 310
Figure 6-18: Distributional maps of populated 12km grid cells across the contiguous United
States in 2045. Darker areas represent the location of grid cells within the highest 5 percent of
baseline concentrations for (a) PM2.5 and (b) ozone 313
Figure 7-1: Costs Relative to First Year Costs using Different Seed Volume Factors 332
Figure 7-2: Repair & Maintenance Cost/Mile Curve (2018 dollars) Error! Bookmark not
defined.
Figure 7-3: Emission Repair Cost/Mile Curve (2018 dollars) Error! Bookmark not defined.
Figure 7-4: Emission Repair Cost Curve in the Baseline Case (dashed line) and the proposed
Option 1 (solid line) for a Short-haul Combination Truck equipped with a HHD Class-8
Diesel Engine Error! Bookmark not defined.
Figure 8-1: Stylized Relationship between the PM2.5 Concentrations Considered in Epidemiology
Studies and our Confidence in the Estimated PM-related Premature Deaths Error! Bookmark
not defined.
Figure 8-2: Estimated Percentage of PM2.5-Related Deaths (Turner et al. 2016) and Number of
Individuals Exposed (30+) by Annual Mean PM2.5 Level in 2045 387
Figure 11-1: Secondary Vehicle Manufacturers, Estimated Impacts as a Percent of Annual
Revenues 437
XXXlll
-------
Executive Summary
The Environmental Protection Agency (EPA) is finalizing new standards and changes to the
heavy-duty highway engine and vehicle emissions control program in order to reduce emissions
of oxides of nitrogen (NOx), particulate matter (PM), hydrocarbons (HC), and carbon monoxide
(CO). These emissions contribute to ozone and PM and their resulting threat to public health,
which includes premature death, respiratory illness (including childhood asthma), cardiovascular
problems, and other adverse health impacts.
This RIA is generally organized to provide overall background information, methodologies,
and data inputs, followed by results of the various analyses. A summary of each chapter of the
RIA follows.
Chapter 1 describes key technologies that manufacturers could use to meet more stringent
emissions standards for NOx, PM, HC, and CO. The chapter focuses on technologies specific to
compression-ignition engines and spark-ignition engines, and also discusses fuel considerations.
Chapter 2 describes the existing test procedures as well as the development process for the
final test procedures for spark- and compression-ignition engine compliance. This includes the
determination of emissions from both engines and hybrid powertrains, as well as the
development of new duty cycles.
Chapter 3 describes the technology feasibility demonstration programs, including engine
technologies and emission control strategies for reducing NOx, PM, NMHC, and CO. The
technologies presented represent potential ways that the industry could meet the final stringency
levels, and they provide the basis for the technology costs and benefits analyses.
Chapter 4 presents a discussion of the health effects associated with exposure ambient
concentrations of ozone, PM, NO2, CO, and air toxics. The discussion of health impacts is
mainly focused on describing the effects of air pollution on the population in general.
Additionally, children are recognized to have increased vulnerability and susceptibility related to
air pollution and other environmental exposures; this and effects for other vulnerable and
susceptible groups are discussed. The chapter also discusses the environmental effects associated
with pollutants affected by this final rulemaking, specifically PM, ozone, NOx and air toxics.
Chapter 5 presents our analysis of the national emissions impacts of the final program for
calendar years 2027 through 2045. In this chapter we quantify emissions from NOx, volatile
organic compounds (VOC), PM2.5, CO, and others. The onroad national emissions inventories
were estimated using the latest public version of EPA's Motor Vehicle Emission Simulator
(MOVES) model (MOVES3). Table ES-1 summarizes the projected reductions in heavy-duty
emissions from the final rule in 2045. In addition to describing the national emission inventories,
this chapter describes the methods used to estimate the spatially and temporally-resolved
emission inventories used to support the air quality modeling analysis documented in Chapter 6.
1
-------
Table ES-1: Projected Heavy-Duty Emission Reductions in 2045 from the Final Program
Pollutant
Percent Reduction in Highway
Heavy-duty Emissions
NOx
48%
Primary PM25
8%
VOC
23%
CO
18%
Chapter 6 presents information on air quality, including a discussion of current air
quality, details related to the methodology used for the air quality modeling analysis, and results
from the air quality modeling analysis. When feasible, we conduct full-scale photochemical air
quality modeling to accurately project levels of criteria and air toxic pollutants, because the
atmospheric chemistry related to ambient concentrations of PM2.5, ozone, and air toxics is very
complex. Air quality modeling was performed for the proposed rule using emission reductions
that compare well with the emission reductions estimated for the final rule, and it demonstrated
improvements in concentrations of air pollutants. Given the similar structure of the proposed
and final programs, we expect consistent geographic distribution of emissions reductions and
modeled improvements in air quality, and that the air quality modeling conducted at the time of
proposal adequately represents the final rule. Specifically, we expect this rule will decrease
ambient concentrations of air pollutants, including significant improvements in ozone
concentrations in 2045 as demonstrated in the air quality modeling analysis. Our analysis
indicates that the largest predicted improvements in both ozone and PM2.5 are estimated to occur
in areas with the worst baseline air quality, and that a substantially larger number of people of
color are expected to reside in these areas. An expanded analysis of the air quality impacts
experienced by specific race and ethnic groups found that non-Hispanic Blacks will receive the
greatest improvement in PM2.5 and ozone concentrations as a result of the standards. The
emission reductions provided by the final standards will be important in helping areas attain and
maintain the National Ambient Air Quality Standards (NAAQS) and prevent future
nonattainment. We also expect reductions in ambient PM2.5, NO2 and CO due to this rule.
Although the spatial resolution of the air quality modeling is not sufficient to quantify it, this
rule's emission reductions will also reduce air pollution in close proximity to major roadways,
where concentrations of many air pollutants are elevated. In addition, the final standards are
expected to result in improvements in nitrogen deposition and visibility.
Chapter 7 presents estimates of the costs associated with the emissions-reduction
technologies that manufacturers could add in response to the final program. We present these not
only in terms of the upfront technology costs per engine as presented in Chapter 3 of this RIA,
but also how those costs would change in the years following implementation. We present the
costs associated with the final program elements of extended regulatory useful life and warranty.
These technology costs are presented in terms of direct manufacturing costs and associated
indirect costs such as warranty and research and development (R&D). The analysis also includes
estimates of the possible operating costs associated with the final program— the addition of new
technology and extension of warranty and useful life periods. All costs are presented in 2017
dollars unless noted otherwise. Table ES-2 presents the technology costs, operating costs and the
sum of the two for the final program in 2045.
2
-------
Table ES-2: Total Program Costs: Undiscounted Annual Costs in 2045 and Annualized Costs through 2045 at
3% and 7% Discount Rates (Billions of 2017 dollars)
Total
Technology
Costs
Total
Operating
Costs
Sum
2045 Annual
$4.1
$0.62
$4.7
Present Value, 3%
$53
$1.4
$55
Present Value, 7%
$38
$0.6
$39
Annualized, 3%
$3.7
$0,099
$3.8
Annualized, 7%
$3.7
$0,058
$3.8
Chapter 8 describes the methods used to estimate health benefits from reducing
concentrations of ozone and PM2.5 For the final rulemaking, we have quantified and monetized
health impacts in 2045, representing projected impacts associated with a year when the program
will be fully implemented and when most of the regulated fleet will have turned over. We also
discuss unquantified benefits associated with the standards that, if quantified and monetized, will
increase the total monetized benefits. Overall, we estimate that the final program will lead to a
substantial decrease in adverse PM2.5- and ozone-related health impacts in 2045. Table ES-3
presents our estimates of total monetized benefits for the final program.
Table ES-3: 2045 Annual Value, Present Value and Equivalent Annualized Value of Benefits of the Final
Program (billions, 2017$)a b
3%
7%
Discount
Discount
2045
$12-$33
$10 -$30
Present Value (2027-2045)
$91 -$260
$53 -$150
Annualized Value
$6.3 -$18
$5.1 -$14
a All benefits estimates are rounded to two significant figures; numbers may not sum due to independent rounding.
The range of benefits (and net benefits) in this table are two separate estimates and do not represent lower- and
upper-bound estimates, though they do reflect a grouping of estimates that yield more and less conservative benefits
totals. The costs and benefits in 2045 are presented in annual terms and are not discounted. However, all benefits in
the table reflect a 3 percent and 7 percent discount rate used to account for cessation lag in the valuation of avoided
premature deaths associated with long-term exposure.
b The benefits associated with the standards presented here do not include the full complement of health and
environmental benefits that, if quantified and monetized, would increase the total monetized benefits.
Chapter 9 compares the estimated range of total monetized health benefits to total costs
associated with the final program. This chapter also presents the range of monetized net benefits
(benefits presented in Chapter 8 minus costs presented in Chapter 7) associated with the same
scenarios (see Table ES-4).
The health- and environmental-related effects associated with heavy-duty vehicle and engine
emissions are a classic example of an externality-related market failure. An externality occurs
when one party's actions impose uncompensated costs on another party. The final standards will
help correct this market failure. EPA expects that implementation of the final rule will provide
society with a substantial net gain in welfare, notwithstanding the health and other benefits we
3
-------
were unable to quantify (see RIA Chapter 8.8 for more information about unquantified
benefits).A
Table ES-4: 2045 Annual Value, Present Value and Equivalent Annualized Value of Costs, Benefits and Net
Benefits of the Final Program (billions, 2017$)a'b
3%
7%
Discount
Discount
Benefits
$12 -$33
$10-$30
2045
Costs
$4.7
$4.7
Net Benefits
$6.9 - $29
$5.8 -$25
Benefits
$91 -$260
$53 -$150
Present Value
Costs
$55
$39
Net Benefits
$36 - $200
$14 - $110
Equivalent
Benefits
$6.3 -$18
$5.1 -$14
Annualized
Costs
$3.8
$3.8
Value
Net Benefits
$2.5 -$14
$1.3 -$11
EPA is required by Executive Order (E.O.) 12866 to estimate the benefits and costs of major
new pollution control regulations. At the same time, EPA notes that this analysis is for purposes
of Executive Order 12866, rather than for purposes of showing that the final rule satisfies the
requirements of the Clean Air Act section 202(a). The Clean Air Act does not require a weighing
of costs and benefits in determining what standards are achievable, and EPA did not do so in
determining what standards to adopt.
Chapter 10 provides an economic analysis of the impacts of the final standards on vehicle
sales and employment. This rulemaking is considered economically significant, because it is
expected to have an annual impact on the economy of $100 million or more but is not expected
to have measurable inflationary or recessionary effects. This chapter presents a peer-reviewed
analysis to develop a relationship between estimated changes in vehicle price due to a new
regulation and corresponding changes in vehicle sales (i.e., pre- and low-buy elasticities). In RIA
Chapter 10.1 we outline an approach to quantify potential impacts on vehicle sales due to new
emission standards; we also illustrate how this method could be used to estimate pre-and low-
buy as a function of the estimated costs of this final rule. Our example results for the final
standards suggest pre- and low-buy for Class 8 trucks may range from zero to approximately 2
percent increase in sales over a period of up to 8 months before the 2027 standards begin (pre-
buy), and a decrease in sales from zero to just under three percent over a period of up to 12
months after the 2027 standards begin (low-buy). Our illustrative analysis suggests that if pre-
buy and low-buy occur, the difference would be very slight and short-lived; we do not expect
long-term fleet turnover impacts from pre-buy or low-buy, including effects on average fleet age.
The employment assessment focuses on the motor vehicle manufacturing and the motor vehicle
parts manufacturing sectors, with some assessment of impacts on additional sectors likely to be
most affected by the standards. The employment assessment includes EPA's qualitative and
quantitative estimates of the partial employment impacts of this rule on regulated industries and
an examination of employment impacts in some closely related sectors.
A EPA does not expect the omission of unquantified benefits to impact the Agency's evaluation of the final program
since unquantified benefits generally scale with the emissions impacts of the final program.
4
-------
Chapter 11 presents our analysis of the potential impacts of the final rule on small entities
that will be subject to the highway heavy-duty engine and vehicle provisions of this final rule.
These are: heavy-duty alternative fuel engine converters and heavy-duty secondary vehicle
manufacturers. Other entities that will be subject to the rule are either not small (e.g., engine and
incomplete vehicle manufacturers) or are not expected to incur any burden from the final rule
(e.g., in sectors other than highway heavy-duty engines and vehicles). Our analysis estimates that
no small entities will experience an impact of 3% or more of their annual revenue as a result of
our final rule.
5
-------
Chapter 1 Technology to Control Emissions from Heavy-Duty Engines
This chapter describes key technologies that manufacturers are likely to use to meet more
stringent emissions standards for oxides of nitrogen (NOx), particulate matter (PM),
hydrocarbons (HC), and carbon monoxide (CO). The chapter introduces technologies specific to
compression-ignition engines and spark-ignition engines, and also discusses fuel considerations,
advanced powertrain technologies, and emission monitoring technologies that may apply across
engine types.
1.1 Compression-Ignition Engine Technologies
The following sections describe the compression-ignition engine technologies that we
considered for reducing criteria pollutant emissions as part of this final rule. Many of the
technologies are described with respect to diesel fuel, but they are expected to be broadly
applicable to all fuels used in compression-ignition engines. Our compression-ignition engine
feasibility demonstration for this final rule is based on some of the technologies presented in this
section.
Chapter 2 of this RIA describes the updated test procedures for diesel-ignition engine
certification. Chapter 3 describes the compression-ignition engine feasibility demonstration
program, including a description of the specific technology packages we evaluated, the
effectiveness of those technologies relative to the final standards and corresponding test
procedures, and our projected direct manufacturing cost of those technologies.
1.1.1 Current Catalyst Technologies
This section addresses technologies that, based on our current understanding, are anticipated
to be available in the 2024 to 2030 timeframe to reduce emissions and ensure robust in-use
compliance. The following discussion introduces the technologies and emission reduction
strategies we considered for the final rulemaking, including thermal management technologies
that can be used to better achieve and maintain adequate catalyst temperatures, and the next
generation of catalyst configurations and formulations that will improve catalyst performance
across a broader range of engine operating conditions.
Modern diesel engines rely heavily upon catalytic exhaust aftertreatment systems (EAS) to
meet exhaust emission standards. Current (MY2018-2022) heavy-duty diesel EAS consist of a
diesel oxidation catalyst (DOC) followed by a catalyzed diesel particulate filter (CDPF), a urea
injector, a urea mixer or other decomposition component, and then one or more selective
catalytic reduction (SCR) monolithic substrates (Figure 1-1, Figure 1-2). Such systems are
capable of reducing PM emissions by greater than 95% under most operating conditions and are
capable of reducing NOx emissions by 90 to 98% at exhaust temperatures above approximately
250 °C.
Unreacted ammonia downstream of the SCR is typically referred to as "ammonia slip". An
ammonia slip catalyst (ASC) can be zone-coated onto the outlet of the rearmost SCR substrate
(the case for most LHDDE and MHDDE and some HHDDE applications) or can be coated onto
a separate catalyst substrate (some HHDDE applications) and uses platinum-group-metal (PGM)
exchanged zeolites to promote reaction of ammonia remaining downstream of the SCR catalysts.
Ammonia is an important air toxic compound and can also contribute to secondary PM
6
-------
formation. The use of closed-loop feedback electronic control of urea dosing using zirconia NOx
sensors for NOx and ammonia feedback and the use of ASC together can reduce ammonia
emissions from modern EAS-equipped heavy-duty diesel engines to less than 4 mg/bhp-hr.1
Some LHDDE applications using chassis dynamometer certification place the urea injector and
SCR between the DOC and CDPF or combine SCR and CDPF functionality into one catalyst,
sometimes referred to as selective catalytic reduction on filter (SCRF).
DOC
CDPF
Urea Injector
SCR + ASC
(zone coated)
SCR
lexhaust flow^
17
Same components integrated
into box-style muffler
Figure 1-1: Functional schematic showing relative positioning of exhaust emission control components
arranged within an in-line exhaust system (top) and integrated into a box-style system (bottom).
7
-------
Figure 1-2: Integrated/series heavy-duty truck exhaust emission control system from Cummins Emission
Solutions (top) and box-style sy stem from Ebcrspachcr (bottom), with cut-away showing some of the internal
components (bottom right).A
The DOC, SCR, and ASC typically use cordierite ceramic flow-through monolithic substrates
that are wash-coated with active materials. The CDPF uses a wall-flow substrate made of either
cordierite, silicon carbide (SiC), or aluminum titanate (AbTiOs) for exhaust filtration (or
"trapping") of particulate matter that is coated with active materials. Alternating cells of the
wall-flow substrate are blocked, forcing the exhaust to flow through the porous substrate wall.
The particulate matter, consisting primarily of elemental carbon soot, is filtered from the exhaust
A Disclaimer: Any mention of trade names, manufacturers or products does not imply an endorsement by the United
States Government or the U.S. Environmental Protection Agency. EPA and its employees do not endorse any
commercial products, services, or enterprises.
8
-------
flow onto and within the wall of the CDPF and can then be oxidized to CO2 using either passive
regeneration with nitrogen dioxide (NO2) or active regeneration in the presence of excess oxygen
in the exhaust. Passive regeneration of the CDPF depends on oxidation of a fraction of nitric
oxide (NO) emissions in the exhaust to NO2. Soot oxidation using NO2 occurs at exhaust
temperatures of approximately 250 °C, thus it does not require external heat addition to the
exhaust under most operating conditions. Active regeneration using excess oxygen in the
exhaust occurs at exhaust temperatures above 500 °C - 600 °CB, and thus requires adding heat to
the exhaust. This can be accomplished using one of several different approaches:
• Late, in-cylinder, post-injection of fuel after the primary combustion event and
subsequent heat addition from the exothermic reaction of the excess fuel over the
DOC and CDPF.
• Direct injection of diesel fuel into the exhaust, with exothermic reaction of the fuel
over the DOC and CDPF.
• Use of an exhaust-integrated, external combustion burner system.
Selective catalytic reduction (SCR) reduces nitrogen oxides NOx (consisting of both NO and
NO2) to N2 and water by using ammonia (NH3) as the reducing agent. The SCR catalyst coatings
used for post-2010 model-year heavy-duty diesel applications in the U.S. are typically copper
(Cu) exchanged or iron (Fe) exchanged zeolites, (e.g., Fe-ZSM-5), and most SCR coatings for
recent (MY2018-2020) applications are one of several different Cu and/or Fe-exchanged
chabazite zeolite structures (e.g., Cu-SSZ-13). The method for supplying ammonia to the SCR
catalyst is to inject a mixture of 32.5% urea in water solution into the exhaust stream. In the
presence of high temperature exhaust gasses (> 180 - 250 °C)C, the urea decomposes to form
both NH3 and iso-cyanic acid (HNCO) by thermolysis, with subsequent hydrolysis of the HNCO
to form additional NH3:
CO(NH2)2 —> NH3 + HNCO
HNCO + H20 —> NH3 + C02
The "standard SCR reaction" of NO (the predominant NOx species from diesel combustion)
over transition-metal zeolite or vanadium SCR catalysts can be represented as:
4NH3 + 4NO + 02 —> 4N2 + 6H2O
Improved reaction kinetics can be achieved at low exhaust temperatures (<300 °C) by
oxidizing a portion of the NO in exhaust using a platinum-group-metal (PGM) coated diesel
oxidation catalyst (DOC) to achieve a 1:1 molar ratio of NO:N02. The resulting "fast SCR
reaction" is:
B The temperature at which soot oxidation temperature occurs differs depending on the catalytic coating used on the
CDPF.
c Note that the urea decomposition temperature is dependent upon spray atomization, exhaust flow turbulence, and
exhaust flow rate.
9
-------
2NH3 + NO + NO2
2N2 + 3H20
An NO2 SCR reaction with limited NO availability also occurs, but at a significantly slower
reaction rate than the standard SCR reaction and is sometimes referred to as the "slow SCR
reaction":
8NH3 + 6 N02 —> 7N2 + 12H20
The details of SCR reactions for NOx reduction over transition-metal zeolite SCR can be
better represented as a series of interrelated reactions that are part of a more complex redox
cycle, such as the one proposed by Rudolf and Jacob.2
Urea dosing control takes into account a number of different factors, including:
• The stoichiometry of NOx reduction by NH3 (1:1 or 4/3:1 molar ratio)
• Molar ratio of N0:N02 at the inlet of the SCR catalyst
• The amount of NH3 stored and released from the zeolites
0 Thermal desorption of stored NH3 can allow NOx reduction to occur at exhaust
temperatures that are often too low for urea injection and decomposition
• The degree of urea/exhaust mixture preparation of the system design
0 Droplet formation and evaporation
0 Induced turbulent mixing of aqueous urea and exhaust to aid droplet breakup
• The efficiency of urea decomposition to NH3 (> 95-98% at >250 °C is typical for
modern injector/mixer designs)
• The probability forming solid deposits at low exhaust temperatures from partially
decomposed urea
0 Urea injection at exhaust temperatures below approximately 180 to 200 °C can
result in significantly increased deposit formation depending on mixture
preparation and other factors
0 Urea injector fouling can occur from deposit build up on the urea injector tip and
other exhaust system surfaces
0 Deposits can temporarily deactivate active catalytic surfaces, requiring higher
temperature operation in order to remove the deposits
Copper (Cu) exchanged chabazite zeolites such as Cu-SSZ-13 have demonstrated good
hydrothermal stability, good low temperature performance, and represent a large fraction of the
transition-metal zeolite SCR catalysts used in heavy-duty diesel applications.3 Improvements to
both the coating processes and the substrates onto which the zeolites are coated have improved
the low-temperature and high-temperature NOx conversion, improved selectivity of NOx
reduction to N2 (i.e., reduced selectivity to N2O), and improved the hydrothermal stability.
Improvements in SCR catalyst coatings over the past decade have included:4'5'6'7'8
10
-------
• Increased washcoat thickness
• Optimization of Silicon/Aluminum (Al) and Cu/Al ratios
• Increased Cu content and Cu surface area
• Optimization of the relative positioning of Cu2+ ions within the zeolite structure
• The introduction of specific co-cations
• Co-exchanging of more than one type of metal ion into the zeolite structure
In the absence of more stringent NOx standards, these improvements have been realized
primarily as reductions in SCR system volume, reductions in system cost, and improvements in
durability since the initial introduction of metal-exchanged zeolite SCR in MY2010. Sales-
weighted average engine-displacement-specific catalyst volumes for MY2019 MHDDE and
HHDDE are shown in Table 1-1.
Table 1-1: Engine-displacement specific catalyst substrate volume for MY2019 MHDDE and HHDDE
Component
MHDDE Specific Volume*
HHDDE Specific Volume*
DOC
0.61
0.74
CDPF
1.39
1.49
SCR
2.11
2.24
ASC
0.38
0.40
Notes:
*Specific Volume = (catalyst total substrate volume) / (engine's piston swept displacement)
NOx reductions greater than 95% are possible with modern SCR systems over a broad range
of operating conditions and at relatively high hours of operation, however SCR functionality is
particularly reduced at lower exhaust temperatures due to difficulties with low-temperature urea
decomposition and due to slower SCR reaction kinetics (Figure 1-3). In the figure, the initial
data point at 140 °C SCR inlet temperature reflects NOX reduction with stored ammonia only
(no urea injection). The hours in the legend represent hours of operation over an accelerated
aging cycle that included both thermal and chemical effects. The 1,430 hours of aging
represented approximately 8,000 hours of equivalent engine operation. Reduced oxidation of NO
to NO2 over the DOC and DPF at low exhaust temperatures (e.g., 200 to 250 °C) reduces the
ability to take advantage of the "fast SCR reaction". As previously mentioned, current SCR
systems limit urea injection to temperatures above 180 °C to 200 °C to prevent urea injector and
catalyst deposits. NOx reduction reactions at temperatures below approximately 200 °C are thus
reliant on use of NH3 stored within the zeolite structure. During extended operation at low
exhaust temperatures, stored NH3 is eventually depleted and if exhaust temperatures cannot be
increased sufficiently to allow initiation of urea injection and effective decomposition to NH3,
then NOx reduction eventually ceases.
11
-------
20%
10%
0%
DOC InletTemperaturefC) -> 180
SCR InletTemperaturefC) — 140
SCR Space Velocity (hr1) — 6.500
280
220
7.000
360
320
13.000
430
370
17.000
400
370
25.000
390
360
33.000
380
360
39.000
390
370
48.0C
Figure 1-3: NOx reduction efficiency of an early, developmental Cu-zeolite SCR formulation relative to DOC
and SCR inlet temperatures and SCR space velocity (adapted from McDonald et al. 20119).
Compression-ignition engine exhaust temperatures are relatively low following cold starts,
during coasting downhill, during sustained idle, or at low vehicle speeds and during light load
operation. Technologies that accelerate warm-up from a cold-start and maintain catalyst
temperature above 200°C can help achieve further NOx reduction from SCR systems under those
types of operation. Technologies that improve urea decomposition to NH3 at temperatures below
200°C can also be used to reduce NOx emissions under cold start, coasting, light load, and low
speed conditions.
1.1.2 Catalyst Durability
The regulatory full-useful-life for HHDDE emissions compliance with the fully-phased-in
2007 Heavy-duty Standards is 435,000 miles, ten years, or 22,000 hours of operation. Zeolite-
based SCR systems have demonstrated high levels of NOx reduction efficiency at the end of
regulatory useful life.D The aging mechanisms of diesel exhaust aftertreatment systems are
complex and include both chemical and hydrothermal changes. Aging mechanisms on a single
component can also cascade into impacts on multiple catalysts and catalytic reactions within the
system due to the interrelated nature of catalytic reactions over upstream components on other
aftertreatment components further downstream. Some aging impacts are fully reversible (i.e.,
conditions occur that can fully mitigate the aging impact). Other aging impacts are only partially
D Note that this would not necessarily be the case for EAS subjected to misfuelling with a high sulfur distillate diesel
fuel, poor maintenance and subsequent severe component failure, or tampering with or removal of key EAS system
components.
12
-------
reversible, irreversible, or can only be reversed with some form of intervention (e.g., changes to
engine calibration to alter exhaust temperature and/or composition).
DOCs undergo reversible aging due to adsorption of hydrocarbons, soot and sulfur species;
and irreversible aging due to phosphorus (P) poisoningE and thermal sinteringF of platinum
group metals (PGM) and other active materials. The catalytic materials on DPFs undergo similar
aging impacts, and also continuously accumulate metallic ash, which typically accumulates
towards the rear of the DPF channels. As accumulated ash migrates towards the front of the DPF
channels, exhaust backpressure increases. Heavy-duty diesel exhaust systems, particularly those
of HHDDE, are designed to allow CDPF removal for ash maintenance, often at approximately
250,000- to 350,000-mile intervals. Systems for LHDDE and MHDDE are typically designed
with sufficient CDPF capacity to forgo ash maintenance within the current regulatory
requirements for full useful life. Ash maintenance involves removal of the CDPF and
application of a either a dry cleaning process (e.g., reverse flushing of the CDPF with
compressed air) or a wet cleaning process (e.g., reverse flushing of the CDPF with water or with
a specific aqueous cleaning solution).
Aging of zeolite SCR is more complex. Hydrothermal aging of Cu-SSZ-13 SCR catalysts
impacts both catalyst acidity and NFb adsorption, transforms active Cu sites into less active
species, and causes Cu migration from exchanged positions within the zeolite structure and
subsequent formation of aggregated CuO.10 The severity of hydrothermal aging increases in the
presence of sulfur.11
Chemical poisoning of SCR can occur from fuel and lubricant contaminants, or via
degradation of upstream components. Sources of chemical poisoning include:
• Lubricant consumption
° Zinc dialkyldithiophosphate anti-wear, antioxidant, and corrosion inhibiting
additives
Phosphorus (P)
Zinc (Zn)
° Sulfur (S)
• Fuel
° Sulfur
° Trace contaminants from biodiesel (alkali metals, e.g., Na, K)
° Adsorption of hydrocarbon species into zeolite structure and subsequent blockage
of pores through soot formation12
• Migration of metals from upstream components
° PGM from the DOC and CDPF
° Transition metal (e.g., Fe, Cu) oxides from upstream SCR components or along
SCR substrates in series
E The sources of P poisoning are from lubricating oil consumption and P-containing lubricating oil additives, such as
zinc dithiophosphate.
F Sintering is a solid-solid phase transition that occur at very high temperatures and can lead to the transformation of
one crystalline phase into another. Phase transformations typically occur in the bulk washcoat, and they dramatically
decrease the surface area of the catalyst.
13
-------
Hydrothermal and chemical aging impacts on the DOC can also impact SCR NOx reduction,
particularly at low temperatures, via inhibition of NO to NO2 oxidation necessary for the fast
SCR reaction. The potential for future SCR durability improvements fall into the following
categories:
• Designing excess capacity into the catalyst (e.g., increased catalyst volume, increased
catalyst cell density, increased active material content and surface area)
• Use of a small-volume initial "sacrificial" substrate to adsorb chemical catalyst
poisons upstream of the initial DOC or SCR substrate
• Continued improvements to zeolite materials
0 Further optimization of Silicon/Aluminum (Al) and Cu/Al ratios
0 Exchange of beneficial co-cations
0 Co-exchanging of more than one type of transition metal into the zeolite structure
0 Reducing pore size to inhibit HC adsorption and pyrolysis
• Direct hydrocarbon dosing downstream of the light-off SCR during active CDPF
regeneration to reduce exposure of the light-off SCR to hydrocarbons and fuel
contaminants
• Use of washcoat additives, changes to substrate porosity, and other improvements to
increase PGM dispersion, reduce PGM particle size, reduce PGM mobility and
reduce agglomeration within the DOC and CDPF washcoatings,
• Improvements to catalyst housings and substrate matting material to minimize
vibration and prevent exhaust gas leakage around the substrate
• Reducing SCR and DOC sectional density, either through increased porosity or
decreased cell wall thickness, thus lowering substrate mass and improving warm-up
characteristics.
• Adjusting engine calibration and emissions control system design to minimize
operation that would damage the catalyst (e.g., improved control of CDPF active
regeneration, increased passive CDPF regeneration, HC dosing downstream of initial
light-off SCR, direct temperature sensor feedback control of active regeneration and
chemical deSOx)
• Use of specific engine calibration strategies for chemical deSOx of SCR (e.g., high
temperature operation with urea dosing)13 to remove strongly-bound sulfur
compounds from zeolite SCR
• Diagnosis and prevention of upstream engine malfunctions that can potentially
damage exhaust aftertreatment components
Increased SCR catalyst capacity, along with incremental improvements to current zeolite
coatings would be primary strategies for improving NOx control over a longer regulatory useful
life requirement. SCR capacity can be increased by approximately 40 to 50% with the use of a
light-off SCR substrate combined with a downstream substrate with a moderate volume increase
and with moderately increased catalytic activity from continued incremental improvements to
chabazite and other zeolite coatings used for SCR. Total SCR volume would thus increase by
approximately 50% to 80% relative to today's systems. SCR capacity can also be increased in
the downstream SCR system using thin-wall (4 to 4.5 mil), high cell density (600 cells-per-
square-inch) substrates.
14
-------
Chemical aging of the DOC, CDPF, and SCR can be reduced by the presence of an upstream
light-off SCR or use of a small "sacrificial" substrate to adsorb chemical poisons. Transport and
adsorption of sulfur (S), P, calcium (Ca), zinc (Zn), sodium (Na), and potassium (K) compounds
and other catalyst poisons are more severe for the initial catalyst within an emissions control
system and tend to reduce in severity for catalysts positioned further downstream. Chemical
deSOx strategies can be used to remove strongly-bound sulfur from zeolite SCR13. This involves
creating a strongly reducing environment via dosing of urea in excess of the typical 1:1 NH3 to
NOx ratio at temperatures of approximately 500 °C to 550 °C. Further evolutionary
improvements to the DOC washcoating materials to increase PGM dispersion and reduce PGM
mobility and agglomeration are also anticipated for meeting increased useful life requirements.
The primary strategy for maintaining CDPF function to a longer useful life is via design of
integrated systems that facilitate easier removal of the CDPF for ash cleaning at regular
maintenance intervals. Accommodation of CDPF removal for ash maintenance is already
incorporated into many existing diesel exhaust system designs.0 Incremental improvements to
catalyst housings and substrate matting material are also expected to be necessary for all catalyst
substrates within the system. Integration into a box-muffler type system is currently being used
by a number of manufacturers and this approach is expected to continue for all catalyst
components (except possibly for an initial close-coupled/light-off SCR) in order to improve
passive thermal management and improve access to the DPF for ash maintenance.
1.1.3 Improving SCR NOx Reduction at Low Exhaust Temperatures
The improvement of SCR NOx reduction under low-speed (<1200 rpm) light-load (< 5-bar
BMEP) conditions or immediately following cold starts will require improvements to both active
and passive thermal management of the EAS.
1.1.3.1 Active Thermal Management
Active thermal management involves using engine hardware and associated control systems
to maintain and/or increase exhaust temperatures. This can be accomplished through a variety of
means, including engine throttling, heated aftertreatment systems, and exhaust flow bypass
systems. Later combustion phasing can also be used for active thermal management.
Diesel engines operate at very low fuel-air ratios (i.e., with considerable excess air), and
particularly so at low load (<5-bar BMEP) conditions. This causes relatively cool exhaust to
flow through the exhaust system at low loads, which cools the catalyst substrates. This is
particularly the case at idle. It is also significant at moderate-to-high engine speeds with little or
no engine load, such as when a vehicle is coasting down a hill. Air flow through the engine can
be reduced by induction and/or exhaust throttling. All modern heavy-duty diesel engines are
equipped with an electronic throttle control (ETC) within the induction system and most are
equipped with a variable-geometry-turbine (VGT) turbocharger, and these systems can be used
to throttle the induction and exhaust system, respectively, at light-load conditions. However,
throttling reduces volumetric efficiency11, and thus has a trade-off relative to CO2 emissions and
fuel consumption.
G Video by Eberspaecher demonstrating DPF removal for ash cleaning maintenance: https://youtu.be/lf_vysKbfaA
H Relative efficiency of the air-exchange process in an internal combustion engine
15
-------
Heat can be added to the exhaust and the EAS by burning fuel in the exhaust system or by
using electrical heating, both of which can increase the SCR efficiency. Burner systems use an
additional diesel fuel injector in the exhaust to combust fuel and create additional heat energy in
the exhaust flow. Electrically heated catalysts (EHC) use electric current applied to a metal foil
monolithic structure in the exhaust to add heat to the exhaust flow. At light-load conditions with
relatively high flow/low temperature exhaust, considerable fuel energy or electric energy is
needed for these systems. This would likely cause a considerable increase in CO2 emissions and
fuel consumption with conventional designs. Heated and higher-pressure urea dosing systems
improve the decomposition of urea at low exhaust temperatures and thus allow urea injection to
occur at lower exhaust temperature (i.e., at approximately 135 °C to 140 °C) and with
considerably less energy-input when compared to burner systems or EHC.
Exhaust flow bypass systems can be used to manage the cooling of exhaust during cold start
and low load operating conditions. For example, significant heat loss occurs as the exhaust gases
flow through the turbocharger turbine. Turbine bypass valves allow exhaust gas to bypass the
turbine and avoid this heat loss at low loads when turbocharger boost requirements are low. In
addition, an EGR flow bypass valve would allow exhaust gases to bypass the EGR cooler when
EGR cooling is not required, such as immediately following a cold start or under cold ambient
conditions. EGR cooler bypass is currently used in light-duty diesel, Light HDE, and Medium
HDE applications.1415
Variable valve actuation (VVA) systems can also be used for active thermal management.
VVA includes a family of valvetrain designs that alter the timing and/or lift of the intake and
exhaust valves. Use of VVA can reduce pumping losses, increase specific power, and control
the level of residual gases in the cylinder.
VVA has been adopted in light-duty vehicles to increase an engine's efficiency and specific
power. It has also been used as a thermal management technology to open exhaust valves early
to increase heat rejection to the exhaust and heat up exhaust catalysts more quickly. This VVA
strategy, called early exhaust valve opening (EEVO), has been applied to the Detroit DD816 to
aid in CDPF regeneration, but a challenge with this strategy for maintaining aftertreatment
temperature is that it reduces cycle thermal efficiency, and thus can contribute to increased C02
emissions.
Cylinder deactivation (CDA), late intake valve closing (LIVC), and early intake valve closing
(EIVC) are three VVA strategies that can also be used to reduce airflow through the exhaust
system at light-load conditions and can reduce the CO2 emissions and fuel consumption trade-off
compared to use of the ETC and/or VGT for throttling.I7-18-19-20
Since we are particularly concerned with catalyst performance at low loads, EPA evaluated
two valvetrain-targeted thermal management strategies that reduce air aspiration of engines at
light-load conditions (i.e., less than 3-4 bar BMEP): CDA and LIVC. Both strategies force
engines to operate at a higher fuel-air ratio in the active cylinders for a given load demand, which
increases exhaust temperatures, with the benefit of little or no fuel consumption increase and
with potential for fuel consumption decreases under some operating conditions. The key
difference between these two strategies is that CDA completely removes airflow from one or
more deactivated cylinders with the potential for exhaust temperature increases of up to 80 °C at
light loads, while LIVC reduces airflow from all cylinders with up to 40 °C hotter exhaust
temperatures.18'19'20
16
-------
One of the challenges of CDA is that it requires proper integration with the rest of the
vehicle's driveline. This can be difficult in the vocational vehicle segment where an engine is
often sold by the engine manufacturer (to a chassis manufacturer or body builder) without
knowing the type of transmission or axle used in the vehicle or the precise duty cycle of the
vehicle. The use of CDA requires fine tuning of the engine calibration as the engine moves into
and out of deactivation to achieve acceptable noise, vibration, and harshness (NVH). Mitigation
strategies include changes to driveline dampening, motor mount and/or chassis dampening, and
the use of dynamic CDA with individual cylinder deactivation control. LIVC may provide
emission reductions similar to fixed CDA, with the added benefits of no significant NVH
concerns and some efficiency improvements under higher load conditions.
1.1.3.2 Passive Thermal Management
Passive thermal management involves changes or modifications to component designs to
increase and maintain the exhaust gas temperatures without the use of active thermal
management. It is done primarily through insulation and/or reducing the mass of EAS and other
exhaust system components so that less exhaust energy input is required to reach catalyst light-
off temperatures and/or the exhaust temperatures at which urea dosing can commence.21 Passive
thermal management strategies generally have little to no impact on CO2 emissions or can
improve CO2 emissions if used to replace an active thermal management strategy. The use of
passive thermal management strategies for improving catalyst light-off in light-duty gasoline
applications has led to significant reductions in cold-start exhaust emissions.22 Passive thermal
management design elements can be equally applied to EAS systems used in heavy-duty
applications.
More specifically, using a smaller sized, initial SCR catalyst within the EAS with a high-
porosity, lower density substrate reduces its mass and reduces catalyst warmup time. Moving the
SCR catalyst nearer to the turbocharger outlet effectively reduces the exhaust system mass and
surface area prior to the SCR inlet, minimizing heat loss and reducing the amount of energy
needed to warm components up to normal operating temperatures. Reducing the mass of the
exhaust system and insulating between the turbocharger outlet and the inlet of the SCR system
using an air-gap or other insulation can also reduce the amount of thermal energy lost through
the exhaust system walls. Close coupling of catalysts is near ubiquitous in modern light-duty
EAS. The use of air-gap construction is also common in light-duty applications.
Dual-walled manifolds and exhaust pipes utilizing a thin inner wall and an air gap separating
the inner and outer wall can be used to simultaneously insulate the exhaust system and reduce the
thermal mass, minimizing heat lost to the walls and decreasing the time necessary to reach
operational temperatures after a cold start. Mechanical insulation applied to the exterior of
exhaust components, including exhaust catalysts, is readily available and can minimize heat loss
to the environment and help retain heat within the catalyst as operation transitions to lighter
loads and lower exhaust temperatures. Integrating the DOC, DPF, and SCR substrates into a
single exhaust assembly can also assist with retaining heat energy.
EPA evaluated several passive thermal management strategies in the diesel technology
feasibility demonstration program. See RIA Chapter 3.1 for detailed discussion of our diesel
technology demonstration programs).
17
-------
1.1.3.3 Advanced SCR System Development
A recent development in SCR system architecture is the development of light-off or dual SCR
systems, which is a variation of passive thermal management.18-23- 24 This system maintains a
layout similar to the conventional SCR configuration discussed earlier, but integrates an
additional small-volume SCR catalyst, which is in some cases also close-coupled to the
turbocharger's exhaust turbine outlet (Figure 1-4). This small SCR catalyst may be configured
with or without an upstream DOC, and with or without a small sacrificial substrate to adsorb
chemical poisons upstream of an initial SCR substrate. A recent example of this system's
architecture was demonstrated as part of "Stage 3" of the California Air Resources Board
(CARB) - Heavy-duty Low-NOx Test Program.25 The CARB Stage 3 research program is
summarized within Chapter 3.1. EPA evaluated dual-SCR catalyst system technology similar to
the CARB "Stage 3" system as part of a diesel technology feasibility demonstration program (see
Chapter 3.1 for more detail).
The benefits of this design result from its ability to warm up the initial light-off SCR substrate
faster as a result of it being relatively low mass and being the first catalyst downstream of the
turbocharger with the EAS. Such light-off SCR catalysts can also be designed to have smaller
substrates with lower bulk density. The reduced mass reduces thermal inertia and allows faster
warmup. The design also positions the urea injection and mixing as the very first components in
the system, thus allowing faster heat up of the urea injector and urea mixer when implementing
active thermal management measures. These designs also require less input of heat energy into
the exhaust to maintain exhaust temperatures during light-load operation. Urea injection to the
close-coupled light-off SCR can also be reduced or terminated once the second, downstream
SCR reaches operational temperature, thus allowing additional NOx to reach the DOC and CDPF
to promote passive regeneration (soot oxidation) on the CDPF, reducing fuel consumption and
CO2 emissions. Very close-coupling of the light-off SCR to the exhaust turbine is possible when
using heated urea dosing system since such systems enable a relatively short mixing length
between the urea dosing system and the inlet of the light-off SCR (see Chapter 3.1 for more
detail).
Urea Inieitur Cu-SSZ-13 SCR + ASC
CU-5SZ-13 SCR
18
-------
Urea Injector
Cu-SSZ-13 SCR Cu-SSZ-13 SCR + ASC
(zone coated)
Figure 1-4: Potential layout of a 2027+ dual-SCR system in an in-line configuration (top) and comparable
components integrated to improve passive thermal management (bottom).
One potential concern about this technology is the durability challenge associated with
placing an SCR catalyst upstream of the CDPF. To address this concern, two light-off SCR
system designs were hydrothermally and chemically aged to an equivalent of 850,000 miles as
part of the EPA Heavy-duty Diesel Low NOx Demonstration Program. Please refer to Chapter
3.1 for additional information regarding this test program.
1.1.4 Crankcases
During combustion, gases can leak past the piston rings sealing the cylinder and into the
crankcase. These gases are called blowby gases and generally include unburned fuel and other
combustion products. Blowby gases that escape from the crankcase are considered crankcase
emissions (see 40 CFR 86.402-78). Current regulations restrict the discharge of crankcase
emissions directly into the ambient air, and blowby gases from gasoline engine crankcases have
been controlled for many years by sealing the crankcase and routing the gases into the intake air
through a positive crankcase ventilation (PCV) valve. However, in the past there have been
concerns about applying a similar technology for diesel engines. For example, high PM
emissions venting into the intake system could foul turbocharger compressors. As a result of this
concern, diesel-fueled and other compression-ignition engines equipped with turbochargers (or
other equipment) were not required to have sealed crankcases (see 40 CFR 86.007-11(c)). For
these engines, manufacturers are allowed to vent the crankcase emissions to ambient air as long
as they are measured and added to the exhaust emissions during all emission testing to ensure
compliance against the emission standards.
Because all new highway heavy-duty diesel engines on the market today are equipped with
turbochargers, they are not required to have closed crankcases under the current regulations.
Manufacturer compliance data show approximately one-third of current highway heavy-duty
diesel engines have closed crankcases, indicating that some heavy-duty engine manufacturers
have developed systems for controlling crankcase emissions that do not negatively impact the
turbocharger. In the final rule associated with this RIA, EPA is finalizing a requirement for
manufacturers to use one of two options for controlling crankcase emissions. One option is
closing the crankcase, as proposed. These emissions could be routed upstream of the
aftertreatment system or back into the intake system. The second option is an updated version of
19
-------
the current requirements for an open crankcase that includes accounting for total emissions
during certification and off-cycle field testing through useful life including full accounting of
crankcase emission deterioration (See Preamble Section III.B).
1.1.4.1 Emissions from Open Crankcases
EPA conducted emissions testing of open crankcase systems on two low mileage, modern
heavy-duty diesel trucks.26 The testing was conducted at EPA's National Vehicle and Fuel
Emissions Laboratory. The two vehicles were tested on a heavy-duty chassis dynamometer
where the crankcase flow and emissions were measured separately from the tailpipe exhaust
emissions. The vehicles were tested over a variety of operating conditions. The cycles included
the ARB Transient cycle with a cold start, repeat ARB Transient cycles, a 10-minute idle cycle,
and a highway cycle at 55 mph and 65 mph.
The crankcase emission rates were calculated for THC, NMHC, CH4, NOx, CO2, and CO
using the densities found in 40 CFR 86.144-94. The average crankcase and tailpipe emission
rates for each of the two trucks (NVFEL 1 and NVFEL 2) by test phase are show in Figure 1-5.
The error bars represent the standard error of the mean. As shown, the crankcase THC and CO
emissions are a notable fraction of the tailpipe exhaust emissions. Table 1-2 includes the average
crankcase emission rates across the cycles for each truck.
Table 1-2: Average Crankcase Emission Rates (gram/hour)
THC
CH4
NOX
CO
Truck 1
0.305
0.001
1.19
0.212
Truck 2
0.067
0.026
1.09
0.832
20
-------
7
b
5, s
t2:
¦ Tailpipe CO (a/hr)
¦ O anfcca«* CO Is/*1''*
I* 1.
Tr*n*i«nt
iHf tUrt
TfinHoH
NVFEL 1
¦ Trill pip#* THC
ll
it
Hip hw *v wf (,
NVFEl 2
tdlB Nt**TW«V
NVFEL 2
I I .
Tr»n»l»ni »r.§«i>icTV u
NVFEL 2
1AO
130
ICO
30
%
CO
40
?o
1
HV«mlcry i«llr
NVFFI 1
¦ Taalpip*- NO* I*/)1")
«¦ Crankcase NQx (g/Hr)
I
Iransicn-t N
NVFEL 1
¦ i
Transient f
NVFFI 7
NVFEL 2
Figure 1-5: Tailpipe Exhaust and Crankcase Emission Rates from Two Heavy-Duty Diesel
Trucks
We were unable to measure PM emissions from the crankcase as part of the EPA test
program. Therefore, in our MOVES model we will continue to use the PM emission rates
21
-------
measured in the Advanced Collaborative Emissions Study (ACES) Phase 1 test program.27 The
average PM emission rate of the four 2007 MY heavy-duty diesel engines was 32.1 mg/hour.
1.1.4.2 Description of Closed Crankcase Technologies
If manufacturers choose to close the crankcase, crankcase emissions can be controlled through
the use of closed crankcase filtration systems or by routing unfiltered blow-by gases directly into
the exhaust system upstream of the emission control equipment. Closed crankcase filtration
systems work by separating oil and particulate matter from the blow-by gases through single or
dual stage filtration approaches, routing the blowby gases into the engine's intake manifold and
returning the filtered oil to the oil sump. These systems were required for new heavy-duty diesel
vehicles in Europe starting in 2000. Oil separation efficiencies in excess of 90 percent have been
demonstrated with production ready prototypes of two stage filtration systems. By eliminating 90
percent of the oil that would normally be vented to the atmosphere, the system works to reduce
oil consumption and to eliminate concerns over fouling of the intake system when the gases are
routed through the turbocharger.
An alternative approach would be to route the blow-by gases into the exhaust system
upstream of the catalyzed diesel particulate filter which would be expected to effectively trap and
oxidize the engine oil and diesel PM. This approach may require the use of low sulfur engine oil
to ensure that oil carried in the blow-by gases does not compromise the performance of the
sulfur-sensitive emission control equipment.
Our feasibility analysis is based on the use of closed crankcase system that includes
technologies designed to filter crankcase gases sending the clean gas to the engine intake for
combustion and returning the oil filtered from the gases to the engine crankcase. These systems
are proven in use.
1,1.5 Opposed-Piston Diesel Engines
While not part of EPA's planned technology demonstration program for this rulemaking, the
agency is tracking ongoing work to develop opposed-piston diesel engine technology for heavy-
duty on-highway vehicle applications.28'29 One example of work with this technology is a project
to develop and demonstrate a 10.6 liter, 450 hp opposed-piston diesel engine and related
aftertreatment technologies for Class 8 line-haul tractors operating at certified emission
performance levels of 0.02 g NOx/bhp-h (90 percent below US 2010) and 432 g C02/bhp-h
(2027 HD Phase 2 engine standard).30 In addition to the emissions demonstration work, a high
level cost study has been conducted by FEV that indicates the direct and indirect cost of an
opposed-piston engine are less than that of a conventional HD diesel engine.31 This project is
supported by a variety of public and private sector partners including: Achates Power, Aramco
Services Company, BASF, CALSTART, CARB, Corning, Delphi, Eaton, Faurecia, Federal
Mogul, PACCAR/Peterbilt, Sacramento Metropolitan Air Quality Management District, San
Joaquin Valley Air Pollution Control District, South Coast Air Quality Management District,
Southwest Research Institute, Tyson Foods, and Walmart.32
Opposed-piston engine technology has not yet been proven feasible in Class 8 on-highway
applications, but if it becomes feasible, then the technology could provide another pathway to
ultra-low NOx, high efficiency engine technology for heavy-duty vehicle fleets. As such, it may
22
-------
be reasonable to anticipate commercialization of heavy-duty opposed-piston diesel engine
technology by model year 2027.
1.2 Spark-Ignition Engine Technologies
The following sections describe the spark-ignition (SI) engine technologies we considered to
reduce criteria pollutant emissions (NMHC, CO, NOx, and PM) as part of this rulemaking. Many
of the technologies are described with respect to gasoline fuel, but they are expected to be
broadly applicable to all fuels used in spark-ignition engines. Our spark-ignition engine
feasibility demonstration for this final rule is based on some of the technologies presented in this
section.
Chapter 2.3 of this RIA describes the test procedures for spark-ignition engine certification.
Chapter 3.2 describes the spark-ignition engine feasibility demonstration program, including a
description of the specific technology packages we evaluated, the effectiveness of those
technologies over our current and new test procedures, and our projected direct manufacturing
cost of those technologies.
1.2.1 Technology Description for NMHC, CO, and NOx Control
A range of technology options exist to reduce NMHC, CO, and NOx emissions from both
heavy-duty highway gasoline fueled spark ignition and diesel engines to levels below the current
EPA 2007/2010 standards. Available options include modifications to the engine calibration,
engine design, exhaust system, and aftertreatment system design. The different available options
each contain 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 spark-
ignition engines covered by this rule. Unique compression-ignition technologies are addressed in
Section 1.1.
1.2.1.1 Summary of the Technology Challenge for NMHC, CO, and NOx Control
Historically, heavy-duty spark-ignition engines running the EPA Federal Test Procedure
(FTP) tests have shown that the majority of NMHC, CO, and NOx emissions occur during the
cold start phase. Real world emissions during warmed-up and hot operation, specifically during
high-load operation that may not be fully exercised by the FTP tests, can significantly contribute
to emissions. Additionally, as described in Chapter 3.2, in-vehicle testing has indicated that
sustained low load conditions such as prolonged idling can result in emission increases due to
reduced aftertreatment temperatures (i.e., cool-off). Prolonged idling is also a real world
condition that is not thoroughly addressed by the FTP. For this rulemaking, we evaluated
strategies that target prolonged idle and FTP and high-load NMHC, CO and NOx emission
control performance. Specifically, significant quantities of NMHC and CO emissions can be
produced if enrichment events occur regularly during high-load operation. Control of NOx
emissions during high-load operation requires designs that provide sufficient catalyst volume to
handle the higher exhaust gas flow rates and also precise control of closed loop fuel biasing for
the catalyst to maintain peak NOx efficiency.
23
-------
In order to achieve significantly lower NMHC, CO and NOx emissions over the FTP,
manufacturers can change the design of their exhaust and catalyst systems, as well as adopt
calibration strategies to reduce catalyst light-off times and reduce warmed-up and hot running
emissions. Design changes to reduce catalyst light-off time (e.g., closer catalyst placement) can
also result in higher catalyst temperatures during high-load operation. To achieve lower NMHC
and NOx levels, manufacturers will need to develop and implement technologies and calibration
strategies to manage catalyst temperatures during high-load operation while minimizing fuel
enrichment.
For the catalyst to effectively reduce NMHC, CO and NOx emissions it must reach a light-off
temperature of approximately 350 °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 and maintaining sufficient catalyst temperature during low load and idle
operation. Reducing warmed-up NOx emissions requires improving the efficiency of the catalyst
system using improved catalyst loading and washcoat technologies in addition to more precise
calibration and software controls. NOx emissions control system effectiveness has generally
improved from a reduction in the sulfur content of the fuel in recent years.1
It is anticipated that engine manufacturers will focus on four 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
staying above the light-off temperature throughout all operation
• Improving the NOx efficiency of the catalyst during warmed-up operation at medium
and high loads
• Minimizing or eliminating enrichment in high-load operation.
We describe strategies to address these four areas in the following sections.
1.2.1.2 Reducing Engine-Out Emissions
When the three-way catalyst (TWC) of an SI engine is "cold" (i.e., the engine temperature
matches ambient and is between 20 and 30 °C (68 and 86 °F)) and the engine is operating under
low loads, the cold engine produces lower concentrations of NOx than NMHC. As the engine
warms up and as the load increases, the concentration of NOx produced by the engine increases
and the concentration of NMHC decreases. We have observed this NMHC-NOx tradeoff during
the cold start portion of the FTP duty cycle as the engine transitions from the first minutes of
operation when the engine is either at idle or low speed and load to higher load with a warmed
engine.
1 Sulfur content in commercially available fuel was reduced starting in 2017 as part of the Tier 3 Motor Vehicle
Emission and Fuel Standards rule (79 FR 23414); this test program was conducted using gasoline with Tier 3 sulfur
content.
24
-------
The design of the air induction system, combustion chamber, spark plug, and fuel injection
system determine 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 NMHC 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 NMHC emissions.
For example, it has been shown that cold start NMHC 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.33 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 the creation of a combustible
mixture. Droplet size can be reduced by modifying the injector orifice plate and by increasing the
fuel pressure. Reducing droplet size to improve 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
light-duty FTP emission test.
The mixture formation process in a GDI engine is different than a PFI engine. In a PFI engine
the fuel can be injected prior to or during the intake valve opening to prepare the fuel in an
optimal manner for emission controls. The fuel generally has time to evaporate during the intake
stroke as the fuel and air are drawn into the cylinder and is mixed with the incoming air. In
addition, as the engine combustion heat from the previous firing events warms the intake valve
and other surfaces in the area, the fuel can be injected into the intake runner and engine heat can
assist in evaporating the fuel prior to the intake valve opening. The GDI engine injects fuel at
higher fuel pressures than PFI engines directly into the combustion chamber. In a GDI 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 GDI engine compared to a PFI engine. An advantage
of the GDI design is that the fuel spray does not impinge on the walls of the intake manifold or
other surfaces in the cylinder.
GDI systems stagger 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 GDI 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.34
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 NMHC emissions significantly during the engine warm up prior to the catalyst
reaching the light-off temperature.
25
-------
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 in coming model
years, which will pave the way for them to be applied to heavy-duty engines.
Because the engine is relatively cold and the operating loads are low during the first 50
seconds of operation, the engines typically do not produce significant quantities of NOx
emissions during this time. 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 NMHC 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.
1.2.1.3 Reducing Catalyst Light-Off Time
The effectiveness of current engine emission 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 350°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. The technologies are described in greater detail below.
1.2.1.3.1 Calibration Changes
Engine calibration changes may be employed to increase the temperature and mass flow of
the exhaust with the goal of reducing the amount of time required for the catalyst to reach the
critical light-off temperature. By reducing the time required for the catalyst to light-off, these
changes can effectively reduce NMHC, CO and NOx emissions. Since the catalyst system in an
SI engine is the predominant method to control emissions and is responsible for over a 95%
reduction from the engine out emissions, any acceleration in the warm-up of the catalyst system
translates into immediate emission reductions at the tailpipe.
Retarding combustion in a cold engine by retarding the spark advance is a well-known
method for reducing the concentration of NMHC emissions in the exhaust and increasing the
exhaust gas temperature.35'36 The reduction in NMHC concentrations is due to a large fraction of
the unburned fuel within the cylinder combusting before the flame is extinguished at the cylinder
wall. Reductions of total hydrocarbon mass of up to 40 percent have been reported from these
studies evaluating the effect of spark retard on exhaust emissions.
In addition to reducing the NMHC exhaust concentration, 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
26
-------
NMHC 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 NMHC 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 NMHC 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 NMHC emissions for all vehicle classes.
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. Calibration changes to the valve timing during the cold start phase can lead to
additional reductions in cold start NMHC emissions.37
1,2,13.2 Exhaust System Thermal Management
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 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 achieve catalyst light-off, thermal management
technologies reduce NMHC, CO and NOx emissions.
Moving the catalyst closer to the cylinder head is a means that 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 closer to the cylinder head reduces the light-off
time significantly.
Another means for reducing heat losses is 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.
27
-------
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
manufacturers will ensure that fuel enrichment is minimized on the FTP. Calibration measures,
other than fuel enrichment, may therefore be needed 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 a means to both quickly light-off the
catalyst and reduce high-load exhaust temperatures.
We expect thermal management to be an effective strategy for manufacturers to lower
NMHC, CO and NOx emission levels. Our feasibility demonstration described in Chapter 3.2
evaluates catalysts located closer to the engine as a method of thermal management. We expect
that manufacturers will further optimize the thermal inertia of the exhaust system to minimize the
time needed for the catalyst to achieve the light-off temperature, while ensuring the high-load
performance does not cause thermal degradation of the catalyst 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 for improving engine cold start
emission performance. It is not anticipated that heavy-duty vehicles with spark-ignition engines
will utilize catalysts that are very close-coupled to the exhaust manifold (i.e., will not use a
strategy similar to close-coupled catalyst locations found on passenger cars and light-duty
trucks). The higher operating loads of these heavy-duty 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.
1.2.1.3.3 Catalyst Design Changes
There are several different catalyst design changes that 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 1,200 cells per square inch (cpsi) with wall thicknesses approaching 0.05 mm.
28
-------
Increasing the surface area of the catalyst improves the performance of the catalyst. Higher
substrate cell densities increase the surface area for a given catalyst volume. Higher surface areas
improve the catalyst efficiency and durability reducing NMHC and NOx emissions.
The key limitation of the higher cell density substrates is 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.
Engine speeds and load are low during the first 50 seconds of the FTP test and it is a
challenge to achieve catalyst light-off during the cold start operation. One method for reducing
the catalyst light-off time is to replace a single catalyst with two catalysts which when combined
total the same volume as the single catalyst. Having a two-catalyst system that includes a close-
coupled, front catalyst with comparatively reduced-volume reduces the heat needed to reach the
light-off temperature for the front catalyst due to its location and reduced thermal mass. The
front catalyst of the two-catalyst system will reach operating temperature before the individually
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 reactions.
Therefore, as the loading levels and composition of the PGM changes, the light-off performance
for both NMHC and NOx need to be evaluated. Improved catalyst substrates and PGM loadings
designs are additional effective approaches to reduce emissions and we anticipate manufacturers
will incorporate advanced catalyst designs in their future emission strategies.38 We used an
advanced catalyst formulation in our HD SI feasibility demonstration for this rule (see Chapter
3.2).
1,2.1.3.4 Secondary Air Injection
By injecting air directly into the exhaust stream, close to the exhaust valve, combustion
(hydrocarbon oxidation) can be maintained within the exhaust, creating additional heat and
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 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. This fuel is added to decrease the air/fuel ratio in the cylinder to a ratio
that is rich of stoichiometric. 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.39'40
Engines which do not use secondary air injection can only operate rich of stoichiometry for a
minimal amount of time after a cold start as the added enrichment will cause increased NMHC
emissions. The richer cold start calibration used with vehicles that have a secondary air injection
system provides a benefit, as combustion 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
29
-------
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 will 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.
Historically, secondary air injection has also been used to control CO and NMHC emissions
during high load rich operation. These designs incorporated a mechanically driven air pump that
continuously injects air into part of the catalyst to oxidize the CO and NMHC emissions, a
particularly important emission control technology used during enrichment operation. With the
improvements in electronic engine controls, all manufacturers discontinued the use of secondary
air injection for the purpose of high load enrichment and instead used software models and other
algorithms that maintain stoichiometric operation for slightly longer periods of operation than
previously. These newer strategies however were designed to only provided a temporary
emission control benefit of stoichiometric operation since the engines eventually go into
enrichment modes for either power improvement or thermal protection at which point the CO
and NMHC emissions are no longer controlled.
1.2.1.4 Improving Catalyst NOx Efficiency during Fully Warmed-up Operation
For engines certified to the EPA 2007/2010 emission standards, significant quantities of NOx
emissions are produced by engines during warmed-up engine operation on the FTP. The
stabilized NOx emission levels will need to be reduced. Improving the NOx performance of the
engine 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 engines 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.
1.2.1.5 Reducing Enrichment
Heavy-duty vehicles tend to operate at high loads and catalyst durability can be 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 could result in increasing the time spent in fuel
enrichment modes to ensure the catalyst temperatures are maintained below design thresholds,
which if allowed to operate too hot could reduce the durability of the catalyst. Using fuel
enrichment to control catalyst temperature while effective, causes significant increases in criteria
pollutant emissions and also significant increases in fuel consumption.
1.2,1.5.1 Exhaust Gas Temperature Measurement
The methodology for determining when temperatures in the exhaust and in the catalyst are
high enough to initiate thermal protection (i.e., enrichment) is almost exclusively done using
software modeling of the thermally-limited components. This methodology can be effective at
triggering enrichment when needed; however, if it is implemented in an excessively conservative
manner where the temperature prediction is higher than the actual critical component
30
-------
temperature, it can result in unnecessary enrichment episodes which lead to substantial increases
in all the emissions. Since the gasoline heavy-duty engines are designed for work and are
expected to operate regularly at high loads where the exhaust temperatures become important
concerns for component durability, any improvement in the accuracy of the methodology to
provide enrichment protection will result in both reductions in emissions and improvements to
durability.
A potential improvement over the current practice of using temperature modeling is use of
hardware in the form of temperature measurement sensors in the exhaust at the most critical
locations, generally at the catalyst inlet. While some light-duty variations are available,
currently, only one gasoline HD product, the Nissan NV series of cargo vans with the 5.6L V8
has implemented temperature sensors.41 However, temperature measurement sensors are very
common in all diesel applications. While this improvement to the accuracy of temperature
measurements in the exhaust may not result in emission reductions during the limited operation
range of today's FTP and even SET certification test cycles, it will likely provide "real world"
emission reductions compared to modeled temperature strategies that may conservatively trigger
enrichment episodes prematurely.
1.2.1,5.2 Continuous Stoichiometric Operation
One concept that has been used in other non-road sectors (e.g., large SI engines operating
indoors or confined areas) is requiring the engine to always operate at stoichiometry where the
three-way catalyst is generally at peak efficiency for all emissions. To apply this strategy to
heavy-duty highway SI engines, manufacturers must prevent the engine from entering
enrichment. Enrichment can be avoided with upgrades to materials of specific components that
are currently limited by high temperature constraints, or using the large degree of modern engine
control authority to prevent the engine from entering areas of operation that require enrichment.
Modern gasoline engines have several engine hardware components and calibration strategies
that can reduce temperatures by modifying combustion or load characteristics, such as EGR,
valve timing, electronic throttle airflow, cylinder deactivation, and other available methods.
Some of these methods to remain in stoichiometric air-fuel control may result in a governing or
detuning of the engine after a period of time to avoid prolonged high-power operation where
stochiometric operation cannot be maintained due to increasing exhaust component
temperatures; however, this may be an acceptable approach for a sector where sustained absolute
engine power may not be necessary or as important as lower emissions and better fuel economy.
While we did not have the hardware and/or software required to directly control enrichment in
our HD SI feasibility demonstration, in Chapter 3.2 we show that emissions can be well-
controlled when an engine maintains stoichiometric operation. We expect engine manufacturers
will continue to optimize their engine calibrations and limit enrichment as an effective means of
reducing NMHC and CO.
1.2.1.6 Additional Emission Control Strategies
A strategy that may provide some degree of emission reductions involves down-speeding or
governing of the engine operating range to keep the engine speeds and loads in areas where
engine hardware and exhaust temperatures minimize needed enrichment for thermal protection.
This strategy will allow the emission controls to remain in stoichiometric air-fuel control (i.e.,
31
-------
closed loop) where the catalysts can maintain peak efficiency for NMHC, CO and NOx for a
broader range of operation.
The down-speeding approach was a technology discussed to reduce CO2 emissions for the HD
Phase 2 GHG rule42 and has been successfully implemented in both gasoline and diesel
applications to reduce GHG emissions. With the advent of modern electronic controls of both
engine and transmission, the opportunity exists to precisely keep the engine in the optimal
operating areas for reduced GHG and criteria emissions. Multi-speed automatic transmissions
available in recent years in HD applications with wide ratios containing 6 or more forward gears
have provided the opportunity to operate the engine in a more optimal fashion with little or no
loss of vehicle performance and capabilities. This strategy is currently used by at least one HD
gasoline engine manufacturer as indicated by its advertised maximum rated test speed (i.e., peak
horsepower) of approximately 4000 RPM compared to the much higher speeds, over 4700 RPM
of other HD applications.43 This lower speed is made possible by transmission strategies
preventing over-speeding, which allows the emission controls to operate in a much more
desirable and lower emitting area of engine operation. We evaluated engine down-speeding as
part of the HD SI feasibility demonstration presented in Chapter 3.2.
1,2.2 Technology Description for PM Control
Particulate matter emitted from internal combustion engines is a multi-component mixture
composed of elemental carbon (or soot), semi-volatile organic compounds (SVOC), sulfate
compounds (primarily sulfuric acid) with associated water, nitrate compounds and trace
quantities of metallic ash. At temperatures above 1,300 K, fuel hydrocarbons without access to
oxidants can pyrolyze to form particles of elemental carbon. Fuel pyrolysis can occur as the
result of operation at richer than stoichiometric air-to-fuel ratio (primarily gasoline PFI engines
and 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.
SVOCs are composed primarily of organic compounds from lubricant and partial combustion
products from fuel. PM emissions from SVOC are typically gas phase when emitted 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 NMHC
emissions. Sulfur and nitrogen compounds are emitted primarily as gaseous species (SO2, 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 SO2 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 SO2 over PGM catalysts.
32
-------
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)
• Injection strategies that eliminate opportunity for PM forming conditions (open valve
injection on PFI)
• 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 (GPF)
SVOC PM emissions can be controlled by:
• Reducing lubricating oil consumption
• Improvements in exhaust catalyst systems used to control gaseous NMHC emissions
(e.g., increased PGM surface area in the catalyst, improvements in achieving catalyst
light-off following cold-starts, etc.)
1.2.3 Technologies to Address Evaporative and Refueling Emissions
As exhaust emissions from gasoline engines continue to decrease, evaporative emissions
become an increasingly significant contribution to overall HC emissions from gasoline-fueled
vehicles. To evaluate the evaporative emission performance of current production heavy-duty
gasoline vehicles, EPA tested two heavy-duty vehicles over running loss, hot soak, three-day
diurnal, on-board refueling vapor recovery (ORVR), and static test procedures. These engine-
certified "incomplete" vehicles meet the current heavy-duty evaporative running loss, hot soak,
and three-day diurnal emission requirements. However, as they are certified as incomplete
vehicles, they are not required to control refueling emissions and do not have ORVR systems.
Results from the refueling testing confirm that these vehicles have much higher refueling
emissions than gasoline vehicles with ORVR controls.44'45 The results for the ORVR tests are
shown in Table 1-3. A discussion on the test procedure limitations and estimated modeled results
from this test program is in Section 2.3.2.
Table 1-3: ORVR results with modeled values for test procedure limitations
SHED (grams)
Modeled
SHED
(grams)
Average (g/gal)
Modeled
Average
(g/gal)
Current
Refueling
Standard
(g/gal)
Ford E-450
114
168
2.3
3.34
0.2
108
2.2
Isuzu NPR
72
86
2.8
71
2.8
33
-------
Opportunity exists to extend the usage of the refueling evaporative emission control
technologies already implemented in complete heavy-duty gasoline vehicles to the engine-
certified incomplete gasoline vehicles in the over-14,000 lb. GVWR category. The primary
technology we are considering is the addition of ORVR, which was first introduced to the
chassis-certified light-duty and heavy-duty applications beginning in MY 2000 (65 FR 6698,
February 10, 2000). An ORVR system includes a carbon canister, which is an effective
technology designed to capture HC emissions during refueling events when liquid gasoline
displaces HC vapors present in the vehicle's fuel tank as the tank is filled. Instead of releasing
the HC vapors into the ambient air, ORVR systems recover these HC vapors and store them for
later use as fuel to operate the engine.
The fuel systems on these over-14,000 pound GVWR incomplete heavy-duty gasoline
vehicles are similar to complete heavy-duty vehicles that are already required to incorporate
ORVR. These incomplete vehicles may have slightly larger fuel tanks than most chassis-
certified (complete) heavy-duty gasoline vehicles and are somewhat more likely to have dual
fuel tanks. These differences may require a greater ORVR system storage capacity and possibly
some unique accommodations for dual tanks (e.g., separate fuel filler locations), but we expect
they will maintain a similar design. Figure 1-6 presents a schematic of a standard ORVR system.
34
-------
Figure 1-6: Schematic of an ORVR system1
1.2.3.1 Filler Pipe and Seal
In an ORVR system, the design of the filler pipe, the section of line connecting the point at
which the fuel nozzle introduces fuel into the system to the gas tank, is integral to how fuel
vapors displaced during a fuel fill will be handled. The filler pipe is typically sized to handle the
maximum fill rate of liquid fuel allowed by law while also integrating one of two methods to
prevent fuel vapors from exiting through the filler pipe to the atmosphere: a mechanical seal or a
liquid seal approach. A dual fuel tank chassis configuration may require a separate filler pipe and
seal for each fuel tank
The mechanical seal is typically located at the top of the filler neck at the location where the
fuel nozzle is inserted into fuel neck. The hardware piece forms a seal against the fuel nozzle by
using some form of a flexible material (usually a plastic material) that makes direct contact with
the fuel station fuel-filling nozzle to prevent fuel vapors from exiting the filler pipe as liquid fuel
is pumped into the fuel tank. In the case of capless systems, this seal may be integrated into the
spring-loaded seal door that opens when the nozzle is inserted into the filler pipe receptacle.
There are concerns with a mechanical seal's durability due to wear over time, and its ability to
maintain a proper seal with unknown service station fill nozzle integrity and variations beyond
design tolerances.
The liquid seal approach uses the size and bends of the filler pipe to cause a condition where
the entire cross-section of the filler pipe is located in the fuel tank or close to the entry into the
fuel tank and is full of the incoming liquid fuel preventing fuel vapors from escaping up and out
1 Stant ORVR System http://stant.com/orvr/orvr-systems/
35
-------
through the filler pipe. By creating a solid column of liquid fuel in the filler pipe, the liquid seal
approach does not require a mechanical contact point with the fill nozzle to prevent escape of
vapors. The liquid seal has been the predominant sealing method implemented in the regulated
fleet in response to the ORVR requirements.
1.2.3.2 ORVR Flow Control Valve
As described above, the sealing of the filler pipe prevents the fuel vapors from escaping into
the ambient air; however, the fuel vapors that are displaced by the incoming liquid fuel need to
be routed to the canister. In order to properly manage the large volume of vapors during
refueling that need to be controlled, most ORVR systems have implemented a flow control valve
that senses that the fuel tank is getting filled with fuel and triggers a unique low-restriction flow
path to the canister. This flow path is specifically used only during the refueling operation and is
unique in that it provides the ability to quickly move larger volumes of fuel vapors into the tank
than normally required under other operation outside of refueling events. The flow control valve
will allow this larger flow volume path while refueling but then return to a more restrictive vapor
flow path under all other conditions, including while driving and while parked for overnight
diurnals.
The flow control valve is generally a fully-mechanical valve system that utilizes connections
to the fuel tank and filler pipe to open and close vapor pathways with check valves and check
balls and pressure switches via diaphragms. The valve may be integrated into the fuel tank and
incorporate other aspects of the fuel handling system ("multi-function control valve" in Figure
1-6) including roll-over valve, fuel and vapor separators to prevent liquid fuel from reaching the
canister, and other fuel tank vapor control hardware. Depending on the design, the filler pipe
may also be integrated with the flow control valve to provide the necessary pressure signals. A
dual fuel tank chassis configuration may require a separate flow control valve for each fuel tank.
1.2.3.3 Canister
The proven technology to capture and store fuel vapors has been activated charcoal. This
technology has been used in vehicles for over 50 years to reduce evaporative emissions from
sources such as fuel tanks and carburetors. When ORVR was originally discussed, existing
activated charcoal technology was determined to be the appropriate technology for the capture
and storage of refueling related fuel vapors. This continues to be the case today, as all known
ORVR-equipped vehicles utilize some type of activated charcoal.
The activated charcoal is contained in a canister, which is made from a durable material that
can withstand the fuel vapor pressures, vibration, and other durability concerns. For vehicles
without ORVR systems, canisters are sized to handle evaporative emissions for the three-day
diurnal test with the canister volume based on the fuel tank capacity. A dual fuel tank chassis
configuration may require a separate canister for each fuel tank.
1.2.3.4 Purge Valve
The purge valve is the electro-mechanical device used to remove fuel vapors from the fuel
tank and canister by routing the vapors to the running engine where they are burnt in the
combustion chamber. This process displaces some amount of the liquid fuel required from the
fuel tank to operate the engine and results in a small fuel savings. The purge valve is controlled
by the engine or emission control electronics with the goal of removing the necessary amount of
36
-------
captured fuel vapors from the canister in order to prepare the canister for subsequent fuel vapor
handling needs of either the next refueling event or vapors generated from a diurnal event. All
on-road vehicles equipped with a canister for evaporative emissions control utilize a purge valve.
Depending on the design, a dual fuel tank chassis configuration may require a separate purge
valve for each fuel tank.
1.2.3.5 Design considerations for Unique Fuel Tanks
The commercial truck market gasoline applications incorporate several fuel tank options that
may require unique ORVR design considerations. While most commercial vehicle fuel tanks are
similar to the already ORVR-compliant complete vehicles in the 8500 to 14,000 GVWR class,
some of the commercial vehicles include larger tank sizes (50 to 70 gallons) or may have a dual
tank option. As described above, the canister sizing will be a function of the required amount of
fuel vapor handling during refueling. Larger fuel tanks will require larger canisters with more
activated charcoal than historically found in other gasoline vehicles. Some design challenges
will likely exist in designing the canister system to handle the large vapor volumes while
balancing the restriction to flow through the larger activated charcoal containing canisters.
Dual fuel tank systems, which have very limited availability, may also require some unique
design considerations. Typically, the canister is located in very close proximity to the fuel tank
to properly manage the refueling fuel vapors efficiently with minimal distance between the tank
and canister. Dual fuel tanks may require duplicate ORVR systems to have the necessary
flexibility to manage the refueling vapors, particularly since the fuel tanks are filled
independently through separate filler pipe assemblies.
A small portion of the commercial truck market gasoline applications have fuel tanks that are
similar in design to diesel fuel tanks located on the outside of the frame. These tanks are
typically cylindrical or rectangular in shape with the gas cap directly on the top of the tank and
do not have a fill neck. These type of fuel tanks may require unique approaches such as a
mechanical seal built into the fuel tank filling location where the fuel cap is normally located, or
they may require a design that adds a filler pipe for a liquid seal approach.
1.3 Fuels Considerations
Both the compression- and spark-ignition engine technologies discussed above are capable of
running on alternative fuels (e.g., natural gas, biodiesel). We have typically applied the gasoline-
and diesel-fueled engine standards to alternatively-fueled engines based on the combustion cycle
of the alternatively-fueled engine: applying the gasoline-fueled engine standards to spark-
ignition engines and the diesel-fueled engine standards to compression-ignition engines. The
sections below discuss some of the available alternative fuels in more depth.
1.3.1 Natural gas
With relatively low natural gas prices (compared to their peak values) in recent years, the
heavy-duty industry has become increasingly interested in engines that are fueled with natural
gas. It has some emission advantages over diesel, with lower engine-out levels of both NOx and
PM. Several heavy-duty CNG engines have been certified with NOx levels better than 90 percent
below US 2010 standards. However, because natural gas must be distributed and stored under
pressure, there are additional challenges to using it as a heavy-duty fuel.
37
-------
Liquified Petroleum Gas (LPG) is also used in certain lower weight-class urban applications,
such as airport shuttle buses, school buses, and emergency response vehicles. LPG use is not
extensive, nor do we project it to grow significantly in the timeframe of this rulemaking.
1,3.2 Biodiesel-
Over the last decade, biodiesel content in diesel fuel has increased under the Renewable Fuels
Standard. In 2010, less than 400 million gallons of biodiesel were consumed in the U.S.,
whereas in 2018, over 2 billion gallons of biodiesel were being blended into U.S. diesel fuel.
While the biodiesel content in diesel fuel averaged around 3.5 percent in 2018, biodiesel levels
range from 0 to 20 volume percent in highway diesel fuel. As discussed further below, with
increasing volumes of biodiesel in the fuel stream, greater attention has been given to the
influence of biodiesel on highway diesel fuel quality. The following sub-sections discuss how
biodiesel is produced and the importance of purification processes for removing potential metal
containments, standards for biodiesel fuel quality, potential impacts of metals in biodiesel fuel on
diesel engines, aftertreatment systems, and emissions, as well as efforts by EPA and others to test
the metal content of biodiesel samples across the country.
1.3.2.1 Biodiesel Production and a Potential for Metals
Biodiesel can be made from various renewable sources, such as vegetable oil, animal fat, or
waste cooking oil. It is produced through transesterification of the oil or fat with methanol,
which results in mono-alkyl esters and the co-product glycerin. The process occurs in the
presence of a catalyst, typically sodium or potassium methoxide or hydroxide.46 Following
transesterification and separation of the glycerin, the biodiesel must be purified, which is usually
done by extracting, distilling, or filtering the impurities into water.
The purification process is essential to address the potential for metal and other impurities in
biofuel. There are number of potential sources of metal contamination in biofuel production.
These include:
1) Vegetable oil seeds used to produce feedstock contain high concentrations of sodium (Na),
potassium (K), calcium (Ca), magnesium (Mg), and potassium (P), as well as aluminum (Al),
iron (Fe), manganese (Mn), zinc (Zn), and smaller concentrations of other metals.47
2) The potassium and sodium methoxide catalysts which break down triglycerides to methyl
esters (NaOH and KOH can also be used) can contribute metals to biodiesel. These metals can
form soaps with free fatty acids, and the soaps in both the metal esters and glycerine forms are
reacted with acid (hydrochloric acid) to convert the soaps to free fatty acids so they can be more
easily removed. Sodium hydroxide is added to neutralize any acid added to eliminate soaps.
3) Methyl esters are washed, distilled or filtered to remove the metals added as catalysts. The
wash water is recycled, and metal ions can accumulate in the wash water. Hard wash water
containing CaCCb, Mg(OH)2, CaSC>4 is found in Rocky Mountain states and the Midwest, and
these water-soluble compounds can accumulate in the residual water found in biodiesel.
K Biodiesel is different from renewable diesel which is much more similar to petroleum refined diesel than it is to
biodiesel.
38
-------
4) The medium used to filter methyl esters could also contribute to metals in the biodiesel.
The filter material is typically made up of diatomaceous earth which is primarily silica
containing alumina, iron oxide and calcium oxide. In addition, small amounts of calcium or
magnesium can be added to the fuel from the purification process.48'49
1.3.2.2 Standards for Biodiesel Fuel Quality
Biodiesel quality, including metal content, is specified by ASTM D6751-19 for B100 fuels.
ASTMD6751-19 specifies a limit of 5 ppm for combined Na and K (group 1A metals) and a
limit of 5 ppm for combined Ca and Mg (group 2A metals) using the EN14538 inductively
coupled plasma optical emission spectroscopy (ICP-OES) measurement method.50 ASTM
D6751-18 also specifies a 10-ppm limit on P (group 5 metal) using the ASTMD4951
inductively coupled plasma atomic emission spectroscopy (ICP-AES) measurement method.51
The limits on metals in ASTM D6751 are meant to be protective when biodiesel is used in
blends (e.g., B20, BIO). Fuel quality for biodiesel blends in the B6 to B20 range is specified by
ASTM D7467-19.52 This specification does not contain a metal limit for these biofuel blends
because, as the method states, the concentration would likely be too low to measure using the
ICP-OES method specified (EN 14538). Similarly, D975 addresses BO to B5 and does not have
a metals specification (just a total ash % limit of 0.01%).53 Thus, the basis for control of metals
in biodiesel blends is control of the B100 blend stock. This is because if the B100 fuel is under
the ASTM D6751-19 limit, the combined Na + K and Mg + Ca will likely be below 1 ppm
respectively for B20 and lower blends. Yet, the actual metal content of today's fuels can be
challenging to quantify when it is lower than the 1 ppm level specified for B20 and lower blends,
because of the detection limit of the current test methods. The detection limit of the EN14538 is
1 ppm for each metal, and the method includes a statement if the metal is below the limit of
detection of the method, then it is not included in the reporting calculation. Efforts to quantify
biodiesel metal contents below 1 ppm are discussed in 1.3.2.5 below.
1.3.2.3 Potential Impacts of Metals on Engine and Emission Control Devices
Across a range of concentrations, metals in biodiesel can be present as ions, abrasive solids or
soluble metallic soaps. Abrasive solids can contribute to wear of fuel system components,
pistons and rings, as well as contribute to engine deposits. Soluble metallic soaps have little
impact on wear but may contribute to diesel particulate filter plugging and engine deposits.
Metal accumulation in diesel particulate filters can increase pressure drops and result in shorter
times between maintenance intervals.54'55 A level of 1 mg/kg (1 part per million) of trace metal
in the fuel result in an estimated accumulation of about 22 g of trace metal in diesel particulate
filters per 100,000 miles (assuming a fuel economy of 15 mpg and 100% trapping efficiency).54
Metallic fuel contaminants can also accumulate on fuel injectors, or be converted to oxides,
sulfates, hydroxides or carbonates in the combustion process, which forms an inorganic ash that
can deposit onto the exhaust emission control devices found in modern diesel engines.56 Alkali
metals are well known poisons for catalysts used in emission control devices, and have been
shown to negatively impact the mechanical properties of ceramic substrates.57'58 Alkali metal
hydroxides such as Na and K are volatilized in the presence of steam and can, therefore,
penetrate the catalyst washcoat or substrate.
39
-------
1.3.2.4 Potential for Emissions Impacts of Metals in Biodiesel
Numerous studies have collected and analyzed emission data from diesel engines operated on
biodiesel blended diesel fuel with controlled amounts of metal COntent.55'59'60'61'62'63'64'65'66 Some
of these studies show an impact on emissions, while others do not. However, four factors need
to be considered when reviewing these studies:
1. These studies were conducted using accelerated aging protocols and exposure to these
metals from the fuel consumed in a more conventional manner could cause different
effects (larger or smaller) than what these studies show,
2. The emissions testing studies were designed to test the effect of metal content in
biodiesel if the metal content was at the ASTM limit, however, as shown below, biodiesel
likely contains a lower metal content than the standard.
3. Different manufactures use different catalyst formulations and different physical layouts
for emission aftertreatment systems, and while one manufacturer might be less
susceptible to metals contamination, others may be more affected. This issue relates to
factor #4:
4. When these studies examined the effect of metals on heavy-duty engines, they studied the
impact of these metals based on current engine and aftertreatment configurations over the
current regulatory useful life. This rule requires heavy-duty engines to comply with a
more stringent NOx standard and a longer useful life. The longer useful life will expose
the aftertreatment devices to increased amounts of metals during the useful life
(compared to many of today's engines that often operate beyond the current regulatory
useful life and are already exposed to more metals after their regulatory useful life).
Also, the engine manufacturers may change the composition and configuration of their
aftertreatment devices, which could affect how fuel metals affect the aftertreatment
devices.
Brookshear et al. 2012 studied the impact of Na on heavy-duty diesel engine aftertreatment
devices.61>L In this accelerated aging study, they doped a B20 fuel to 5,000 ppm each of Na and
S and aged to an equivalent 435,000 miles. They found impacts on SCR function if the SCR was
positioned before the DPF. There was no impact on the DOC or DPF.
Lance et al. 2016 also studied the effect of Na on heavy-duty diesel engine aftertreatment.63'M
They doped their B20 fuel with Na to a level of 14 ppm or 14 times the pseudo 1 ppm limit of a
B20 fuel and accelerated aged the aftertreatment out to 435,000 miles. The results indicated an
acceleration of DPF ash buildup and platinum group metal migration from the DOC/CDPF to the
SCR. The results of the system performance, including degradation in performance, are shown
in Figure 1-7. The results indicated that the degradation in NOx performance can be attributed to
degradation of all aftertreatment components.
L Author affiliations: University of Tennessee and Oak Ridge National Lab
M Author affiliations: Oak Ridge National Lab, Cummins, and National Renewable Energy Lab
40
-------
0.50
0.45
_ 0.40 ¦
j=
£¦0.35
1
-0.30
*
c
? 0.25
Z. 0.20
«
5 0.15
£
~ 0.10
0.05
0.00
Figure 1-7: SCR performance over the hot start HDDE FTP63- N
Williams et al. 2011 studied the effect of Na and Ca on a 2008 non-road 8.8L Caterpillar
diesel engine, a MAN D2066 10.5 L diesel engine, and a 2008 Cummins 8.3L diesel engine.66 0
They doped their B20 fuel to 27 times the pseudo 1 ppm Na and Ca limit of a B20 fuel and
accelerated aging of the emission control systems out to 150,000 and 435,000 miles. The results
showed no significant degradation in the thermo-mechanical properties of cordierite, aluminum
titanate, or silicon carbide DPFs after exposure to 150,000-mile equivalent biodiesel ash and
thermal aging. It is estimated that the additional ash from 150,000 miles of biodiesel use would
also result in moderate increases in exhaust backpressure for a DPF. A decrease in DOC activity
was seen after exposure to 150,000-mile equivalent aging, resulting in higher HC slip and a
reduction in NO2 formation. The exposure of a cordierite DPF to 435,000-mile equivalent aging
resulted in a 69% decrease in the thermal shock resistance parameter. The metal-zeolite SCR
catalyst experienced a slight loss in activity after exposure to 435,000-mile equivalent aging.
This catalyst, placed downstream of the DPF, showed a 5% reduction in overall NOx conversion
activity over the HDDT test cycle.
I
yi
G
'Ex
-o
o
CI
"O
u ta
o y
"2 A
CL
O c
—
"8 as
Is
I -
— —
c 2
M&.
5 -
U n
°2 n
c 2
c -
U T3
g ^
- <
c 0.
0
Q
Note that DG indicates that the specific component is not an aged part, but a new degreened part
Author affiliations: National Renewable Energy Lab, Manufacturer of Emission Controls. BASF. Caterpillar, and
Umicore AG
41
-------
Williams et al. 2013 studied the effect of Na, K and Ca on a 2011 LD 6.7L diesel engine
aftertreatment.56'p They doped their B20 fuel to 14 times the pseudo 1 ppm Na and Ca limit of a
B20 fuel and accelerated aged the emission control systems out to 150,000 miles. The authors
aged sets of production exhaust systems that included a DOC, SCR catalyst, and DPF. Four
separate exhaust systems were aged, each with a different fuel: ULSD containing no measurable
metals, B20 containing sodium, B20 containing potassium, and B20 containing calcium.
Analysis of the aged catalysts included Federal Test Procedure emissions testing with the
systems installed on a Ford F250 pickup, bench flow reactor testing of catalyst cores, and
electron probe microanalysis (EPMA). The thermo-mechanical properties of the aged DPFs were
also measured.
EPMA imaging of aged catalyst parts found that both the Na and K penetrated into the
washcoat of the DOC and SCR catalysts, while Ca remained on the surface of the washcoat.
Bench flow reactor experiments were used to measure the standard NOx conversion, NH3
storage, and NH3 oxidation for each of the aged SCR catalysts. Flow reactor results showed that
the first inch of the SCR catalysts exposed to Na and K had reduced NOx conversion through a
range of temperatures (Figure 1-8 and Figure 1-9) and also had reduced NH3 storage capacity.
The SCR catalyst exposed to Ca had similar NOx conversion and NH3 storage performance
compared to the catalyst aged with ULSD.
Chassis dynamometer vehicle emissions tests were conducted with each of the aged catalyst
systems installed onto a Ford F250 pickup. Regardless of the evidence of catalyst deactivation
seen in flow reactor experiments and EPMA imaging, the vehicle successfully passed the 0.2
gram/ mile NOx emission standard with each of the four aged exhaust systems. This indicates
that if catalyst volumes are chosen to account for degradation, the emission control system can
accommodate some loss in catalyst activity since deactivation occurred only in the first inch of
the catalyst and did not affect overall NOx emissions.56
p Author affiliations: National Renewable Energy Lab, Oak Ridge National Lab, Manufacturers of Emission
Controls, BASF, Ford, and University of Tennessee
42
-------
-10
100 150 200 250 300 350 400 450 500 550 600 650
Catalyst Bed Temperature (°C)
Figure 1-8: SCR NQx conversion for the first inch of aged SCR catalysts.56
100 150 200 250 300 350 400 450 500 550 600 650
Catalyst Bed Temperature (°C)
Figure 1-9: SCR NOx conversion for the seventh inch of aged SCR catalysts.56
1.3.2.5 Testing for Metals in Biodiesel
The National Renewable Energy Laboratory (NREL) has conducted several studies on the
metal content of biodiesel. The NREL studies generally look at the fuel quality in fuel samples
taken across from the country. In some cases, samples are taken at the refinery where B100 was
sampled and in other cases they are taken at the pump. These studies provided analysis of metal
content for fuel samples collected and analyzed in the 2006, 2007, 2008, 2011, and 2016
timeframes. The 2006, 2007, and 2011 studies analyzed the metal content of B100 fuel, while
the 2008 and 2016 studies analyzed the metal content of biodiesel blends up to approximately
B20. Some samples taken during the 2006 study were acquired prior to the finalization of the
combined Na + K limit in ASTM D6751 (May 2006). 48-49-67-68-69-70-71
43
-------
These results indicate that over the 2006 to 2018 time frame the incidence of off specification
biofuel decreased over time. A summary of the off-specification samples can be found in Table
1-4.
Table 1-4: NREL Fuel Samples off Specification for ASTM D6751 (or equivalent B20) limit for Na + K and
Ca + Mg.
Number
Number
of
of
NREL Fuel Study
Biodiesel Content
Samples
Samples
Total Number of
Year
off spec
off spec
Samples
for Na
for Ca
+ K
+ Mg
2006 (pre-D6751)
B100
7
1
24
2006 (post-D6751)
B100
0
1
15
2007
B100
3
3
55
2008
B0.2 to B90
6
0
34
2011
B100
1
0
67
2016
B0 to B22
0
1
35
2017
B100
1
0
459
2018
B100
0
0
491
The NREL studies prior to 2016 focused on identifying gross exceedances of the blends and
blend stocks. As noted above, the analytical method specified in ASTM D6751-18 (EN 14538)
affords a detection limit of 1 ppm, which is adequate to ensure whether or not biofuel blend
stocks (B100) are compliant with the Na + K and Ca + Mg limits. In addition to determining
compliance with the ASTM D6751-18 limit, it is also important to determine the actual metal
content of these fuels in order to assess the levels that aftertreatment systems will be exposed to
over their full useful life. The 2016 NREL study used measurement equipment and procedures
capable of testing down to lower metal levels, which is useful for understanding the actual metal
content of biodiesel blends. Table 1-5 summarizes the test procedures and level of detection
(LOD) levels for each year of the NREL studies.
Table 1-5: Test procedure LODs for NREL studies.
Year of NREL Study
Test Procedure
Level of Detection (ppm)
Na
K
Ca
Mg
2006
ASTM D5185
?
?
?
?
2007
ASTMD7111
1
1
0.1
0.1
2008
ASTMD7111
0.5
0.5
0.5
0.1
2011
ASTMD7111
1
1
0.1
0.1
2016
UOP-389
0.1
0.1
0.1
0.1
2016
ICP-MS
0.029
0.001
0.0005
0.001
2017
EN 14538
1
1
1
1
2018
EN 14538
1
1
1
1
44
-------
In the 2016 study, NREL took steps to improve their understanding of testing accuracy and
improve the testing resolution of the analysis by utilizing three different measurement methods,
two of which afforded very low detection limits. In this study, biodiesel blends were analyzed
for metal content using the Universal Oil Products Method 389 (UOP-389), Microwave Plasma
Atomic Emission Spectroscopy (MP-AES), and Inductively Coupled Plasma Mass Spectroscopy
(ICP-MS) methods.72 This allowed for comparisons of metal content from the three different
testing methods. Almost all the results for the MP-AES testing method were below the LOD,
though, so this method will not be further discussed. The results of the study for UOP-389 and
the ICM-MS are presented in Table 1-6.
45
-------
Table 1-6: NREL 2016 Metals results for UOP-389 and ICP-MS72
State
Biodiesel
Content
Error
UOP-389 ICP-AES
ICP-MS
Na
K
Ca
Mg
Na
K
Ca
Mg
Fe
Vol %
Vol %
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
FL-1 (fleet)
22.01
1.76
0.14
0.04
0.08
0.01
(0.029)
0.021
0.008
0.026
0.065
FL-2
18.38
1.59
0.04
0.01
0.05
0.01
(0.029)
0.033
0.002
0.014
0.146
GA
19.36
1.63
0.06
0.02
0.06
0.01
(0.029)
0.034
0.002
0.011
1.44
NC-1 (fleet)
20.4
1.68
0.06
0.01
0.09
0.02
(0.029)
0.02
0.002
0.01
0.004
NC-2
0
0.71
0.02
0.01
0.02
0.01
No Data
No Data
No Data
No Data
No Data
PA-1
17.73
1.56
0.09
0.01
0.05
0.02
(0.029)
0.014
0.002
0.009
0.012
PA-2 (fleet)
20.31
1.68
0.08
0.02
0.04
0.03
(0.029)
0.031
0.002
0.019
0.017
MA
21.35
1.73
0.04
0.01
0.04
0.01
(0.029)
0.02
0.002
0.015
0.016
VA (fleet)
18.56
1.6
0.08
0.02
0.06
0.02
(0.029)
0.028
0.012
0.011
0.086
IA (fleet)
16.1
1.48
0.09
0.03
0.04
0.01
(0.029)
0.1
0.017
0.014
0.249
IL
15.3
1.44
0.07
0.01
0.03
0.01
(0.029)
0.02
0.007
0.019
0.005
IN
20.34
1.68
0.2
0.07
0.1
0.03
(0.029)
0.237
0.106
0.031
0.075
KS
17.27
1.53
0.09
0.02
0.53
0.06
(0.029)
0.11
1.366
0.166
1.557
KY
20.11
1.67
0.05
0.01
0.03
0.01
(0.029)
0.022
0.002
0.015
0.069
MI
20.51
1.69
0.11
0.02
0.06
0.01
(0.029)
0.045
0.122
0.025
3.117
MN-1
10.68
1.22
0.04
0.01
0.04
0.01
(0.029)
0.016
0.002
0.025
0.172
MN-2
10.61
1.22
0.02
0.01
0.03
0.01
(0.029)
0.016
0.002
0.048
0.003
MO
19.44
1.64
0.09
0.01
0.05
0.01
(0.029)
0.034
0.01
0.012
0.019
OH-1
20.4
1.68
0.05
0.01
0.05
0.01
(0.029)
0.013
0.002
0.011
0.033
OH-2
8.84
1.13
0.01
0.01
0.03
0.01
(0.029)
0.015
0.002
0.007
0.003
TN-1
20.24
1.68
0.1
0.01
0.06
0.01
(0.029)
0.019
0.003
0.006
0.085
TN-2 (fleet)
20.02
1.67
0.24
0.02
0.08
0.01
(0.029)
0.029
0.012
0.009
0.361
LA
12.26
1.3
0.04
0.01
0.03
0.01
No Data
No Data
No Data
No Data
No Data
TX-1
17.14
1.53
0.06
0.01
0.08
0.01
(0.029)
0.01
0.002
0.005
0.003
TX-2
1.4
0.78
0.07
0.01
0.04
0.01
(0.029)
0.012
0.002
0.006
0.004
TX-3
1.2
0.77
0.04
0.01
0.03
0.01
(0.029)
0.011
0.002
0.005
0.004
CO
21.99
1.76
0.09
0.01
0.05
0.01
(0.029)
0.023
0.094
0.035
0.037
ID (fleet)
19.64
1.65
0.22
0.03
0.12
0.03
(0.029)
0.085
0.065
0.034
0.044
AZ (fleet)
20.39
1.68
0.07
0.01
0.04
0.01
(0.029)
0.085
0.065
0.034
0.044
CA-1
18.84
1.61
0.05
0.02
0.03
0.01
(0.029)
0.016
0.002
0.019
0.009
CA-2
22.12
1.77
0.07
0.01
0.04
0.01
(0.029)
0.016
0.002
0.011
0.03
CA-3 (fleet)
19.9
1.66
0.06
0.01
0.04
0.01
(0.029)
0.033
0.002
0.018
0.005
NM
19.61
1.65
0.1
0.02
0.05
0.01
(0.029)
0.025
0.002
0.005
0.284
OR-1
20.1
1.67
0.06
0.02
0.03
0.02
(0.029)
0.012
0.002
0.005
0.009
OR-2
20.15
1.67
0.05
0.01
0.05
0.01
(0.029)
0.021
0.002
0.005
0.011
Average
0.079
0.016
0.064
0.014
(0.029)
0.037
0.058
0.021
0.243
Average
Na + K and Ca + Mg
0.047
0.039
0.033
0.040
Note: Values in parenthesis are below the detection limit and are reported at the detection limit.
Differences in the results between the two methods could be due to interferences by other
metals, or other aspects of the molecules, with the measuring method. The measurement
46
-------
differences could also be due to the unique ionization efficiency of each element and how well
the instalment ionizes the element of interest. Even small differences could impact the results at
sub 1-ppm levels. The difference in sample preparation techniques can also have a significant
effect on the results. The UOP-389 method uses acid digestion followed by ashing, while the
ICP-MS method used a simpler preparation of sample dilution and direct analysis. The UOP-389
method was developed for the analysis of petroleum products and blending components,
including biodiesel blends, and uses a wet ashing method that is unique to this procedure, which
is why it was selected for this project.
Although several samples contained elevated amounts of Ca and Mg (KS, MI, IN, and CO)
that were well above the level of other samples, these trace levels would still be very low in the
B100 blend stock, with the exception of one sample (KS). Overall, the NREL data suggest that
metal contents in biodiesel have decreased over time and, as of 2018, are generally very low
across samples. Nevertheless, small sample sizes could be biasing the results.46'47'67'68'69'70'71
To that end, in 2019 an engine manufacturer raised concerns to EPA that biodiesel is the
source of high metal content in highway diesel fuel, and that higher biodiesel blends, such as
B20, are the principal problem.73 The engine manufacturer observed higher than normal
concentrations of alkali and alkaline earth metals (Na, K, Ca, and Mg) in their highway diesel
fuel samples, and observed fouling of the aftertreatment control systems of their engines, which
caused an associated increase in emissions. The engine manufacturer sampled the ash that was
fouling their fuel injectors and aftertreatment devices and determined the ash to be composed of
sodium sulfate, sodium carboxylates, and sodium chloride, which they claimed were from
biodiesel. The engine manufacturer recommends limiting biodiesel blends to 5 percent biodiesel
(B5). After hearing engine manufacturer concerns about the metal content in biodiesel in early
2019, EPA began to focus a previously developed fuel sampling program on biodiesel metal
content. Below, we summarize the information that we obtained through that sampling program.
Separate from hearing about engine manufacturer concerns with biodiesel metal content, EPA
began a process to determine the metal content of different fuels in early 2016. The most
prominent concern to EPA at that time was the blending of less refined natural gas liquids with
ethanol to produce E85 blends; however, EPA recognized the need to understand the metal
content of all fuels. EPA initiated a sampling effort in late 2016 to obtain samples of diesel fuel,
gasoline, natural gas liquids, jet fuel, biodiesel and ethanol. The samples were collected in acid-
washed glass bottles using a clean hands/dirty hands sampling procedure to reduce the chance for
contamination during the sampling process. Fuel samples were collected from both small and
large fuel production plants, in case facility size plays a role in the metal content of the fuel.
Samples were also taken in different geographical regions. Fuel production facilities likely use
different feedstocks based on their geographical region, and the different feedstock types could
affect the fuels metal content. Because of the cost and effort involved in obtaining these fuel
samples, a limited number of samples of each fuel type were obtained. This screening study
could be expanded upon the detection of high metal content in any of the fuels.
Approximately 100 samples were obtained across the various fuel types. A subset of these
samples (27 B100 samples) were recently sent to the California Department of Food and
Agriculture (CDFA) laboratory. These samples were analyzed for biodiesel regulated metals,
including Na, K, Ca, Mg and P, and also tested for Molybdenum (Mo), Boron (B), Barium (Ba),
Copper (Cu), Manganese (Mn), Silica (Si), Titanium (Ti), Vanadium (V) and Zinc (Zn). The
47
-------
California lab utilized the ASTM D7111-16 ICP-AES method that returned detection limits of
0.023 (Na), 0.052 (K), 0.013 (Ca), 0.004 (Mg), 0.001 (P), 0.006 (Mo), 0.013 (B), 0.001 (Ba),
0.005 (Cu), 0.001 (Mn), 0.017 (Si), 0.003 (Ti), 0.002 (V), and 0.005 (Zn) ppm.74 The results of
the analysis are shown in Table 1-7 and Table 1-8.
Na was above the detection limit for 22 of the samples, with the highest result at 564 ppb. K
was only above the detection limit for 3 of the samples, with the highest result at 660 ppb. Ca
was above the detection limit for 9 of the samples, with the highest result at 551 ppb. Mg was
only above the detection limit for 5 of the samples, with the highest result at 133 ppb. The
highest result for combined Na and K was 744 ppb, while the highest result for combined Ca and
Mg was 662 ppb.
All of the 27 B100 fuel samples from this test program were compliant with the ASTM
D6751-18 limit of 5 ppm for Na + K and Ca + Mg respectively, and the results showed that
levels were at less than 20% of the limit for two of the samples, while the rest were at less than
10% (and in most cases well below that) of the limit. A reduction of 80% in metal content for
B20 and a reduction of 95% in metal content for B5 fuel blends would result in a maximum Na +
K content of 149 ppb and 37 ppb respectively for the B100 fuel with the highest Na + K content.
Ca + Mg would be 132 ppb and 33 ppb respectively.*2
Q This assumes no contribution from the diesel fuel used to formulate the blends.
48
-------
Table 1-7: EPA 2017 Metals results for ICP-AES analysis of Na, K, Ca, and Mg performed by CDFA.
Sample ID
Area of US
Na
(ppm)
K
(ppm)
Ca
(ppm)
Mg
(ppm)
Na + K
(ppm)
Ca + Mg
(ppm)
25982
East
0.171
0.211
0.131
0.133
0.382
0.264
25988
East
0.084
0.660
(0.013)
0.018
0.744
[0.0311
25998
Midwest
0.241
(0.052)
(0.013)
0.005
[0.293]
[0.018]
26004
Midwest
(0.023)
(0.052)
(0.013)
(0.004)
(0.075)
(0.017)
26006
Midwest
0.081
(0.052)
(0.013)
(0.004)
[0.1331
(0.017)
26083
Midwest
0.181
(0.052)
0.026
(0.004)
[0.233]
[0.030]
26084
Midwest
0.188
(0.052)
0.025
(0.004)
[0.2401
[0.0291
26088
Midwest
0.111
(0.052)
(0.013)
(0.004)
[0.1631
(0.017)
26090
Midwest
0.135
(0.052)
(0.013)
(0.004)
[0.187]
(0.017)
26092
Midwest
0.201
(0.052)
(0.013)
(0.004)
[0.2531
(0.017)
26095
Midwest
0.378
(0.052)
(0.013)
(0.004)
[0.4301
(0.017)
26164
South
0.193
(0.052)
0.040
(0.004)
[0.245]
[0.044]
26165
South
0.044
(0.052)
(0.013)
(0.004)
[0.0961
(0.017)
26166
South
(0.023)
(0.052)
(0.013)
(0.004)
(0.075)
(0.017)
26217
West
0.108
(0.052)
(0.013)
(0.004)
[0.160]
(0.017)
26218
West
0.564
(0.052)
0.027
(0.004)
[0.6161
[0.0311
26219
West
0.278
(0.052)
0.143
0.031
[0.3301
0.174
26248
West
0.303
(0.052)
0.551
0.111
[0.355]
0.662
26250
South
0.031
0.186
(0.013)
(0.004)
0.217
(0.017)
26253
South
(0.023)
(0.052)
(0.013)
(0.004)
(0.075)
(0.017)
26254
South
(0.023)
(0.052)
(0.013)
(0.004)
(0.075)
(0.017)
26256
South
(0.023)
(0.052)
(0.013)
(0.004)
(0.075)
(0.017)
26283
South
0.331
(0.052)
(0.013)
(0.004)
[0.3831
(0.017)
26830
South
0.379
(0.052)
0.016
(0.004)
[0.431]
[0.020]
26833
East
0.290
(0.052)
(0.013)
(0.004)
[0.3421
(0.017)
27581
East
0.226
(0.052)
0.044
(0.004)
[0.2781
[0.0481
27955
West
0.270
(0.052)
(0.013)
(0.004)
[0.322]
(0.017)
Average
0.182
0.085
0.046
0.014
0.267
0.060
*Values in (parenthesis) are below the detection limit and are reported at the detection limit.
**Na + K and Ca + Mg values in [square brackets] include one element that is below the detection limit and is
included in the calculation at the detection limit.
49
-------
Table 1-8: EPA 2017 Metals results for ICP-AES analysis of Mo, P, B, Ba, Cu, Mn, Si, Ti, V, and Zn
performed by CDFA.
Sample
Area of US
Mo
P
B
Ba
Cu
Mn
Si
Ti
V
Zn
ID
(ppm)
(ppm)
(ppm)
(ppm)
(ppm)
(ppm)
(ppm)
(ppm)
(ppm)
(ppm)
25982
East
0.070
0.301
2.577
0.057
0.157
0.141
2.364
0.138
0.134
0.041
25988
East
0.012
0.305
1.220
0.006
0.032
0.014
2.896
0.004
0.010
(0.005)
25998
Midwest
(0.006)
0.207
0.126
0.002
0.006
0.002
0.927
(0.003)
0.003
(0.005)
26004
Midwest
(0.006)
0.073
0.154
(0.001)
(0.005)
(0.001)
0.034
(0.003)
(0.002)
(0.005)
26006
Midwest
0.008
0.138
0.081
(0.001)
(0.005)
(0.001)
0.100
(0.003)
(0.002)
(0.005)
26083
Midwest
(0.006)
0.188
0.062
0.003
0.022
0.004
0.241
(0.003)
(0.002)
(0.005)
26084
Midwest
(0.006)
0.182
0.053
0.002
0.023
0.004
0.233
(0.003)
(0.002)
(0.005)
26088
Midwest
(0.006)
0.735
0.029
(0.001)
(0.005)
(0.001)
0.020
(0.003)
(0.002)
(0.005)
26090
Midwest
0.008
0.226
0.017
(0.001)
0.019
(0.001)
0.172
(0.003)
(0.002)
(0.005)
26092
Midwest
0.010
0.165
0.038
(0.001)
0.046
(0.001)
1.168
(0.003)
0.004
(0.005)
26095
Midwest
(0.006)
0.100
0.029
(0.001)
0.054
(0.001)
1.043
(0.003)
0.004
(0.005)
26164
South
0.022
0.232
0.022
(0.001)
0.024
0.003
0.220
(0.003)
(0.002)
(0.005)
26165
South
(0.006)
0.949
(0.013)
(0.001)
(0.005)
(0.001)
0.046
(0.003)
(0.002)
(0.005)
26166
South
(0.006)
0.019
0.014
(0.001)
(0.005)
(0.001)
0.032
(0.003)
(0.002)
(0.005)
26217
West
0.011
0.101
0.024
(0.001)
0.035
(0.001)
0.350
(0.003)
0.021
(0.005)
26218
West
(0.006)
(0.001)
(0.013)
(0.001)
0.025
(0.001)
(0.017)
(0.003)
(0.002)
0.049
26219
West
(0.006)
0.302
0.014
(0.001)
0.035
(0.001)
0.104
(0.003)
(0.002)
(0.005)
26248
West
(0.006)
0.365
(0.013)
0.004
0.037
(0.001)
1.776
(0.003)
(0.002)
(0.005)
26250
South
(0.006)
(0.001)
(0.013)
(0.001)
(0.005)
(0.001)
0.130
(0.003)
(0.002)
(0.005)
26253
South
(0.006)
0.025
0.023
(0.001)
(0.005)
(0.001)
1.077
(0.003)
(0.002)
(0.005)
26254
South
(0.006)
0.021
0.027
(0.001)
(0.005)
(0.001)
1.239
(0.003)
(0.002)
(0.005)
26256
South
(0.006)
(0.001)
(0.013)
(0.001)
(0.005)
(0.001)
0.306
(0.003)
(0.002)
(0.005)
26283
South
(0.006)
0.133
(0.013)
(0.001)
0.193
(0.001)
0.592
(0.003)
0.020
(0.005)
26830
South
(0.006)
0.193
(0.013)
(0.001)
0.066
(0.001)
1.516
(0.003)
0.004
(0.005)
26833
East
0.012
0.237
(0.013)
(0.001)
0.035
0.007
0.397
(0.003)
0.013
(0.005)
27581
East
(0.006)
0.038
(0.013)
(0.001)
(0.005)
(0.001)
0.461
(0.003)
0.004
(0.005)
27955
West
(0.006)
(0.001)
(0.013)
(0.001)
(0.005)
(0.001)
0.115
(0.003)
(0.002)
(0.005)
Average
0.010
0.194
0.172
0.004
0.032
0.007
0.651
0.008
0.009
0.008
*Values in parenthesis are below the detection limit and are reported at the detection limit.
The California Air Resources Board (ARB) CDFA inspectors carried out a biodiesel sampling
campaign throughout California during the spring and fall of 2019 collecting three hundred fifty-
five (355) biodiesel and diesel fuel samples from both #2 diesel labeled pumps and biodiesel
labeled pumps in the state of California.75
These samples were analyzed by the same lab as the 27 EPA samples mentioned above and
afforded the same detection limits. The primary focus of analysis was to examine the average
and observed range of concentration for Na, K, Ca, Mg and P of the biodiesel samples and the
diesel samples.
Statistical analysis of the samples showed that the Na, K, Ca, Mg and P concentrations in all
of the 355 collected fuel samples across California were significantly lower than the worst case
expected concentrations for a B20 fuel blended from B100 blend stock that is at the ASTM
50
-------
D6751-18 limit. Only three P samples, one Mg + Ca sample, and thirteen Na + K samples across
the entire sample set exceeded worst case expected absolute concentrations for a B5 blended
from B100 blend stock that is at the ASTMD6751-18 limit.
Na was the most abundant metal observed and was above the detection limit for 273 of 355
samples with sample 30077 exhibiting the highest result at 837 ppb. The rest of the metals were
largely below detection limits. K was only above the detection limit for 14 of 355 samples with
sample 15162 exhibiting the highest result at 172 ppb. Ca was above the detection limit for 24
of 355 samples with sample 30062 exhibiting the highest result at 168 ppb. Mg was only above
the detection limit for 32 of 355 samples with sample 15162 exhibiting the highest result at 238
ppb. Sample 30077 exhibited the highest result for combined Na and K at 889 ppb, while sample
30062 exhibited the highest result for combined Ca and Mg at 353 ppb. P was above the
detection limit for 92 of 355 samples, with sample 30077 exhibiting the highest result at 862 ppb.
Tables containing the maximum and average concentrations with standard deviation can be
found in the CARB comments to the ANPRM in the docket.75
A review of the NREL, EPA, and ARB data sets indicate that biodiesel fuel is compliant with
the ASTM D6751-18 limits for Na, K, Ca, and Mg. While the test results indicate that there is
an occasional B100 blend stock that is off specification with respect to the ASTM D6751-18
limits, and occasional biodiesel blends that are off specification to the pseudo limits, these
occurances are the exception. The NREL 2016, EPA, and CARB data sets all use measurement
methods that afford low levels of detection (sub-100 ppb), and these data sets further indicate
that the Na, K, Ca, and Mg content of biodiesel blends is extremely low in general, on the order
of less than 100 ppb. While these metals are present in biodiesel blends and testing has shown
that exposure to metals can adversely affect emission control system performance, data suggest
that the low levels measured in today's fuels are not enough to adversely affect system
performance out to full regulatory useful life values we are finalizing in this action, provided that
the engine manufacturer properly sizes the catalysts to account for the low-level exposure.
Chapter 1 References
1 Khalek, I. A., Blanks, M. G., Merritt, P. M., & Zielinska, B. (2015). Regulated and unregulated
emissions from modern 2010 emissions-compliant heavy-duty on-highway diesel engines.
Journal of the Air & Waste Management Association, 65(8), 987-1001.
2 Rudolph, J., Jacob, C.R. (2019) "Computational Insights into the Mechanism of the Selective
Catalytic Reduction of NOX: Fe- versus Cu-Exchanged Zeolite Catalysts." ACS Omega, 4,
7987-7993.
3 Lambert, C.K. "Perspective on SCR NOX control for diesel vehicles." Reaction Chemistry &
Engineering, 2019, 4, 969.
4 Fan, C., Chen, Z., Pang, L., Ming, S., Zhang, X., Albert, K. B., . . . Li, T. (2018). The influence
of Si/Al ratio on the catalytic property and hydrothermal stability of Cu-SSZ-13 catalysts for
NH3-SCR. Applied Catalysis A: General, 550, 256-265.
51
-------
5 Fedyko, J. M. and H.-Y. Chen (2015). Zeolite Catalyst Containing Metals. U. S. Patent No.
US20150078989A1, Johnson Matthey Public Limited Company, London.
6 Cui, Y., Wang, Y., Walter, E. D., Szanyi, J., Wang, Y., & Gao, F. (2020). Influences of Na+
co-cation on the structure and performance of Cu/SSZ-13 selective catalytic reduction catalysts.
Catalysis Today, 339, 233-240.
7 Fedyko, J. M. and H.-Y. Chen (2019). Zeolite Catalyst Coating Containing Metals. U. S. Patent
No. US 20190224657A1, Johnson Matthey Public Limited Company, London, UK.
8Wang, A., Wang, Y., Walter, E. D., Washton, N. M., Guo, Y., Lu, G., et al. (2019). NH3-SCR
on Cu, Fe and Cu+ Fe exchanged beta and SSZ-13 catalysts: Hydrothermal aging and propylene
poisoning effects. Catalysis Today, 320, 91-99.
9 McDonald, J. F., Schenk, C., Sanchez, L. J., & Nelson, B. J. (2011). Testing of catalytic
exhaust emission control systems under simulated locomotive exhaust conditions. SAE
Technical Paper No. 2011-01-1313.
10 Pereira, M. V. L., Nicolle, A., & Berthout, D. (2015). Hydrothermal aging effects on Cu-
zeoliteNH3-SCR catalyst. Catalysis Today, 258, 424-431.
11 Luo, J., Wang, D., Kumar, A., Li, J., Kamasamudram, K, Currier, N., & Yezerets, A. (2016).
Identification of two types of Cu sites in Cu/SSZ-13 and their unique responses to hydrothermal
aging and sulfur poisoning. Catalysis Today, 267, 3-9.
12 Kumar, A., Kamasamudram, K, & Yezerets, A. (2013). Hydrocarbon Storage on Small-Pore
Cu-Zeolite SCR Catalyst. SAE International Journal of Engines, 6(2), 680-687.
13 Kumar, A., Smith, M. A., Kamasamudram, K, Currier, N. W., & Yezerets, A. (2016).
Chemical deSOx: An effective way to recover Cu-zeolite SCR catalysts from sulfur poisoning.
Catalysis Today, 267, 10-16.
14 Steinparzer, F., Hiemesch, D., Kranawetter, E., Salmansberger, E., Stiitz, W. (2020) The
technical concept of the new BMW 6-cylinder 2nd generation modular Diesel engines
Proceedings of the 41st International Vienna Motor Symposium, Vienna, Austria, 22-24 April
2020
15 BorgWarner. "BorgWarner's EGR Technologies Reduce Emissions for Ford's Power Stroke®
Diesel Engines." Accessed on the Internet on 9/23/2022 at the following URL:
https://www.borgwarner.com/newsroom/press-releases/2017/06/15/borgwarner-s-egr-
technologies-reduce-emissions-for-ford-s-power-stroke-diesel-engines
16 Detroit. "DETROIT DD8" Available online: https://demanddetroit.com/engines/dd8/
17 Ding, C., Roberts, L., Fain, D., Ramesh, A.K., Shaver, G.M., McCarthy, J., et al.(2015). "Fuel
efficient exhaust thermal management for compression ignition engines via cylinder deactivation
and flexible valve actuation." Int.J.Eng. Res. doi:10.1177/1468087415597413.
18 Neely, G., Sharp, C., Pieczko, M., and McCarthy, J., (2020) "Simultaneous NOX and C02
Reduction for Meeting Future CARB Standards Using a Heavy-Duty Diesel CDA-NVH
Strategy," SAE Int. J. Engines 13(2).
52
-------
19 McDonald, Joseph. "Engine Modeling of LIVC for Heavy-duty Diesel Exhaust Thermal
Management at Light-load Conditions" Memorandum to Docket EPA-HQ-OAR-2019-005 5.
November 21, 2019.
20 Farrell, L., Frazier, T., Younkins, M., Fuerst, J. (2020) Diesel Dynamic Skip Fire (dDSFTM):
Simultaneous C02 and NOx Reduction. Proceedings of the of the 41st International Vienna
Motor Symposium
21 Hamedi, M., Tsolakis, A., and Herreros, J., "Thermal Performance of Diesel Aftertreatment:
Material and Insulation CFD Analysis," SAE Technical Paper 2014-01-2818, 2014,
doi: 10.4271/2014-01-2818.
22 U.S. EPA (2014) Chapter 1 - Vehicle Program Technological Feasibility. Control of Air
Pollution from Motor Vehicles: Tier 3 Motor Vehicle Emission and Fuel Standards Final Rule -
Regulatory Impact Analysis. EPA Document No. EPA-420-R-14-005.
23 Harris, T. M., Mc Pherson, K., Rezaei, R., Kovacs, D., Rauch, H., & Huang, Y. (2019).
Modeling of close-coupled SCR concepts to meet future cold start requirements for heavy-duty
engines. SAE Technical Paper No. 2019-01-0984.
24 Singh, N., Adelman, B., Malagari, S., & Hickey, K. (2019). Ultra-Low NOX Emission
Prediction for Heavy Duty Diesel Applications Using a Map-Based Approach. SAE Technical
Paper No. 2019-01-0987.
25 Zavala, B., Sharp, C., Neely, G., and Rao, S. (2020) CARB Low NOX Stage 3 Program -
Aftertreatment Evaluation and Down Selection, SAE Technical Paper 2020-01-1402,
doi: 10.4271/2020-01-1402.
26 U.S. EPA. "Crankcase Emissions for 2010+ Model Year Heavy-Duty Diesel Trucks - Test
Report." Docket EPA-HQ-OAR-2019-0055. April 2020.
27 Khalek, I., Bougher, T., Merritt, P. "Phase 1 of the Advanced Collaborative Emissions
Study." CRC Report: ACES Phase 1. Alpharetta, GA: Coordinating Research Council. 2009.
28 CARB; On-Road Advanced Technology Demonstration Project, Page 2, September 26, 2017.
https://www.arb.ca.gov/msprog/aqip/solicitations/092617onroadatdapplicationsfyl617.pdf
29 South Coast Air Quality Management District; Execute Contract to Develop and Demonstrate
Near-Zero Emission Opposed Piston Engine, December 2, 2017.
http://www.aqmd.gov/docs/default-source/Agendas/Governing-Board/2017/2017-decl-006.pdf
30 Achates Power; Opposed Piston Engine HD Demo Program, September 26, 2019.
https://ww3.arb.ca.gov/msprog/hdlownox/files/workgroup_20190926/guest/l 0_achates_hd_dem
o_program.pdf
31 Achates Power Cost Study White Paper, March 2020, Achates Power Cost Study White
Paper_March 2020.pdf
32 CARB; Opposed Piston Engine Class 8 Heavy-Duty On-Road Demonstration, January 26,
2018. https://www.arb.ca.gov/msprog/lct/pdfs/opposedpiston.pdf
53
-------
33 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.
34 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.
35 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.
36 Eng, James A. "The Effect of Spark Retard on Engine-out Hydrocarbon Emissions." SAE
Technical Paper 2005- 01-3867.
37 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.
38 Ball, D., Zammit, M., Wuttke, J., Buitrago, C. "Investigation of LEV-III Aftertreatment
Designs." SAE Technical Paper 2011-01-0301.
39 Serrano, D., Lavy, J., Kleeman, A., Zinola, S., Dumas, J., Le Mirronet, S., Heitz, D. "Post
Oxidation Study During Secondary Exhaust Air Injection for Fast Catalyst Light-Off" SAE
Technical Paper 2009-01-2706.
40 Lee, D., Heywood, J. "Effects of Secondary Air Injection During Cold Start of SI Engines."
SAE Technical Paper 2010-01-2124.
41 Schell, T., Holle, C., Wunsch, R. et al. (2020) "M 254 - the Mercedes-Benz 4-Cylinder
Gasoline Engine of the Future." Proceedings of the 41st Vienna Motor Symposium.
42 81 FR 73478, October 2, 2016.
43 U.S. EPA. "Heavy-Duty Highway Gasoline and Diesel Certification Data (Model Years:
2015 - Present)". Available online: https://www.epa.gov/sites/production/files/2020-01/heavy-
duty-gas-and-diesel-engines-2015-present.xlsx. Accessed June 2020.
44 SGS-Aurora, Eastern Research Group, "Light Heavy-Duty Gasoline Vehicle Evaporative
Emissions Testing." EPA-420-R-19-017. December 2019.
45 U.S. Environmental Protection Agency. "Summary of "Light Heavy-Duty Gasoline Vehicle
Evaporative Emissions Test Program"" EPA-420-S-19-002. December 2019.
46 Graboski, M.S. and McCormick, R.L., "Combustion of Fat and Vegetable Oil Derived Fuels
in Diesel Engines," Progress in Energy and Combustion Science, 24 125-163, 1998.
47 Chaves, Eduardo S. et al; Metals and phosphorus determination in vegetable seeds used in the
production of biodiesel by ICP OES and ICP-MS; Microchemical Journal 96 (2010) 71 - 76.
54
-------
48 Alleman, T.L., McCormick, R.L., "Results of the 2007 B100 Quality Survey," NREL/TP-
540-42787, Golden, CO: National Renewable Energy Laboratory, March 2008.
49 Alleman, Theresa L., Quality Parameters and Chemical Analysis for Biodiesel Produced in
the United States in 2011; NREL/TP-5400-57662; March 2013.
50 CSN EN 14538 Fat and oil derivatives - Fatty acid methyl ester (FAME) - Determination of
Ca, K, Mg and Na content by optical emission spectral analysis with inductively coupled plasma
(ICP OES).
51 ASTM, 2018. "Standard Specification for Biodiesel Fuel (BlOO)Blend Stock for Distillate
Fuels", American Society for Testing and Materials, D6751-18.
52 ASTM, 2019. "Standard Specification for Diesel Fuel Oil, Biodiesel Blend (B6 to B20)",
American Society for Testing and Materials, D7467-19.
53 ASTM, 2019. "Standard Specification for Diesel Fuel", American Society for Testing and
Materials, D975 - 19b.
54 Jaaskelainen, H. Biodiesel Standards & Properties.
https://www.dieselnet.com/tech/fuel_biodiesel_std.php
55 Sappok, A., Wong, V., 2007. "Impact of Biodiesel on Ash Emissions and Lubricant
Properties Affecting Fuel Economy and Engine Wear Comparison with Conventional Diesel
Fuel", US DOE, 13th Diesel Engine-Efficiency and Emissions Research (DEER) Conference,
Detroit, MI, August 2007,
https://wwwl.eere.energy.gov/vehiclesandfuels/pdfs/deer_2007/session5/deer07_sappok.pdf
56 Williams, Aaron; Impact of Fuel Metal Impurities on the Durability of a Light-Duty Diesel
Aftertreatment System; SAE 2013-01-0513; 4/08/2013.
57 Dou, D. and Balland, J., "Impact of Alkali Metals on the Performance and Mechanical
Properties of NOX Adsorber Catalysts," SAE Technical Paper 2002-01-0734, 2002,
doi: 10.4271/2002-01-0734.
58 Cavataio, G., Jen, H., Dobson, D., and Warner, J., "Laboratory Study to Determine Impact of
Na and K Exposure on the Durability of DOC and SCR Catalyst Formulations," SAE Technical
Paper 2009-01-2823, 2009, doi:10.4271/2009-01-2823.
59 McCormick, Robert; Biodiesel Performance with Modern Engines, CRADA #: CRD-05-153;
March 2016
60 Schroder, Jorg; Accelerated performance and durability test of the exhaust aftertreatment
system by contaminated biodiesel; International Journal of Engine Research; Vol. 18 (10) 1067-
1076, 2017
61 Brookshear, Daniel W.; Investigation of the effects of biodiesel-based Na on emission control
components; Catalysis Today 184 (2012) 205-2018
62 Iojoiu, E.; Biofuel Impact on Diesel Engine After-Treatment: Deactivation Mechanisms and
Soot Reactivity; Emission Control Science and Technology; (2018) 4:15-32; December 18 2017.
55
-------
63 Lance, Michael; Evaluation of Fuel-Borne Sodium Effects on a DOC-DPF-SCR Heavy-Duty
Engine Emission Control System: Simulation of Full-Useful Life; SAE 2016-01-2322.
64 Kienkas, Liene; Effect of Biofuel Impurities on the Diesel Oxidation Catalyst; Degree Project
Stockholm 2017.
65 Guo, Jing; Effects of Biodiesel Blending on Exhaust Emissions; Dissertation; June 2, 2011.
66 Williams, Aaron; Impact of Biodiesel Impurities on the Performance and Durability of DOC,
DPF and SCR Technologies; SAE 2011-01-1136; 04/12/2011.
67 Alleman, T.L., 2006 B100 Quality Survey Results; Milestone Report NREL/TP-540-41549;
May 2007.
68 Alleman, Teresa L., Analysis of Biodiesel Blend Samples Collected in the United States in
2008; NREL/TP540-46592.
69 Alleman, Teresa L., Metals Analysis of Biodiesel Blends; Technical Report NREL/TP-5400-
72341; May 2019
70 Alleman, Teresa L. Assessment of BQ-9000 Biodiesel Properties for 2017; NREL/TP-5400-
75795; January 2020.
71 Alleman, Teresa L. Assessment of BQ-9000 Biodiesel Properties for 2018; NREL/TP-5400-
75796; January 2020.
72 ASTM, 2015. "Trace Metals in Organics by ICP-OES", UOP-389-15.
73 Recker, Alissa, "Fuel Quality Impacts on Aftertreatment and Engine;" Daimler Trucks, July
29, 2019.
74 ASTM, 2016. " Standard Test Method for Determination of Trace Elements in Mddle
Distillate Fuels by Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AES)",
American Society for Testing and Materials, D7111-16.
75 CARB response to comments to ANPRM February 24, 2020.
56
-------
57
-------
Chapter 2 Compliance Provisions
2.1 Compression-Ignition Engine Dynamometer Test Procedures
Test procedures are a crucial aspect of the heavy-duty criteria pollutant program. This
rulemaking is establishing several new test procedures for spark and compression ignition engine
compliance. This chapter will describe the existing test procedures as well as the development
process for the final test procedures. This includes the determination of emissions from both
engines and hybrid powertrains as well as the development of new duty cycles.
2.1.1 Current CI Test procedures
Heavy-duty compression-ignition engines currently are certified for non-greenhouse gas
(GHG) pollutants using the Heavy-Duty Diesel Engine Federal Test Procedure (HDDE FTP) and
Supplemental Emission Test Ramped Modal Cycle (SET). For 2007 and later Heavy-Duty
engines, 40 CFR part 86 - "Control of Emissions from New and In-Use Highway Vehicles and
Engines" and 40 CFR part 1065 - "Engine Testing Procedures" detail the certification process.
40 CFR 86.007-11 defines the standard settings of Oxides of Nitrogen, Non-Methane
Hydrocarbons, Carbon Monoxide, and Particulate Matter. The duty cycles are defined in Part
86. The HDDE FTP is defined in 40 CFR part 86, appendix I. The SET is defined in 40 CFR
86.1362(a). All emission measurements and calculations are defined in Part 1065, with
exceptions as noted in 40 CFR 86.007-11. The data requirements are defined in 40 CFR 86.001-
23 and 40 CFR 1065.695.
The measurement method for CO is described in 40 CFR 1065.250. For measurement of
NMHC, refer to 40 CFR 1065.260. For measurement ofNOx, refer to 40 CFR 1065.270. For
measurement of PM, refer to 40 CFR 1065.140, 1065.170, and 1065.290. Table 1 of 40 CFR
1065.205 provides performance specifications that we recommend analyzers meet. Note that 40
CFR 1065.307 provides linearity verifications that the system must meet. For the calculation
method for brake specific mass emissions for CO, NMHC, NOx and PM, refer to 40 CFR
1065.650.
2.1.1.1 HDDE FTP
The Heavy-Duty Diesel Engine Federal Test Procedure (HDDE FTP) is a transient test
consisting of second-by-second sequences of engine speed and torque pairs with values given in
normalized percent of maximum form. The cycle was computer generated from a dataset of 88
heavy-duty trucks in urban operation in New York and Los Angeles. This procedure is well-
defined, mirrors in-use operating parameters, and continues to be appropriate for assessment of
criteria pollutant emissions from heavy duty engines.
A complete HDDE FTP involves three test sequences. First, a 20-minute test is run over the
duty-cycle with the engine at the same ambient temperature as the test cell (between 68°F and
86°F). The engine undergoes a 10-minute hot-soak following the cold-start. A 20-minute hot
start test is run over the duty-cycle following the hot-soak. The HDDE FTP emission level for
the engine is determined by weighting the cold start emissions by 1/7 (about 14 percent) and the
hot-start emission results by 6/7 (about 86 percent).
58
-------
2.1.1.2 SET
The Supplemental Emission Test Ramped Modal Cycle (SET) is a 13-mode, steady-state
engine dynamometer test that replaced the steady-state modal run SET and is based on the
European Stationary Cycle (ESC). The engine is tested on an engine dynamometer over a
sequence of steady-state modes. Emissions are collected over both the steady-state and transition
portions (20-second transitions) of the test and the results are integrated to produce a single
emission result to show compliance with the standard.
The current weighting of modes within the SET for engines complying with the 2010 NOx
and Phase 1 GHG standards is given in Table 2-1. A, B, and C speeds are determined according
to 40 CFR 1065.610.
Table 2-1: SET Mode Weighting Factors for the 2010 NOx and Phase 1 GHG Standards
Speed, % Load
Weighting factors of SET
(%)
Idle
15
A, 100
8
B, 50
10
B, 75
10
A, 50
5
A, 75
5
A, 25
5
B, 100
9
B, 25
10
C, 100
8
C, 25
5
C, 75
5
C, 50
5
Total
100
Idle Speed
15
Total A Speed
23
Total B Speed
39
Total C Speed
23
2.1.1.3 Powertrain
Powertrain test procedures were created under EPA's Heavy-duty Greenhouse Gas: Phase 2
(Phase 2) rulemaking for vehicle certification.1'2 At the time of their development, no
certification procedure existed for powertrain certification of heavy-duty hybrid vehicles to any
engine standards. The powertrain test was updated for powertrain certification of the engine to
engine standards for GHG pollutants in the Heavy-Duty Engine and Vehicle Test Procedures,
and Other Technical Amendments rulemaking (hereinafter "HD Technical Amendments").3 The
powertrain certification test was finalized for certification to both the SET and FTP and is carried
out by following 40 CFR 1037.550 as described in 40 CFR 1036.510 and 1036.512 and is
59
-------
applicable for powertrain systems with the hybrid function located in the PO, PI, P2, and P3
positions.
The development of these test procedures required the addition of a speed and road grade
profile to the existing HDDE FTP, HDOC FTP, and newly finalized LLC duty cycles in the
recently finalized 40 CFR part 1036, appendix II, and to the SET in 40 CFR 1036.510.3 It also
required the development of vehicle parameters to be used in place of those in 40 CFR 1037.550;
namely vehicle test mass, vehicle frontal area, vehicle drag area, vehicle coefficient of rolling
resistance, drive axle ratio, tire radius, vehicle curb mass, and linear equivalent mass of rotational
moment of inertias. Determination of system and continuous rated power along with the
maximum vehicle speed (C speed) is also required using 40 CFR 1036.520. The combination of
the generic vehicle parameters, the engine duty-cycle vehicle speed profile, and road grade
profile fully defines the system load and this is designed to match up the powertrain load with
the HDDE vFTP, HDOC vFTP, vSET, and vLLC load for an equally powered engine.
The development of these test procedures was previously described in detail in the HD
Technical Amendments.3
2,1.2 Final updates to CI Test procedures
2.1.2.1 HDDE FTP
We are finalizing no changes to the HDDE FTP weighting factors or the duty-cycle torque
values from the duty-cycle that currently applies to criteria pollutant regulations in 40 CFR part
86 Appendix I (f)(2). We are finalizing a change to the engine speed values that does not
influence the ultimate denormalized speed, as noted below and finalized in the HD Technical
Amendments, apply to criteria pollutant certification as well. We started the migration of some
heavy-duty highway engine standard setting part test procedures to 40 CFR part 1036 in the HD
Technical Amendments. This included the migration of the HDDE FTP drive schedule to
Appendix II (c) of part 1036 in order to add vehicle speed and road grade to the duty-cycle to
facilitate powertrain testing of hybrids for compliance with the Phase 2 GHG standards.
The change that was made for GHG and is being finalized to apply to criteria pollutant
certification as well. We took the normalized vehicle speeds over the HDDE FTP duty-cycle and
multiplied them by 100/112 to eliminate the need to divide by 112 in the diesel engine
denormalization equation in 40 CFR 86.1333(a)(l)(i). This eliminated the need for inclusion of
a denormalization equation in the standard setting part and allows commonization of (between
compression and spark ignition engines) the use of the denormalization equation in 40 CFR
1065.610(c)(1) (equation 1065.610-3), with no effect on stringency.
2.1.2.2 SET
The SET weighting currently used for certification of heavy-duty highway compression
ignition engines for criteria pollutants and Phase 1 GHG has a relatively large weighting in C
speed. The C speed is typically in the range of 1800 rpm for current heavy heavy-duty engine
designs. However, it is becoming much less common for engines to operate at such a high speed
in real-world driving conditions, especially during cruise vehicle speeds in the 55 to 65 mph
vehicle speed range. This trend has been corroborated by engine manufacturers' in-use data that
has been submitted to the agencies in comments to previous rules and presented at technical
conferences.4 Thus, although the current criteria pollutant and HD Phase 1 GHG SET represents
60
-------
highway operation better than the FTP cycle, improvements have been made via the HD Phase 2
GHG program by adjusting its weighting factors to better reflect modern trends in in-use engine
operation. The most recent trends for compliance with the Phase 2 GHG standards indicate that
manufacturers are configuring drivetrains to operate engines at speeds down to a range of 1050-
1200 rpm at a vehicle speed of 65mph.
To address this trend toward in-use engine down-speeding, the agency is finalizing the
application of the refined SET weighting factors and resulting SET developed for the HD Phase
2 GHG standards to the criteria pollutant standards. The Phase 2 GHG SET mode weightings
move most of the C weighting to the A speed, as shown in Table 2-2. To better align with in-use
data, these changes also include a reduction of the idle speed weighting factor. This will apply
the Phase 2 mode weightings to both criteria pollutants and the Phase 2 CO2 emission and fuel
consumption standards starting in model year 2027.
Table 2-2: New SET Mode Weighting Factors in Phase 2
Speed/% Load
Weighting Factor in Phase 2
(%)
Idle
12
A, 100
9
B, 50
10
B, 75
10
A, 50
12
A, 75
12
A, 25
12
B, 100
9
B, 25
9
c, 100
2
C, 25
1
C, 75
1
C, 50
1
Total
100
Idle Speed
12
Total A Speed
45
Total B Speed
38
Total C Speed
5
The Phase 2 SET mode weighting moves most of the C speed weighting to A speed and
reduces the weighting factor on idle speed. These values are based on data from vehicle
manufacturers that have been claimed as confidential business information. These revised SET
weighting factors better reflect the lower engine speed operation of modern engines, which
frequently occurs at tractor cruise speeds.
To evaluate how current engines perform on this cycle, we tested a 2018 Detroit DDI 5 and a
2018 Cummins B6.7. For both engines, there was no significant difference for any of the
measured criteria pollutants between the two cycles. These results are summarized in Chapter 3.1
61
-------
of the RIA. To assess the effect of stringency between the two SET cycles, the CARB Stage 3
demonstration engine was also run on both versions of the SET. The results from these tests can
be found in Chapter 3.1 of the RIA.
2.1.2.3 LLC
Current certifications cycles (FTP and SET) and not-to-exceed (NTE) compliance
requirements do not account for emissions over sustained low load operation. This is either
because the idle time in the duty-cycle is too short, or, in-use, the operation is excluded from
compliance requirements.
We are finalizing a new low load certification cycle to address deficiencies in our current
certification duty-cycles and NTE field testing program with respect to emission control at low
load.
The National Renewable Energy Lab (NREL) and Southwest Research Institute (SwRI) under
contract with the California Air Resources Board (CARB) developed a suite of candidate low
load cycles (LLC) from urban tractor and vocational vehicle real-world activity data. The goal
of the cycle development was to develop a duty-cycle that was representative of real-world urban
tractor and vocational vehicle operations that are characterized by low engine loads, have
average power and duration adequate for demonstrating that hardware and controls needed to
deal with low load challenges are present and functional, and set an emission standard that
balances the need for NOx emission reductions and any associated GHG emission impacts.5'6'7'8
NREL combined their Fleet DNA and CARB's heavy-duty diesel vehicle activity datasets,
incorporating a total of 751 unique vehicles across the United States, to develop the LLC. The
combined dataset included vehicles form 25 distinct locations, 26 combined vocational
designations, and 55 unique fleets incorporating both urban tractor and vocational applications.
A breakdown of the applications that were included can be found in Table 2-3.
62
-------
Table 2-3: Breakdown of vehicles from combined NREL Fleet DNA and CARB datasets.
Vehicle Application
Number of Vehicles
Parcel delivery
100
Refuse pickup
90
Line Haul
84
Beverage delivery
65
Mass transit
61
Food delivery
60
Drayage
41
Utility
32
Linen delivery
30
Transfer truck
29
Tanker
25
Telecom
24
Freight
22
Public work
13
School bus
11
Agricultural
10
Snow plow
9
Warehouse delivery
9
Construction
8
Dump truck
7
Refrigerated truck
6
Local delivery
5
Towing
4
Concrete
3
Dry van
3
Delivery (other)
1
NREL initially developed a list of drive cycle metrics, including engine load specific
calculations to describe engine/vehicle operation. Vehicle operation was then broken down into
microtrips for operation over a given shift day, where a single microtrip was defined as the
duration over which a vehicle speed increases from 0 mph to the time where the vehicle stops,
including operation until the vehicle speed starts to increase above 0 mph again. 10 microtrips
were averaged using moving average windows based on a sensitivity analysis of moving average
windows of 5, 10, and 15 microtrips. The total number of microtrips from the combined data
sets was approximately 1.25 million and resulted in a trimodal distribution with two large
primary peaks and a tertiary lower load peak as shown in Figure 2-2.
63
-------
30000
% average load
Figure 2-1
Figure 2-2: Window size comparison and load distribution profile.
The microtrips were then equated to operational profiles and profiles that contained average
loads of 20% or less were considered further for construction of the LLC. These remaining
profiles were then subjected to cluster analysis to identify unique groups of operation and to
assess for outliers. K-means clustering was chosen to be applied to the dataset due to its
computational efficiency and ability to identify and remove outliers during pre-processing.
Elbow analysis was then applied to determine optimal cluster number. The resulting optimal
cluster number was three.
Representative profiles from each cluster were then selected for Greenhouse Gas Emissions
Model (GEM) simulation. Results for each cluster were ranked based on their distance to cluster
center to identify the most representative profiles. Profiles were examined for behavior and final
suitability for testing, starting with profiles closest to cluster center. Five primary modes of
operation were generally observed in the low load profiles: sustained low load, long idle,
motoring/short idle cooling (high to low load), post-cooling breakthrough (low to high load), and
mid-speed cruise motoring. Profiles found to have outlying behavior were removed and not used
for GEM modeling. These outliers were found to have one or more of the following: prolonged
periods of idle, long key off periods, and missing data. The load data broadcast by engines is not
accurate enough to allow translation of vehicle-based to engine-based profiles to create the
engine duty-cycle, so the Phase 2 GEM simulation model was used to develop the normalized
engine duty-cycle. A representative summary of ten GEM generated profiles is shown in Table
2-4.
64
-------
Initial LLC candidate duty-cycles were constructed using DRIVE to include at least one
example of each of the five primary modes of operation, incorporating five of the GEM
generated duty-cycles. It should be noted that these candidate cycles did not always include the
entire GEM generated profile if the candidate cycle could be completed in a shorter amount of
time by removing portions of the profile that did not adversely affect the target modes of
operation. Figure 2-3 gives an example of a candidate cycle incorporating representative profiles
v9892 cO, vl 1660 c5, v073 cl, v9892 cl, and vl 1806 c5.
CARB narrowed the range of candidate LLC duty-cycles to the three that best represented the
target modes of low load. The three were LLC Candidates 7, 8, and 10. Candidate cycle 7 is 90
minutes in duration, has 30 minutes of sustained low load operation, and retains the v073 cl mid-
speed cruise/motoring segment. Candidate cycle 8 is 81 minutes in duration, has 30 minutes of
sustained low load operation, and has a shorter v073 cl mid-speed cruise/motoring segment to
assess breakthrough only. Candidate cycle 10 is 70 minutes in duration, has 20 minutes of
sustained low load operation, and has a shorter v073 cl mid-speed cruise/motoring segment to
assess breakthrough only. Emission results for a representative engine compliant to EPA 2010
NOx standard for the three-candidate duty-cycles can be found in Table 2-5.
CARB selected candidate cycle 7, with an option to add auxiliary load (1.5, 2.5, and 3.5 kW
for LHD, MHD, and HHD engines respectfully), as the LLC that it moved forward with in its
Omnibus heavy-duty low NOx program. This cycle was chosen by CARB for inclusion of the
highest percentage of non-idle operation when compared to the other two candidate cycles.
CARB's decision to move forward with the option to add auxiliary load was based on making the
duty-cycle more realistic with respect to real world operation as it affords more effective
function of technologies such as cylinder deactivation (CD A) during idle.
We agree with CARB's assessment of the candidate low load cycles and their adoption of
cycle 7. Thus we are finalizing to adopt LLC 7, however we are finalizing to require the use of
auxiliary load, to bolster the current FTP and SET duty-cycles and to better align laboratory
duty-cycles with the finalized changes we are making to off-cycle testing and compliance
requirements, which afford better low load NOx reduction. We are requiring the use of auxiliary
load because this results in more representative testing with respect to real world operation. This
duty-cycle also contains vehicle speed and road grade profiles to facilitate powertrain
certification of hybrid powertrains to the engine standard. These profiles were developed in the
same manner as the HDDE FTP and HDOC FTP as well as the SET as discussed in the HD
Technical Amendments.3 The LLC can be found in 40 CFR part 1036, appendix B(d).
65
-------
Table 2-4: Representative Summary of GEM Generated Profiles for the Engine Duty-cycle
Profile
Vehicle
Cluster
Length
Avg
% Speed
Avg
% Torque
Repeats
in SwRI
Test Runs
Class
Chassis
Engine
Trans
Gears
Vocation
1
v9892
0
800
26.9
6.9
4
8
4x2
Volvo
D13
AMT
12
Food
Service
2
V11660
0
1295
21.4
6.6
3
8
6x4
Mack
MP8-415C
MT
13
Drayage
3
v075
0
1130
26.3
7.4
3
8
6x4
Mack
MP8-415C
AMT
10
Drayage
4
vll815
1
1949
11.5
8.8
3
8
6x4
Cummins
ISX 15
MT
13
Transfer
Truck
5
V11646
1
904
15.9
10.7
4
4
4x2
Cummins
ISB 6.7
AT
6
Parcel
Delivery
6
v073
1
1410
33.8
18.1
3
8
6x4
Mack
MP8-415C
AMT
10
Drayage
7
v9892
1
1616
27
10.6
3
8
4x2
Volvo
D13
AMT
12
Food
Service
8
V11660
5
615
16.2
3.5
4
8
6x4
Mack
MP8-415C
MT
13
Drayage
9
V11806
5
1810
7.5
6.8
3
8
6x4
Cummins
ISX 12
AMT
10
Transfer
Truck
10
vll817
5
739
15.3
7.7
4
8
6x4
Cummins
ISM 11
AMT
10
Transfer
Track
v9892_c0 vll660_c5 v073_cl v9892_cl vll806_c5
I A y ^ y ^ y ^ |
—Speed —Torque
11
J lib ill .I! .1 I l|[ III J I .
1
1 Mm 11 oil II MM\ J L i I
1
i 11 in UK ilr* nmlfi fill ¦ HI m k
11II. Hi lifflllL HI 1 Tfi UII t llil I] A )
nl
Sustained Low Lo Ho- Lo\lj Speed LHi-to-Lo _oi)- Hi-to-Lo Lc
- 0-
Hi 1 li fj'iHb Hi 4i 1 H
1
i 1IL ft
-------
Table 2-5: NOx Emission Levels for a 2010 Compliant Engine on Three Candidate LLCs
Candidate #
Cycle Duration
NOx Conversion
Efficiency (%)
Engine Out NOx
(g/hp-hr)
Tailpipe NOx
(g/hp-hr)
7
90
74
3.2
0.8
8
81
77
2.9
0.7
10
70
69
3.2
1.0
2.1.2.4 Powertrain
We are finalizing the use of powertrain testing as an option for certification of hybrid
powertrains to criteria pollutant standards. It is envisioned that this option will capture CO2 and
NOx co-optimization benefits for hybrid integration; including engine start/stop, electric motor
assist, electric vehicle mode, and brake energy recovery. The powertrain test procedures to
facilitate this type of certification testing have already been developed during the HD Technical
Amendments and are directly applicable to criteria pollutant testing. Duty-cycles that
incorporate vehicle speed and road grade for the HDDE FTP, HDOC FTP, and the SET were
finalized in the HD Technical Amendments where development of these cycles was discussed.3
The finalized addition of the low load cycle in Chapter 2.1.2.3 also included a discussion of the
vehicle speed and road grade profile development for powertrain testing.
2.2 Manufacturer-Run Off-Cycle Field Testing Program for Compression-Ignition Engines
The manufacturer run field testing program is crucial to ensuring compliance with the heavy-
duty criteria pollutant program. This rulemaking establishes a new test procedure for evaluating
off-cycle compression ignition engine compliance.A This chapter describes the existing test
procedure as well as the development process for the final test procedure.
2.2.1 Current Field Testing Program and Off-Cycle Standards
EPA's current regulatory program for on-highway heavy-duty engines has evolved over the
past four decades from relatively simple standards and test procedures appropriate for the
mechanically controlled engines of the 1970s and 1980s, to a multi-faceted program designed to
reduce emissions from modern computer-controlled engines. However, throughout the years, the
compliance paradigm has focused heavily on the pre-production certification process, during
which the manufacturers demonstrate that their engines will meet the applicable standards and
related requirements.
Until fairly recently, compliance was determined only by testing engines in a laboratory on an
engine dynamometer. Prior to model year 2004, engines were evaluated over a single transient
A Duty-cycle test procedures measure emissions while the engine is operating over precisely defined duty cycles in
an emissions testing laboratory and provide very repeatable emission measurements. Off-cycle test procedures
measure emissions while the engine is not operating on a specified duty cycle; this testing can be conducted while
the engine is being driven on the road (e.g., on a package delivery route), or in an emission testing laboratory. When
off-cycle testing is conducted on the road it is referred to as "field testing"; "In-use" is used to denote that testing is
done on an engine that has entered into commerce. Both duty-cycle and off-cycle testing are conducted pre-
production (e.g., for certification) or post-production (i.e., in-use) to verify that the engine meets applicable duty-
cycle or off-cycle emission standards throughout useful life (see Section III for more discussion).
67
-------
test cycle commonly referred to as the Heavy-Duty Diesel Engine FTP (HDDE FTP) cycle.
Beginning with the 2004 standards, EPA added the engine dynamometer-based Supplemental
Emission Test (SET)B, and Not-to-Exceed (NTE) emission limits that are evaluated on heavy-
duty highway engines while operating over the road.
Heavy-duty diesel engines are currently subject to Not-To-Exceed (NTE) standards that are
not limited to specific test cycles, which means they can be evaluated in the field. Data from
field testing are collected by manufacturers as described in section 2.2.1. The data is then
analyzed pursuant to 40 CFR 86.1370 and 40 CFR 86.1912 to generate a set of engine-specific
NTE events, which are intervals of at least 30-seconds when engine speeds and loads remain in
the control area. The express purpose of the NTE test procedure is to apply the emission
standard to engine operation conditions that could reasonably be expected to be seen by that
engine in normal vehicle operation and use, including a wide range of real ambient conditions.
The NTE zone defines the range of engine operation where the engine must comply. The NTE
zone is based on engine speed and load and includes some carve outs that include low load
operation (excludes load points less than 30% of rmax and /Jmax, and less than 15% of the
European Stationary Cycle speed), as described in 40 CFR 86.1370. In addition, there are carve
outs for altitude (> 5500 ft), maximum ambient temperature (100 °F at sea level, 86 °F at 5500
ft), aftertreatment temperature for NOx aftertreatment and oxidizing catalysts (carves out
operation at temperatures < 250°C), and provides a cold temperature operating exclusion for
EGR equipped engines (calculation based on intake manifold temperature, engine coolant
temperature, and intake manifold pressure).0
Heavy-duty engines are required to comply with not-to-exceed (NTE) emission limits during
real world operation. Engine manufacturers must acquire and submit data through the
manufacturer run in-use testing program. These off-cycle NTE emission limits are 1.5 (1.25 for
CO) times the laboratory certification standard or family emission limit (FEL) for NOx, NMHC,
PM and CO and can be found in 40 CFR 86.007-11. A measurement allowance value is added
on to the standard to account for measurement inaccuracies that are associated with field
measurement over short time periods and can be found in 40 CFR 86.1912. The engine
standards and measurement allowances are in Table 2-6.
Table 2-6: Engine Standards and Field Testing Measurement Allowance
NOx
PM
CO
NMHC
(g/hp-hr)
(g/hp-hr)
(g/hp-hr)
(g/hp-hr)
Engine Standards
0.20
0.01
15.5
0.14
NTE Standards
0.30
0.015
19.4
0.21
Measurement Allowance
0.15
0.006
0.25
0.01
A valid NTE event is described as an event that is 30 seconds or more in duration under
engine, aftertreatment, and ambient conditions that are within the NTE zone (i.e., do not occur
B The SET was later modified to run as a single continuous test (similar to how a transient
cycle is run) and renamed the Supplemental Emission TestRamped-Modal Cycle.
c For more on our NTE provisions, see 40 CFR 86.1370.
68
-------
during the aforementioned exclusion conditions), see 40 CFR 86.1370 and 40 CFR 86.1912. The
engine must meet a vehicle pass ratio of 90% of valid NTE events (i.e., 90% of the valid NTE
events must comply with the off-cycle standard (0.45 g/hp-hr forNOx) for the engine to be
considered compliant) as described in 40 CFR 86.1912.
We have concerns with the current NTE regulations that compliance can be achieved over the
entire operating regime of the engine due to low temperature aftertreatment exclusions, and the
narrow engine operation that has to be met for at least 30 consecutive seconds, as shown in
Figure 2-4. Removal of these exclusions will require the engine manufacturer to act to maintain
aftertreatment temperature at low load modes of operation, and this in turn will lead to better real
world performance with respect to emission compliance.9
1000 RpM 1500
Points excluded by reason:
• Intake Manifold Temperature • Aftertreatment Out Temperature
• Power • Torque
• RPM • Duration
1 NTE Event
Figure 2-4: Sample of valid and invalid NTE events, separated by exclusion zones.
Data submitted under the current regulatory program has been acquired over more than
120,000 miles of vehicle testing. The data was generated by more than 540 unique vehicles and
submitted by 14 different engine manufacturers. The percentage of test results meeting or
exceeding the 0.9 pass ratio threshold since 2005 is presented in Table 2-7. A typical data set
representing one test for an engine will contain twenty to forty valid NTE events, but under some
routes no valid NTE events are measured.
69
-------
Table 2-7: Percentage of Tests Meeting or Exceeding 0.90 Pass Ratio Threshold
Constituent
Percentage of Tests Meeting
or Exceeding the 0.9 Pass Ratio
NOx
96.0
CO
99.6
PM
99.4
The average pass ratio for each of the three constituents, for all data submitted since 2005, is
presented in Table 2-8.
Table 2-8: Average Pass Ratios for Data Submitted Since 2005
Constituent
Overall Constituent Pass Ratio
NOx
0.980
CO
0.996
PM
0.997
NTE standards have been successful in broadening the types of operation for which
manufacturers design their emission controls to remain effective. Our analysis of field-testing
data indicates that EPA's existing NTE limits did not apply to 95% of the elapsed operating time
in these studies. Furthermore, we found that emissions are high during many of the excluded
periods of operation, such as when the aftertreatment temperature drops below the catalyst light-
off temperature. For example, 96 percent of tests from 2014, 2015, and 2016 field testing orders
passed with NOx emissions for valid NTE events well below the 0.3 g/hp-hr NTE standard.
When we used the same data to calculate NOx emissions over all operation measured, not just
for valid NTE events, the NOx emissions were more than double (0.5 g/hp-hr).10 The results
were higher when we analyzed the data to only consider NOx emissions that occur during low
load events. These results suggest there may be great potential to improve real world
performance by considering more of the engine operation when we evaluate off-cycle
compliance. The average value of idle NOx for the 2014, 2015 and 2016 Test Orders is
approximately 16 g/hr.
2.2.2 Information evaluated for final updates
2.2.2.1 CE-CERT Program description
The CE-CERT study involves instrumenting about 100 trucks and collecting real world
second by second activity data over the course of one month from on-board GPS units and ECU
scan tools. The instrumented vehicles are equipped with MY 2010 and later heavy-duty diesel
engines utilizing SCR. Over 170 parameters are recorded including GPS based location, speed,
elevation, ECU based vehicle speed, ECU based engine parameters (RPM, MAP, MAF, load),
and aftertreatment variables (temperature, NOx ppm). The vehicles cover 19 different groups
based on vocational use, GVWR, and geographic operation.
Key findings of this study: Average speed varies from 41 mph for interstate line haul trucks to
9 mph for drayage trucks. Intrastate line haul trucks average 32 mph because they spend less
time on freeways and double the time idling, compared to interstate line haul trucks. Most
70
-------
vocational vehicles spent about 33% of time operating at idle, irrespective of the time they spent
on the freeway. SCR temperature for line haul operation has a bi-modal distribution with peaks
at 100°C and 260-280°C. Drayage truck operation in Northern California exhibits a peak SCR
operating temperature of around 110°C. Overall, the vehicles in this study spent 42-91% of their
operating time with SCR temperatures below 250°C. Based on a generic SCR emission control
efficacy versus SCR temperature curve, the average engine-out NOx reduction could be 16% and
69% for agricultural trucks and refuse trucks, respectively.
In addition to giving a picture of heavy-duty vehicle activity, the study also provided data on
NOx sensors "on-time". It is widely known that NOx sensors are not turned on when the
humidity level of the exhaust is high enough to produced condensed water, but there is little data
on how much time the sensors are off during heavy-duty vehicle operation. This information is
useful if NOx sensors are to be used as a compliance tool for an off-cycle standard. The sensors
were not operational, or the tailpipe NOx sensors were not reporting valid concentrations for
40% of the operating time based on the data collected. Figure 2-5 shows the time it takes until
the sensors start reporting data from a cold start. The majority of the engine out sensors report
data within the first 10 minutes after a cold start. The tailpipe sensors report data within the first
30 minutes, much later than the engine out sensors. This is understandable as the catalysts act as
a heat sink, absorbing heat from the exhaust during warm-up and preventing this thermal energy
from making it to the tailpipe sensor.
35%
3" 30%
v 25%
u
c
15 20%
tn
_c
C 15%
ns
10%
~o
o
u 5%
0%
UJJJL,,.
-¦ J-. ... . J
6 9 12 15 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60
Time Until NOxSensorOn (mins)
I Aftertreatment Inlet Sensor
Aftertreatment Outlet Sensor
Figure 2-5: NOx sensor time to on after engine cold-start
2.2.2.2 Summary of HDIUT Data
To evaluate the efficacy of current technology NOx emissions controls, EPA analyzed the
data from engines selected for testing in calendar years 2010 through 2016. This dataset covers
44 engine families, model years 2010 to 2015, from 11 manufacturers. The dataset includes
about 8 million seconds of quality-assured second-by-second data collected during 68,000 miles
of driving. The operational conditions include a wide range of driving speeds, transient and
steady-state conditions, engine loads, and exhaust temperature conditions that have implications
71
-------
for emissions control efficacy, particularly for NOx.11 For the HHD class, out of a total 159
vehicles, 109 were line-haul, 46 were delivery, and the remaining were marked as "Other" in the
metadata. Table 2-9 illustrates the distribution of family emission limit (FEL) values for the 291
vehicles tested. An FEL is a manufacturer-specified value that represents the maximum emission
rate from the engines in that group during certification testing.
Table 2-9: Number of Diesel Vehicles with MY 2010+ Engines by NOx FEL Group from the heavy-duty off-
cycle field testing Program
Regulatory
Class
NOx FE1
^ Group
Total
0.20
0.35
0.50
HHD
49
0
15
64
MHD
26
23
9
58
LHD
93
31
35
159
Urban Bus
0
10
0
10
Total
168
64
59
291
The following sections describe three analyses that use the HDIUT data to investigate the
relationship between aftertreatment temperatures and NOx emissions, and identify operations
where emissions are elevated.
2.2.2.3 HDIUT Data by MOVES OpMode
For the first analysis, the data were grouped by operating mode (OpMode) used by EPA's
MOtor Vehicle Emission Simulator (MOVES) emissions inventory model. MOVES OpModes
are defined in terms of power output using a scaled tractive power (STPt) parameter, shown in
Equation 2-1, and vehicle speed.
Equation 2-1
Avt + BvJ + Cvt + M ¦ vt(at + g • sinOt)
Where:
STPt = the scaled tractive power at time t [scaled kW or skW]
A = the rolling resistance coefficient [kW-sec/m],
B = the rotational resistance coefficient [kW-sec2/m2],
C = the aerodynamic drag coefficient [kW-sec3/m3],
m = mass of individual test vehicle [metric ton],
/scale = fixed mass factor (LHD = 5, MHD = 7, HHD = 10),
vt = instantaneous vehicle velocity at time t [m/s],
at = instantaneous vehicle acceleration [m/s2]
g = the acceleration due to gravity [9.8 m/s2]
sin 8t = the (fractional) road grade at time t
OpMode 0 is reserved for deceleration and braking events. OpMode 1 represents idle,
defined as vehicle speeds less than 1.0 mph. The remaining OpModes are defined by STPt ranges
and vehicle speed (vt) ranges of 1 to 25 mph, 25 to 50 mph, and greater than 50 mph. Figure 2-6
is a graphical representation of the MOVES OpModes, showing their relationship to STPt and vt.
72
-------
See the MOVES Technical Report on heavy-duty emission rates for more information about
OpModes.12
30
24
ao
c
CD
U
CD
Q
CUD
c
Speed: [1-25) mph
o
v
Q_
CUD
C
*+-»
to
03
O
u
1
I
I
I
Speed: [25-50) mph
o
v
Q_
CUD
C
*+-»
CO
03
O
u
1
I
I
I
I
I
I
1
Speed > 50 mph
1
I
I
I
I
o
5
o
Q_
.1 18
"+¦»
u
2 12
T5
-S! 6
ro
u
to
0
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
MOVES Operating Mode
Figure 2-6: MOVES Operating Modes (OpModes) by scaled tractive power and vehicle speed
Figure 2-7 shows the NOx emission rates in g/s for 65 vehicles from five manufacturers (by
color) measured during off-cycle field testing. The engines are categorized into eight HHD
engine families (E-l through E-8) by shape with a NOx FEL certification level of 0.20 g/bhp-hr.
Real-world operation is known to produce large variability in emission rates, compared to
certification testing performed in a lab. The graph shows that there is significant inter-engine
family and intra-engine family variability in these field tests. For example, rates for E-l are
consistently lower than E-8. There is also variability within an engine family as represented by
errors bars for each point. The spread is larger for engine families with higher emission rates, for
example E-8. Similar trends are seen in MHD and LHD engine families.
73
-------
(A
X
O
0.25
0.20 -
0.15 -
0.10
0.05
0.00
ii
. >
ill
i V
i"
i S t T t
I <¦
ili
i-i-l
E-l
E-2
E-3
E-4
E-5
E-6
E-7
E-8
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
MOVES OpMode
Figure 2-7: MOVES OpMode Emission Rates from HHD Engine Broken Down by Engine Family
Figure 2-8, Figure 2-9, Figure 2-10 show the OpMode-based NOx emission rates for vehicles
with NOx FELs of 0.20 g/bhp-hr or better by aftertreatment temperature. The figures do not
include all tested vehicles, as some tests did not report aftertreatment temperature. Figure 2-8 is
based on data from 81 vehicles with HHD engines, Figure 2-9 is from 20 vehicles with MHD
engines, and Figure 2-10 is from 42 vehicles with LHD engines. For all engines and most
OpModes, the NOx emission rates when a vehicle is operating with an aftertreatment
temperature below 250°C are more than double compared to operation with an aftertreatment
temperature above 250°C. The figures also show that high NOx emissions occur across all
OpModes and are not limited to low-speed or idle operation.
74
-------
0.15
0.12
J* 0.09
jag
x
2 0.06
0.03
0.00
de id
[1-25) mph
[25-50) mph
>50 mph
i~_L
T < 250 C
IT> 250 C
i
i, I. h
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
OpMode
Figure 2-8: NOx Emission Rates from 81 Vehicles with 0.20 g/bhp-hr FEL HHD Engines by MOVES
OpMode and Aftertreatment Temperature
jag
x
O
0.05
0.04
0.03
0.02
0.01
0.00
de id
[1-25) mph
[25-50) mph
>50 mph
T < 250 C
T> 250 C
ji
iiLii
0 1 :11 12 13 14 15 16!21 22 23 24 25 27 28 29 30!33 35 37 38 39 40
OpMode
Figure 2-9: NOx Emission Rates from 20 Vehicles with 0.20 g/bhp-hr FEL MHD Engines by MOVES
OpMode and Aftertreatment Temperature
75
-------
de id [1-25) mph [25-50) mph >50 mph
0.04
3s 0 03
x
o
2 0.02
0.01
0.00
Figure 2-10: NOx Emission Rates from 42 Vehicles with 0.20 g/bhp-hr FEL LHD Engines by MOVES
OpMode and Aftertreatment Temperature
Figure 2-11, Figure 2-12, and Figure 2-13 compare the same data in terms of operational time
by aftertreatment temperature. As expected, when the vehicles are operating at idle or low
speeds, more time is spent at the lower temperature bin. However, even at high speeds, a
nontrivial amount of time is spent at aftertreatment temperatures below 250°C.
¦ T < 250 C
¦ T > 250 C
1
1
T I
T 1
. . 1 1
i
lli.il i i i
i
I'i'iii
ll
0 1 111 12 13 14 15 16'21 22 23 24 25 27 28 29 30!33 35 37 38 39 40
OpMode
76
-------
de id
[1-25) mph
[25-50) mph
>50 mph
0.35
0.30
rc
+->
,0
•- 0.25
M-
0
§ 0.20
'¦0
2 0.15
u_
0}
1 0.10
F
0.05
0.00
0 1 !11 12 13 14 15 16!21 22 23 24 25 27 28 29 30!33 35 37 38 39 40
OpMode
Figure 2-11: Time Fraction from 81 Vehicles with 0.20 g/bhp-hr FEL HHD Engines by MOVES OpMode and
Aftertreatment Temperature
1
¦ T < 250 C
¦ T > 250 C
I
ii
I
1 11 *= -- -a — -m
1
1
1
. i,
ro
+->
£
<4-
o
c
o
"¦p
u
ro
at
E
0.45
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
de id
[1-25) mph
[25-50) mph
>50 mph
T < 250 C
IT > 250 C
ilia L i- i. - I. I. i, ix L ii i, i, ialii ia H jj jj j
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
OpMode
Figure 2-12: Time Fraction from 20 Vehicles with 0.20 g/bhp-hr FEL MHD Engines by MOVES OpMode
and Aftertreatment Temperature
77
-------
de id
[1-25) mph
[25-50) mph
>50 mph
0.45
_ 0-40
"ro
£ 0.35
h-
*5 0.30
.1 025
"¦p
ro 0.20
^ 0.15
ji o.io
0.05
0.00
0 1 !11 12 13 14 15 16!21 22 23 24 25 27 28 29 30!33 35 37 38 39 40
OpMode
Figure 2-13: Time Fraction from 42 Vehicles with 0.20 g/bhp-hr FEL LHD Engines by MOVES OpMode and
Aftertreatment Temperature
1 1
¦ T < 250C
¦ T> 250 C
I
I ,
L,
1
i i. .. i. *1 i- i- ia ¦¦ -- ii l| J
. i :
By combining the NOx emission rate data with the time data, we can estimate the total NOx
contribution by operation, as shown in Figure 2-14, Figure 2-15, and Figure 2-16. The imbedded
tables in each figure display the fraction of data in each temperature bin by operation time and
NOx mass. The emissions increase from low aftertreatment temperatures is not uniform across
all operating modes. For HHD engines (Figure 2-14), the aftertreatment temperatures spent
nearly as much time below 250°C as above it, but the contribution to total NOx is much higher
from the lower temperature operation due to the higher emission rates. The MHD and LHD
engines spent much more time in low aftertreatment temperatures conditions, and it is reflected
in a higher contribution to NOx. For all engines, the low- and mid-speed operating ranges
contribute the most NOx emissions. These figures highlight the need to consider both activity
and emission rate to effectively reduce NOx.
78
-------
de id
[1-25) mph
[25-50) mph
>50 mph
¦ T < 250 C
¦ T> 250 C
Ratio of Total
Time NOx
T < 250 degC
T > 250 degC
0.45
0.55
0.65
0.35
f
I
a-
1 -1 -1-. l ix L i,
i ii i1
1
¦
1
Li
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
OpMode
Figure 2-14: Total NOx Contribution from 81 vehicles with 0.20 g/bhp-hr FEL HHD Engines by MOVES
OpMode and Aftertreatment Temperature
de id
[1-25) mph
[25-50) mph
>50 mph
¦ T < 250 C
Ratio of Total
¦ T > 250 C
Time NOx
T < 250 degC
0.72
0.83
T > 250 degC
0.28
0.17
M
|
X
T
i i i i I.
ill1'
L 1 .. r. ii Ii Ii il
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
OpMode
Figure 2-15: Total NOx Contribution from 20 vehicles with 0.20 g/bhp-hr FEL MHD Engines by MOVES
OpMode and Aftertreatment Temperature
79
-------
0.45
0.40
"ro
£ 0.35
h-
*£ 0.30
.2 0 25
"¦p
ro 0.20
t 0.15
o
z o.io
0.05
0.00
Figure 2-16: Total NOx Contribution from 42 vehicles with 0.20 g/bhp-hr FEL LHD Engines by MOVES
OpMode and Aftertreatment Temperature
2.2.2.4 HDIUT Data by Speed
Figure 2-17, Figure 2-18, and Figure 2-19 show the data binned by speed for HHD, MHD,
and LHD vehicles, respectively, with NOx FEL at or below 0.20 g/bhp-hr. Operation below 5
km/hr was excluded to fully remove any idle time or low creep operation, where emission rates
are drastically higher.0 The 0.20 g/bhp-hr FEL vehicles' average NOx emission rate in the
lowest speed bin is much higher compared to higher speeds. The overall average NOx emission
rate, shown in the right-most bin, is more than twice the certification value of 0.20 g/hp-hr. The
figure suggests the difference is even greater for vehicles that spend significant time at lower
speeds. The average time fraction (diamond marker) and the average NOx fraction (square
marker) reveal the outsized contribution to total NOx from the low and medium speed bins
where the time fraction is lower than the NOx fraction.
¦ T < 250 C
Ratio of Total
¦ T> 250 C
Time NOx
T < 250 degC
0.59
0.70
T > 250 degC
0.41
0.30
I T ii
i
* i i i_ s i, i :i i, I, Ii ii 1
\ i .
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
OpMode
D Average NOx emission rate for HHD, MHD, and LHD are 6.38 g/bhp-hr, 7.65 g/bhp-hr, and 4.61 g/bhp-hr,
respectively.
80
-------
4,00
3.50
3.00
i 2.50
Q.
2.00
J!9
x 1.50
o
1.00
0.50
0.00
O
~
o
r\
Y
1.40
f
0,20 g/bhp-hr 0
O
0.71
0-5 km/h
~0-3mph
5-40 km/h
" 3-25 mph
40-80 km/h
" 25-50 mph
o
0
n
O Avg bsNOx
o Avg Time Fraction
~ Avg NQx Fraction
0.24
0.45
80-145 km/h
~ 50-90 mph
0-145 km/h
™ 0-90 mph
1.0
0,9
0.8
0,7 15
£
0.5 j-
0.5
0.4
0.3
0.2
0.1
0.0
o
c
o
"¦p
u
ro
Figure 2-17: Brake-specific NOx by Vehicle Speed Bins for 93 Vehicles with Hill) Diesel Engines and an FEL
of 0.20 g/bhp-hr
81
-------
3,00
2.50
2.00
Q.
-C
-Q
txo
1.50
x
O 1.00
0.50
O.OO
O Avg bsNQx
Q Avg Time Fraction
~ Avg NOx Fraction
o
~
8
o
0.72
o
~
0.20 g/bhp-hr yP1
O
O
;D
O
0.47
O
6
Q
0.24
0.47
1.0
0.9
0.8
0.7 15
£
0.5 j-
0.5
0.4
0.3
0.2
0.1
0.0
o
c
o
"¦p
u
ro
0-5 km/h 5-40 km/h 40-80 km/h 80-145 km/h 0-145 knVh
"0-3 mph ~ 3-25 mph ~ 25-50 mph ~ 50-90 mph " 0-90 mph
Figure 2-18: Brake-specific \ (K by Vehicle Speed Bins for 26 Vehicles with MHD Diesel Engines and an FEL
of 0.20 g/bhp-hr
3.00
2.50
2.00
¦
a.
-C
XI
ao
x
O
1.50
1.00
0.50
0.00
o-
o
Q
o
O Avg bsNOx
O Avg Time Fraction
~ Ave NOx Fraction
0.81
0.55
0.43
~ a
0.20 g/bhp-hr j|p
1^
0.24
1.0
0.9
0.8
0.7 15
0.6
0.5
0.4
0.3
0.2
0.1
0.0
o
c
o
v>
(J
10
0-5 km/h 5-40 km/h 40-80 km/h 80-145 km/h 0-145 km/h
™0-3 mph 3-25 mph ~ 25-50 mph ™ 50-90 mph " 0-90 mph
Figure 2-19: Brake-specific NOx by Vehicle Speed Bins for 49 Vehicles with LHD Diesel Engines and an FEL
of 0.20 g/bhp-hr
82
-------
2.2.2.5 HDIUT Data by Work-Based Window
Figure 2-20 shows a comparison of brake-specific NOx (g/bhp-hr) calculated using the
standard method ("bsNOx method") and CO2 based method ("NOx/CCh method"). This graph is
based on measurement data from 85 HHD diesel vehicles with NOx FEL < 0.20 g/bhp-hr. The
binning is by average power of the window over engine rated power. The windows are
continuous and non-exclusive - window n+l starts at time t=n+1. The amount of time in each
window is based on the engine power demand during the window. This analysis led to 2.90
million windows. The height of the columns represents the mean of all the windows in that
power bin and the error bars represent standard deviation of the mean. The 95% confidence
interval is not shown since windows are not independent. The CCh-based method is more robust
at very small loads, such as the 0-5% and 5-10% average power windows. In these cases, the
small amount of work done (bhp-hr) leads to higher brake-specific NOx values for the standard
method while also causing very large standard deviation. The CO2 based method addresses the
low load in the denominator issue while also not penalizing vehicles that have lower CO2
emissions, by normalizing against CO2 overwork (the second term in the equation). Another
takeaway is that emissions are much higher at lower loads, lowest at loads near the FTP and SET
cycles, and then creep up at higher loads. This suggests the engines are tuned to perform best at
loads/conditions similar to the certification cycle while less optimized for other real-world
operation.
4.00
3.50
.C 3.00
I
Q.
^ 2.50
bfl
2.00
X
2 1-50
IO
1.00
0.50
0.00
Figure 2-20: Brake-specific NOx by Window Average Power Bins for 85 Vehicles with HHD Diesel Engines
and an FEL of 0.20 g/bhp-hr
2.2.2.6 HDIUT Data by Simulated Cycle
Figure 2-21 and Figure 2-22 show analysis using simulated cycles. Figure 2-21 shows the
drive cycles, which are converted to an OpMode time distribution, which is then combined with
I bsNOx mean bsNOxmethod =
NOx,
window
bhp — hr
window
NOx/C02 mean —method = NOx™nd™ * C02testday
CO2 C02wlndow bhp - hrtestday
~ HD FTP ~ Ramped Mode Cycle
I |1 iKti 1/11 .. j' I' r r
83
-------
the OpMode based emission rates for each vehicle in the HDIUT data set. The MOVES national
run OpMode distribution is also included for comparison in Figure 2-22. The Vehicle FTP and
UDDS drive cycles have similar speed traces and thus their OpMode distributions are also
similar. The transient drive cycle has a considerable amount of low-speed transient operation,
which shows up as higher time spent in OpModes 11-14 and particularly OpMode 12. The
MOVES national run for combination-long haul trucks has most of its operation at highway
speeds, and thus most of the time is allocated to OpModes 33 and above.
Vehicle FTP Cycle
-Speed (mph)
6.41 miles
1,200 sees
1200
100
200
300 400
Time (s)
500
600
700
70
60
SO
£
E
4U
T3
a;
30
cu
a.
i/i
2U
10
0
UDDS Cycle
5.55 miles
1.060 sees
4-
4-
400 600 800
Time (s)
Each driving cycle was converted to an OpMode based
time distribution based on the MOVES default road-
load coefficients for combination long-haul trucks.
Road grade was set to zero for the UDDS and Transient
Cycles.
Figure 2-21: Vehicle Speed Profile of HD Duty Cycles
84
-------
0.45
_ 0.40
"ro
jg 0.35
0.30
0.25
O
.2 °-20
+¦>
u
ro 0.15
u_
Q, 0.10
E
H 0.05
0.00
¦ Sim. Vehicle FTP
¦ Sim. UDDS
0 Sim. Transient
¦ MOVES Combi Long-Haul
lLU
1 ' 11 12 13 14 15 16'21 22 23 24 25 27 28 29 30' 33 35 37 38 39 40
MOVES OpMode
Figure 2-22: MOVES OpMode Time Fraction for each Simulated HD Combination Long-Haul Duty Cycle
Figure 2-23 and Figure 2-24 show the cycle average NOx emission rates calculated using the
per vehicle OpMode based emission rate and the drive cycle OpMode time distribution shown in
Figure 2-22. The average rate and spread for Vehicle FTP and UDDS are similar since the cycles
are similar. The transient cycle produces the highest average rate because of low-speed operation
that has lower aftertreatment temperature. The transient cycle also has the largest spread, similar
to the larger spread for the low-speed analysis in Figure 2-17. The MOVES cycle has the lowest
average and spread because the operation is predominantly in the high-speed zone, where
emission rates are better controlled.
85
-------
3.50
3.00
_ 2.50
i-
.C
2.00
-Q
1.50
X
o
Z 1.00
0.50
0.00
A
A Avg of 93 Vehicles
&
A ^
1.10 A
0.75
£ 0.80
s
0.20 g/bhp-hr 1
¦
Simulated
Vehicle FTP
Simulated
UDDS
Simulated
Transient
MOVES National
Combi Long-Haul
Figure 2-23: Brake-specific NOx emissions by simulated cycle for HHD diesel engines with NOx FEL of 0.20
g/bhp-hr
14.00
12.00
^ 10.00
Q)
E 8.00
ao
x 6.00
o
z
4.00
2.00
0.00
2.84
o
2.97
o
8
~ Avg of 93 Vehicles
.4.64
1.39
Simulated
Vehicle FTP
Simulated
UDDS
Simulated
Transient
MOVES National
Combi Long-Haul
Figure 2-24: Distance-specific NOx emissions by simulated cycle for HHD diesel engines with NOx FEL of
0.20 g/bhp-hr
86
-------
2,2.3 Final Updates to CI Engine Off-Cycle Test Program and Off-Cycle Standards
The focus of the current off-cycle NTE standards and off-cycle NTE compliance testing is
operation at relatively high load; the data analysis procedure thus excludes a wide range of
vehicle operations that occur in the real world, in particular operations at lower loads.
Importantly, the excluded portion of the data makes up the bulk of vehicle operation, specifically
areas where NOx production is high.
To improve the coverage of the off-cycle standard, test procedures, and field testing program,
we have finalized updates to the off-cycle standard and test procedures to include a broader
range of vehicle operation that is now covered by the regulated off-cycle standard. To keep the
results representative of actual engine/aftertreatment performance and minimize issues with
temporally misaligned data, we have finalized an analysis methodology based on a series of
moving average windows (MAW).
2.2.3.1 Background on Euro VI MAW
The European Union Euro VI emission standards for heavy-duty engines requires testing for
in-service conformity starting with model year (MY) 2014 engines using portable emission
measurement systems (PEMS).13'14 The intent is to confirm whether heavy-duty engines
continue to comply with the emissions standards while in use, under normal operation, over time,
under real world conditions. Manufacturers must check for "in-service conformity" by operating
their engines over a mix of urban, rural, and freeway driving on prescribed routes using portable
emission measurement system (PEMS) equipment to measure emissions. Compliance is
determined using a work-based windows approach where emissions data are evaluated over
segments or "windows." A window consists of consecutive 1 Hz data points that are summed
until the engine performs an amount of work equivalent to the European transient engine test
cycle (World Harmonized Transient Cycle).
Engines are tested over a mix of urban, rural, and freeway driving. Testing starts at 18
months and a minimum of 25,000 km, and continues every two years thereafter out to seven
years and 700,000 km. There are no carve-outs for engine load or aftertreatment low
temperature operation. There are carve-outs for altitude (> 5577 ft), maximum ambient
temperature (100 °F at sea level), and minimum ambient temperature > 20 °F. There is no cold
start emission measurement requirement. Emission and work integration start when engine
coolant temp >70°C and is stable to within ± 2°C, or 20 minutes after engine start, whichever is
first. Vehicle payload must be 50 to 60 % of maximum.
Compliance is determined using a work-based windows approach where emissions are
evaluated over data segments or windows. For the window to be considered valid, the average
power within a window must be > 15% of engine maximum power for MY 2014 -2018 and >
10% for MY 2019 and later. The vehicle must accumulate at least 5 complete windows over its
shift day and 50% of the windows must be valid. Compliance is demonstrated at 1.5 times the
EURO VI emission limit and there is no separate measurement allowance to account for field
testing measurement uncertainty (it is built into the conformity factor multiplier of 1.5).
EPA and others have compared the performance of US-certified engines and Euro VI-
certified engines and concluded that the European engines' NOx emissions are comparable to US
2010 standards-certified engines under city and highway operation, but lower in light-load
87
-------
conditions.15 This suggests that manufacturers responded to the Euro VI test procedures by
designing their emission controls to perform well over broader operation. EPA's final
rulemaking expands our off-cycle standard and test procedures to capture nearly all real-world
operation. Our final approach is similar to the European field testing program, with key
distinctions that improve upon the Euro VI approach, as discussed below.
2.2.3.2 Final Updates
The updated off-cycle testing data analysis method uses a MAW methodology similar to that
established in the Euro VI emission standards. However, most carve-outs are eliminated.
Additionally, in order to adequately capture all vehicle operation, there is no minimum power
requirement for valid windows. There are no prescribed routes for our field testing compliance
program, as the previous NTE program required data to be collected in real-world operation. In
what we believe to be an improvement to a work-based window, the final rule uses a moving
average window (MAW) approach consisting of time-based windows. Instead of basing window
size on an amount of work, the MAW includes a window size of 300 seconds.E The time-based
windows are intended to equally weight each data point collected.
We also recognize that it is difficult to develop a single standard that is appropriate for the
entire range of operation that heavy-duty engines experience. For example, a numerical standard
for CO2 specific NOx that is technologically feasible under worst case conditions such as idle,
would be higher than the levels that are feasible when the aftertreatment is functioning optimally.
Thus, the final rule has separate standards for distinct modes of operation. The 300-second
windows constructed from the second-by-second field data are grouped into one of two bins
using the nominal "normalized average CO2 rate" from the certification test cycles to identify the
boundaries. The normalized CO2 rate is defined as the average CO2 rate in the window divided
by the engine's maximum CO2 rate. The engine's maximum CO2 rate is defined as the engine's
rated maximum power multiplied by the engine's family certification level (FCL) for the FTP
certification cycle.
Windows with a normalized average CO2 rate of 6 percent or less (6 percent is equivalent to
the average power of the low-load certification cycle) are classified as idle and binned together
(bin 1). Windows with a normalized average CO2 rate greater than 6 percent are classified as
non-idle operation and binned together (bin 2).
The emissions performance of the binned data in bin 1 is determined using the sum of the
total mass of emissions divided by the sum of CO2 mass emissions. Emissions performance for
the binned data in bin 1 is determined using the mass rate (total mass of emissions divided by
total time) of the emissions. This "sum-over-sum" approach successfully accounts for all
emissions; however, it requires the measurement system (PEMS) to be accurate across the
complete range of emission concentrations.
As mentioned previously, there is a separate MAW-based standard for each bin. In the NTE-
based program, the NTE standards are 1.5 times the certification duty-cycle standards. Similarly,
E Our evaluation includes weighing our current understanding that shorter windows are more sensitive to
measurement error and longer windows make it difficult to distinguish between duty cycles.
88
-------
for the MAW-based standards, the off-cycle standards for each bin correspond to one or more
laboratory-based cycles, with each bin having its own standard.
2.2.3.3 Data Collection and Exclusion
For the HDUIT, emissions data are to be collected from key-on (t = 0) until the end of the
shift day when the engine is turned off. Data are to be collected once every second (i.e., 1-Hz
data).
From the data collected at 1 Hz, some data points are to be excluded from the remaining
process. Data to be excluded are (1) data collected when an analyzer or flow meter is performing
in-service zero and span drift checks or zero and span calibrations and emission data cannot be
collected, (2) data collected where the engine is off except in some circumstances where the
vehicle under testing is equipped with stop-start and/or automatic engine shutdown systems as
described in the preamble Section III.C, (3) data collected during infrequent regeneration events,
(4) data collected when any approved AECD for emergency vehicle applications is active, and
(5) data collected when ambient temperatures are below 5 °C, or when ambient temperatures are
above the altitude-based value determined using Equation 40 CFR 1036.530-1.
2.2.3.4 Defining Windows
With the extended idle times frequently present in HDIUT samples, a work-based window
approach would include longer periods of time for these windows, and the methodology would
be very sensitive to small inaccuracies in power measurement. To ensure the final set of
windows more accurately reflects the operation of the vehicle, we have adopted a time-based
window approach, where each window contains an equal amount of time rather than an equal
amount of work, as in the Euro VI work-based window approach.
For this methodology, a window will consist of the summation of 300 consecutive 1-Hz data
points (i.e., a 300-second window). The windows are continuous and non-exclusive, with
subsequent windows beginning one second after previous windows (i.e., at the next data point).
The first window will begin at initial key-on (t = 0), and the final sequential window will begin
300 seconds before the last data point taken. To limit the impact of instances where data
exclusions would reduce the weighting of an individual data point, exclusions of < 600 seconds
are removed and the remaining data concatenated. For exclusions > 600 seconds, the final pre-
exclusion window begins 300 seconds before the exclusion, and the next subsequent window
begins immediately after the exclusion.
Except for the data points at the beginning and end of the test and those around long data
exclusions, this methodology equally weights emissions at each data point during the off-cycle
testing. We believe this is appropriate, as the under-weighted data points consist of a small
percentage of the HDIUT data, which contain a minimum of 10,800 1-Hz data points.
2.2.3.5 Emission bins
The agency recognizes that including operation currently excluded from the standard,
including low-load operation and low aftertreatment temperature, will result in a higher range of
variability in both the vehicle operation represented, and in the data captured during testing.
Thus, we are differentiating the data collected by vehicle operation, and independently set
standards for each operational characteristic.
89
-------
To differentiate among various types of operation, the windows are divided among two bins
that are characterized by the normalized average CO2 rate: an idle bin (bin 1) and a non-idle load
bin (bin 2). The normalized CO2 rate of each window is defined as the total window CO2 mass
divided by the 300-second window length and then divided by the maximum CO2 rate of the
engine. The engine's maximum CO2 rate is defined as the engine's rated maximum power
multiplied by its family certification level (FCL) for the FTP certification cycle.
The two bins are defined as follows:
• Bin 1: window normalized average CO2 rate < 6 %
• Bin 2: window normalized average CO2 rate > 6 %
2.2.3.6 Bin Size and Test Validity
For a test to be considered valid, bin 1 must contain a minimum of 2,400 windows and bin 2
must contain at least 10,000 windows. To ensure there are enough windows in bin 1, the engine
may be idled at the end of the shift day. If the vehicle has tamper-resistant idle-reduction
technology that prevents idling, populate bin 1 with additional windows by setting the 1-Hz
emission rate for all regulated pollutants to zero as described in § 1036.415(g) to achieve exactly
2,400 bin 1 windows. If bin 2 contains fewer than 10,000 windows, or bin 1 contains less than
2,400 windows after inclusion of the optional end-of-day idle period, the vehicle must be tested
over an additional shift day. The resulting windows from the second or subsequent shift day are
added to the appropriate bin, so that all windows from all shift days are included.
Using data from 168 previous HDUIT tests of one shift day each, the 1-Hz data from these
tests were collected into windows and binned according to the above process as seen in Figure
2-25. Of these single shift day tests, 98% contained over 10,000 windows in bin 2, and 80%
contained over 2,400 windows in bin 1. From these data, we estimate that nearly all tests would
be valid with a single shift days' worth of data, assuming manufacturers take advantage of the
optional end-of-day idle period.
90
-------
14400
12000
c
1q
u
ro
-------
Equation 2-3
/ g \ Total bin emission
el— J =
^hr' Total bin time
2.2.3.8 Bin Emissions and Standards
These standards apply to the sum-over-sum emissions value for the entire bin, as shown in
Equation 2-2 and Equation 2-3. This methodology accounts for all emissions included in a
particular bin equally and reduces the influence of potential errors in data collection.
2.3 Spark-Ignition Test Procedures and Standards
Spark-ignition test procedures are a crucial aspect of the heavy-duty criteria pollutant
program. This rulemaking establishes new test procedures and requirements for spark-ignition
engine compliance. This section describes the existing test procedures as well as the
development process for the new requirements. This includes the determination of emission
levels as well as the development of new duty cycles.
2,3.1 Current SI Test procedures
Heavy-duty spark-ignition engines currently are certified for criteria pollutants using the
Heavy-Duty Otto-Cycle Engine Federal Test Procedure (HDOE FTP). For 2007 and later
Heavy-Duty engines, 40 CFR part 86 - "Control of Emissions from New and In-Use Highway
Vehicles and Engines" and 40 CFR part 1065 - "Engine Testing Procedures" detail the
certification process. 40 CFR 86.007-11 defines the standard settings of Oxides of Nitrogen,
Non-Methane Hydrocarbons, Carbon Monoxide, and Particulate Matter. The HDOE FTP duty
cycle is defined in 40 CFR part 86, appendix I. All emission measurements and calculations are
defined in Part 1065, with exceptions as noted in 40 CFR 86.007-11. The data requirements are
defined in 40 CFR 86.001-23 and 40 CFR 1065.695.
The measurement method for CO is described in 40 CFR 1065.250. For measurement of
NMHC refer to 40 CFR 1065.260. For measurement of NOx refer to 40 CFR 1065.270. For
measurement of PM, refer to 40 CFR 1065.140, 1065.170, and 1065.290. Table 1 of 40 CFR
1065.205 provides performance specifications that we recommend analyzers meet. Note that 40
CFR 1065.307 provides linearity verifications that the system must meet. For the calculation
method brake specific mass emissions for CO, NMHC, NOx and PM refer to 40 CFR 1065.650.
2.3.1.1 HDOE FTP
The current HDOE FTP is a transient test consisting of second-by-second sequences of engine
speed and torque pairs with values given in normalized percent of maximum form. The cycle
was computer generated from a dataset of 88 heavy-duty trucks in urban operation in New York
and Los Angeles. This procedure is well-defined, mirrors in-use operating parameters, and
continues to be appropriate also for the continued assessment of criteria pollutant emissions from
heavy duty engines.
A complete HDOE FTP involves three test sequences. First, a 20-minute test is run over the
duty-cycle with the engine at the same ambient temperature as the test cell (between 68°F and
86°F). The engine undergoes a 10-minute hot-soak following the cold-start. A 20-minute hot
start test is run over the same duty-cycle following the hot-soak. The HDOE FTP emission level
92
-------
for the engine is determined by weighting the cold start emissions by 1/7 (14 percent) and the
hot-start emission results by 6/7 (86 percent).
2.3.1.2 Test Procedure Engine Mapping Improvements
The heavy-duty FTP test cycle is composed of second-by-second speed and torque targets that
are based on the engine's design operating speeds and the torque levels produced over the full
range of allowed speeds. In order to determine the torque level at any engine speed, a mapping
of the engine is performed prior to the actual FTP testing. The mapping is a sweep across the
mechanically- or electronically-allowed operating speeds to determine the highest possible
torque level the engine can produce at any specific speed. From this "maximum torque" sweep,
the FTP targets are determined for the subsequent transient FTP test and for any additional test
(e.g., SET).
As noted above, the measured torque values are intended to represent the maximum torque the
engine can achieve under fully warmed-up operation with the appropriate design fuel grade (i.e.,
regular grade octane fuel) across the allowed range of engine speeds from idle to the typically
electronically limited highest RPM. The intent is to reflect a torque value that is maintained if
the engine is stabilized at a specific speed over a longer period of time such as what might be
observed in a "real world" condition when an engine is held by a transmission in a specific RPM
by the selected gear.
Electronic control of all aspects of engine hardware and operation has resulted in some
challenges to performing the mapping of the engine. Variable torque levels have been observed
in engine testing related to such things as electronic control response to fuel octane, anticipated
exhaust thermal conditions, transmission torque limiting models, and other electronic features
incorporated into the engine management strategies. This torque variability has been particularly
evident after the change to the mapping test procedures in 40 CFR 1065.510(b)(5)(i), which
requires that an engine be mapped by performing a transient sweep from idle to maximum rated
speed. Prior to this change in the test procedure, the mapping procedure required the engine to
stabilize at discrete engine speed breakpoints before recording the engine torque value for that
engine speed that would be used for the FTP and other testing.
There are two potential improvements to reduce the torque measurement variability. One
option could be to perform the sweep in both directions (i.e., idle to rated maximum speed and
back to idle) and determine the torque value at any speed to be the highest measured torque level.
Typically, we would expect any torque limiting AECDs (e.g., power enrichment delay) to be
active during the sweep up, allowing the maximum torque to be determined during the sweep
down after any torque limiting AECD has concluded. A second option could require
manufacturers to turn off the torque limiting AECDs during the torque mapping sequence and
allow them to perform the procedure as written today. For example, an AECD that results in
delay of component thermal protection due to a temperature model would not be allowed for the
power mapping sequence since any stabilized operation at a temperature limited speed point
would typically require thermal protection. During the transient FTP test however, the AECD
could be active to simulate real world operation and the implemented thermal management
strategies.
93
-------
2,3.2 Summary of Updates Considered for SI Test Procedures and Standards
We are updating the location of our highway heavy-duty engine regulations, moving from the
current 40 CFR part 86 to part 1036. As part of this process, we clarified our nomenclature and
no longer refer to "otto-cycle" engines; instead, these engines are more accurately labeled spark-
ignition engines throughout part 1036. This section provides additional details related to the test
procedures in Section III.D. of the preamble to this rulemaking. Refer to the preamble for
direction on the final procedures.
2.3.2.1 HD SI FTP
As part of our migration to part 1036, the FTP duty cycle maintains the weighting factors for
the duty cycle speed values from the current HDOE FTP duty cycle that applies to criteria
pollutant regulation in 40 CFR part 86, Appendix 1(f)(1). We changed to the negative torque
values, as noted below. The HD Technical Amendments that were published in June 2021
finalized the migration of some heavy-duty highway engine standard setting part test procedures
from 40 CFR part 86 to part 1036. This included the migration of the HDOE FTP drive schedule
to 40 CFR part 1036, appendix B(b) in order to add vehicle speed and road grade to the duty-
cycle to facilitate powertrain testing of hybrid powertrains for compliance with the HD Phase 2
GHG standards.
We are finalizing in this rule changes that better align certification to GHG standards with
criteria pollutant testing; specifically, the removal of and footnoting of the negative normalized
vehicle torque values over the HDOE FTP duty-cycle. The footnote denotes that these torque
points are controlled using closed throttle motoring, which matches how negative torque values
have been controlled in the HDDE FTP. This change reflects the way that engine manufacturers
were already controlling to negative torque from spark-ignition engines and harmonizes the
methodology with the HDDE FTP, with no effect on stringency.
The spark-ignition engine denormalization equation in 40 CFR 86.1333(a)(l)(ii) contains a
divide by 100 which equates it to the denormalization equation in 40 CFR 1065.610(c)(1)
(equation 1065.610-3); thus the elimination of the 40 CFR part 86 equation from the standard
setting part will have no consequence.
2.3.2.2 Engine mapping
We are finalizing a change to the procedure for SI engine torque mapping in 40 CFR part
1065.510. In order to determine the torque level at any engine speed, a mapping of the engine is
performed prior to the actual FTP testing. The mapping is a sweep across the mechanically- or
electronically-allowed operating speeds to determine the highest possible torque level the engine
can produce at any specific speed. From this "maximum torque" sweep, the FTP targets are
determined for the subsequent transient FTP test and for any additional test (e.g., SET).
The measured torque values are intended to represent the maximum torque the engine can
achieve under fully warmed-up operation with the appropriate design fuel grade (i.e., regular
grade octane fuel) across the allowed range of engine speeds from idle to the typically
electronically limited highest RPM. The intent is to reflect a torque value that is maintained if
the engine is stabilized at a specific speed over a longer period of time such as what might be
observed in a "real world" condition when an engine is held by a transmission to a specific RPM
by the selected gear.
94
-------
Variable torque levels have been observed in engine testing related to such things as
electronic control response to fuel octane, anticipated exhaust thermal conditions, transmission
torque limiting models, and other electronic features incorporated into the engine management
strategies. This torque variability has been particularly evident with the change to the mapping
test procedures in 40 CFR part 1065.510(b)(5)(i), which allowed an engine to be mapped by
performing a transient sweep from idle to maximum rated speed. Prior to this change in the test
procedure, the mapping procedure required the engine to stabilize at discrete engine speed
breakpoints before recording the engine torque value for that engine speed that is used for the
FTP and other testing.
We are finalizing the requirement that the engine achieve a stabilized torque reading at
different speeds prior to recording the final torque values. This is accomplished by disabling any
controls that limit or reduce torque during the engine mapping test.
2.3.2.3 Supplemental Emissions Test for HD SI
As noted in Section 2.1.1.2, the compression-ignition engines currently comply with SET-
based standards that represent high-speed and high-load operation. The SET duty cycle is a
ramped modal cycle in which the engine is tested on an engine dynamometer over a sequence of
steady-state modes. As we show in Section 4.2.3.2, there are opportunities to reduce emissions in
high-load operating conditions where engines often experience enrichment for either catalyst
protection or a power boost. We are finalizing SET-based standards for HD SI engines to ensure
that emission controls are properly functioning in the high load conditions covered by that duty
cycle. We are finalizing the same Cl-based SET procedure, summarized in Section 2.1.2.2, for
HD SI engines, including the existing weighting factors shown in Table 2-2.
2.3.2.4 Onboard Refueling Vapor Recovery
The current ORVR test procedure, which can be found in 40 CFR part 1066, subpart J, for
measuring emissions from chassis-certified vehicles during a refueling event, requires that the
testing occur in a sealed housing evaporative determination (SHED) enclosure containing the
complete vehicle. This procedure applies to all light-duty and heavy-duty complete vehicles
subject to the ORVR standards, and manufacturers designed and built the SHEDs at their test
facilities for these vehicles.
During a recent test program, EPA discovered that very few SHEDs are available that could
fit vehicles in the over-14,000 lb GVWR class because of a combination of the length and height
of these work vehicles. Additionally, the limited large volume SHEDs that were available at
third-party laboratories proved to have challenges measuring the refueling emissions because of
the very large volume inside the enclosures.16
Large background volumes of ambient air create a challenge for evaporative emissions testing
because a measurement is only considered representative if emissions are able to reach a
homogeneous distribution throughout the cell prior to initiating a measurement by the emission
analyzers. In EPA's test program, we found that the two heavy-duty test vehicles required a
substantially longer mixing time than the current test procedure developed for light-duty
vehicles.16'17 Another challenge to adapting the existing procedures for larger vehicles is that the
calculations must account for the volume displaced in the SHED by the test vehicle, which can
be highly variable for the range of commercial vehicle designs. If a manufacturer opts for
95
-------
performing a full SHED test for ORVR certification, we considered adjusting the duration of the
test to achieve a representative emissions distribution in the larger SHED, as well as adjustments
to calculating the displaced volume of these diverse vehicle designs in order to get an accurate
measurement of the refueling emissions.1617
Figure 2-26 and Figure 2-27 show examples of the estimated extrapolated mixing time for the
two trucks that were tested in the summer of 2018. The data does not show how much additional
mixing time is necessary to achieve stabilization. Extrapolated test results from this test program
suggest that at least three additional minutes of mixing time would be needed.
Ford - Estimated Additional Time Required for Mixing
1 Art
Thf« addfrrtftll
160
140
AcUi>l End of Ttsi
i vi Oati iAllfUT anMr 11 1 A .vr^rrvrl
minutes
One additional ^ \
mi nut# >*'
~
J*
minute
X '
1 \
UU ni I CI IV?
£
£ icq
B
<3 /
U _ Jr
¦* uri /
MOVES models
3.34 l/g*
a< 2.4b extra
mlmisw —
(Itftgrams)
_I_ W yf
40
10
OttflO 01:00 02:00 Q&OO 04:00 05:00
Time mfout«s:secon
06*10 07:00
ds
09:00
09:00
Figure 2-26: Estimated Projected Ford E-450 ORVR Results based on Extrapolation
96
-------
Isuzu - Estimated Additional Time Required for Mixing
i an
Refueling ends
140
MDYbS models
3.3d g/gat
One Additional
120
£ 100
£*
O
fiO
at flictra
secondi-——k
(R5 gram^J
minuLc-
\
Actual End of Test
z,B3g/eai
[72 grams)
X
¥
**
:Dtl
ClfttlD 151:1X1 D7:fXl ClfcflD CMsCID IM
Time mlnutcs:scconc)s
Figure 2-27 Estimated Projected Isuzu ORVR Results based on Extrapolation
Since there is limited availability of large volume SHED equipment, we considered the
benefits and challenges of other options to demonstrate compliance with refueling standards for
these large commercial vehicles. One option we considered was testing the complete ORVR
system with all of the components (fuel tank, filler pipe, canister, control valves, etc.)
independent of the actual vehicle using an existing SHED designed for smaller vehicles, such as
light-duty applications. These existing SHED enclosures are widely available to test and certify
other vehicles that are currently subject to refueling standards. This approach of only testing the
components associated with refueling emissions would remove the challenge of finding a SHED
with sufficient dimensions to contain the vastly larger (i.e., longer and taller) commercial
vehicles that are part of this final rule. Testing the refueling related components independent of
the vehicle would also eliminate the challenge of minimizing other hydrocarbon sources not
associated with fuel or the fuel system (i.e., tires, plastics, paints, etc).
Another option we considered was allowing independent ORVR hardware described above to
be tested in a small enclosure designed for only component specific testing (i.e., mini-SHED)
similar to the methodology for the "rig" test, allowed for other evaporative testing by California
and accepted by EPA. Similar to the ORVR system-based concept described previously, the
mini-SHED approach would provide a simpler test methodology that would capture just the
refueling-related emissions. Furthermore, this smaller scale, component-based test would
eliminate much of the variability encountered when attempting to test a full vehicle or hardware
in a large SHED.
For both system- and component-based ORVR testing, the canister would require a specific
conditioning cycle to get it into a representative state of a real-world loading entering a refueling
event. This could be performed on or correlated to an actual vehicle driven over an existing EPA
97
-------
test cycle or a "real world" drive cycle. The current preparatory cycle used by today's ORVR-
required vehicles is designed for light-duty vehicle driving patterns and vehicles with typically
much smaller fuel tanks and canisters. The current conditioning procedure is designed to
challenge the purge system in scenarios such as heavy traffic, slow speeds and start-stop events.
Heavy-duty vehicles, with larger fuel tanks and canisters, may drive more miles and have greater
power demands that may help purge the larger canisters more easily than allowed for in the
current light duty vehicle test. Commercial vehicles may still drive under heavy traffic scenarios
but the expectation is that they will drive more miles daily and operate under higher loads on a
regular basis which help to purge larger amount of vapors from the system.
Another option we considered was allowing engineering analysis in lieu of testing, similar to
what is done for evaporative emission standards today. The current 2-day and 3-day evaporative
standards program for vehicles greater than 14,000 lb GVWR acknowledges that the lighter
classes (i.e., under 14,000 lb GVWR) often use the same hardware-and purge-related
calibrations. Because of this similarity between the classes and a high degree of consistent
performance, the agency currently allows the data and testing from the lighter vehicles to be
accepted for the certification compliance demonstration of the larger class of vehicles. We
expect manufacturers will apply ORVR systems similar to their lighter vehicle classes on the
larger vehicles as a result of this rule. In these cases, an engineering analysis could provide a
similar level of emission control assurance if applied as a means to demonstrate compliance with
new refueling standards for the larger vehicles.
2.3.2.5 Idle Test Procedures Considered
It is important to ensure that the main component of the emission control system, the catalyst,
remains effective during prolonged idle situations. The current heavy-duty FTP test does not
include extended idle conditions indicative of real-world behaviors when work vehicles are
started and allowed to idle while warming-up the engine or when the vehicle operator requires
the interior temperature to stabilize with either heat or air conditioning. This prolonged idle can
also occur when the vehicle is brought to a stop for unloading or a similar situation.
We considered the addition of a new test procedure to ensure emission controls are
maintained during idle. We considered the FTP or SET run as a pre-conditioning cycle to
stabilize the engine and emission control system, followed by 10-30 minutes of idle. Previous
idle provisions, established for in-use inspection and maintenance programs, required a 0.50
percent by volume CO limit over a 30-minute idle period, but did not set a limit for NMHC or
NOx.18
We also considered options that take advantage the existing SET duty cycle to avoid
introducing a new test procedure. One option was to reevaluate the weighting factors of the SET
to place a greater emphasis on the idle modes, but this option has two drawbacks. First,
providing more weight to the idle mode deemphasizes the high-load modes that we believe are
critical to encouraging reduced fuel enrichment. Second, the existing SET procedure collects the
emissions from all modes in a single bag for analysis, which means idle performance would be
masked by the composite result and the additional weighting at idle would require us to
reevaluate the SET standard feasibility. Another option was to require two bags for SET such
that the idle modes would be collected in the second bag. This option would isolate the idle
results but would require two separate standards. While this option reduced the need for a new
98
-------
test procedure, manufacturers would still likely need to make adjustments to their test cells to
accommodate the second bag.
Finally, we considered ensuring catalyst operation at idle by driving manufacturers to
maintain an effective catalyst bed operating temperature that promotes appropriate idle emission
controls, particularly during real world conditions that include prolonged idling typically
followed by a driveaway. Current technology shows that 350 °C is a typical low limit for
effective catalyst operation. If manufacturers control the engine operation such that the catalyst
bed temperature is maintained above this light-off temperature, proper emissions control during
idle conditions and drive-away events is expected to be achieved.
2.4 Useful Life
In addition to emission standards and test procedures, appropriate regulatory useful life
periods are critical to assure emission performance of heavy-duty highway engines. Our
regulations require manufacturers to perform durability testing to demonstrate that engines will
meet emission standards not only at certification but also over the full useful life periods
specified by EPA. Useful life represents the period over which emission standards apply for
certified engines, and, practically, any difference between the regulatory useful life and the
generally longer operational life of in-use engines represents miles and years of operation
without an assurance that emission standards will continue to be met. In this section, we present
a summary of the history of our regulatory useful life provisions and describe our estimates of
the length of operational lives of heavy-duty highway engines, which are almost double the
current useful life mileages in EPA's regulations for all primary intended service classes.
2.4.1 History of Regulatory Useful Life
The Clean Air Act specifies that emission standards under section 202(a) "shall be applicable
to such vehicles and engines for their useful life ... whether such vehicles and engines are
designed as complete systems or incorporate devices to prevent or control such pollution."
Practically, this means that to receive an EPA certificate of conformity under the CAA, a
manufacturer must demonstrate that an engine or vehicle, including the aftertreatment system,
will meet each applicable emission standard, including accounting for deterioration, over the
useful life period specified in EPA's regulations. In addition, CAA section 207(c) requires
manufacturers to recall and repair vehicles or engines if the Administrator determines that "a
substantial number of any class or category of vehicles or engines, although properly maintained
and used, do not conform to the regulations prescribed under [section 202(a)], when in actual use
throughout their useful life (as determined under [section 202(d)])." Taken together, these
sections define two critical aspects of regulatory useful life: (1) the period over which the
manufacturer must demonstrate compliance with emissions standards to obtain EPA
certification, and (2) the period for which the manufacturer is subject to in-use emissions
compliance liability, e.g., for purposes of recall. Manufacturers perform durability testing to
demonstrate that engines will meet emission standards over the full useful life. Manufacturers
may perform scheduled maintenance on their test engines only as specified in the owner's
manual and consistent with our maintenance regulations. As part of the certification process,
EPA approves such scheduled maintenance, which is also subject to minimum maintenance
intervals as described in the regulation.
99
-------
EPA prescribes regulations under CAA section 202(d) for determining the useful life of
vehicles and engines. CAA section 202(d) provides that the minimum useful life for heavy-duty
vehicles and engines is a period of 10 years or 100,000 miles, whichever occurs first. This
section authorizes EPA to adopt longer useful life periods that we determine to be appropriate.
Under this authority, we established useful life periods for heavy-duty engines by primary
intended service class. Heavy-duty highway engine manufacturers identify the primary intended
service class for each engine family by considering the vehicles for which they design and
market their engines.F Heavy-duty compression-ignition engines are distinguished by their
potential for rebuild and the weight class of the final vehicles in which the engines are expected
to be installed.0 Heavy-duty spark-ignition engines are generally classified as a single "spark-
ignition" service class unless they are designed or intended for use in the largest heavy-duty
vehicles and are thereby considered heavy heavy-duty engines.11 The following useful life
periods currently apply to the criteria pollutant emission standards for heavy-duty highway
engines:u
• 110,000 miles or 10 years for heavy-duty spark-ignition engines and light heavy-duty
compression-ignition engines
• 185,000 miles or 10 years for medium heavy-duty compression-ignition engines
• 435,000 miles, 10 years, or 22,000 hours for heavy heavy-duty compression-ignition
engines
In our 1983 rulemaking, which first established class-specific useful life values for heavy-
duty engines and vehicles, EPA adopted the principle that useful life mileage values should
reflect the typical mileage to the first rebuild of the engine (or scrappage of the engine if that
occurs without rebuilding).19 Significantly, this approach was adopted at a time when diesel
engine emission control strategies relied mainly on in-cylinder engine combustion controls.
Over time, mileage values became the primary metric for useful life duration. This is
because, due to advancements in general engine durability, nearly all heavy-duty engines reach
the mileage value in-use long before 10 years have elapsed. The age (years) value has meaning
for only a small number of low-annual-mileage applications, such as refuse trucks. Also,
manufacturer durability demonstrations generally target the mileage values, since deterioration is
a function of engine work and hours rather than years in service and a time-based demonstration
would be significantly longer in duration than one based on applicable mileage value.
F See 40 CFR 1036.140 as referenced in the definition of "primary intended service class" in 40 CFR 86.090-2.
G As specified in the current 40 CFR 1036.140(a), light heavy-duty engines are not designed for rebuild and are
normally installed in vehicles at or below 19,5000 pounds GVWR; medium heavy-duty engines may be designed for
rebuild and are normally installed in vehicles from 19,501 to 33,000 lbs GVWR; heavy heavy-duty engines are
designed for multiple rebuilds and are normally installed in vehicles above 33,000 pounds GVWR.
H See 40 CFR 1036.140(b).
1 See 40 CFR 86.004-2. EPA adopted useful life values of 110,000, 185,000, and 290,000 miles for light, medium,
and heavy heavy-duty engines, respectively, in 1983 (48 FR 52170, November 16, 1983). The useful life for heavy
heavy-duty engines was subsequently increased to 435,000 miles for 2004 and later model years (62 FR 54694,
October 21, 1997).
1 The same useful life periods apply for heavy-duty engines certifying to the greenhouse gas emission standards,
except that the spark-ignition standards and the standards for model year 2021 and later light heavy-duty engines
apply over a useful life of 15 years or 150,000 miles, whichever comes first. See 40 CFR 1036.108(d).
100
-------
In the 1997 rulemaking that most recently increased heavy-duty engine useful life, EPA
included an hours-based useful life of 22,000 hours for the heavy heavy-duty engine intended
service class. This unique criterion was added to address the concern that urban vehicles,
particularly urban buses, equipped with heavy heavy-duty engines had distinctly different driving
patterns compared to the line-haul trucks from which the agency based its useful life value of
435,000 miles.20 Commenters in that rulemaking indicated that urban bus average speed was
near 13 miles per hour. Considering that speed, many of these bus engines would reach the end
of their operational life or be candidates for rebuild before the applicable mileage value or the
10-year criterion is reached. The 22,000 hours value was adopted in lieu of a proposed minimum
useful life value of 290,000 miles for heavy heavy-duty engines. Considering the current 435,000
useful life mileage for heavy heavy-duty engines, the 22,000-hour useful life value only has
significance for the small subset of vehicles equipped with heavy heavy-duty engines with an
average speed of less than 20 miles per hour.
In the Phase 1 GHG rulemaking, we promulgated useful life periods for the GHG emission
standards for heavy-duty highway engines and their corresponding heavy-duty vehicles that
aligned with the current useful life periods for criteria pollutant emission standards.21 In the HD
Phase 2 GHG rulemaking, we extended the useful life for Light HDV, light heavy-duty engines,
and spark-ignition engines for the GHG emission standards to 15 years or 150,000 miles to align
with the useful life of chassis-certified heavy-duty vehicles subject to the Tier 3 standards.22 See
40 CFR 1036.108 and 40 CFRpart 1037, subpartB, for the GHG useful life periods that apply
for heavy-duty highway engines and vehicles, respectively.
2.4.2 Identifying Appropriate Useful Life Periods
Emission standards apply for the engine's useful life and manufacturers must demonstrate the
durability of engines to maintain certified emission performance over their useful life. Thus, the
emission standards must be considered together with their associated useful life periods. Larger
useful life mileage values require manufacturers to demonstrate emission performance over a
longer period and should result in effective emission control over a greater proportion of an
engine's operational (sometimes referred to as "service") life. Consistent with our approach to
adopting useful life mileages in the 1983 rulemaking, we continue to consider a comprehensive
out-of-frame rebuild to represent the end of a heavy-duty CI engine's "first life" of operation. For
SI engines that are less commonly rebuilt, engine replacement is a more appropriate measure of
an engine's operational life.
2.4.2.1 Compression-Ignition Engine Rebuild Data
In 2013, EPA commissioned an industry characterization report on heavy-duty diesel engine
rebuilds.23 The report relied on existing data from MacKay & Company surveys of heavy-duty
vehicle operators. In this report, an engine rebuild was categorized as either an in-frame overhaul
(where the rebuild occurred while the engine remained in the vehicle) or an out-of-frame
overhaul (where the engine was removed from the vehicle for more extensive service).K The
study showed that the mileage varied depending on the type of rebuild. Rebuilding an engine
K Note that these mileage values reflect replacement of engine components, but do not include aftertreatment
components. At the time of the report, the population of engines equipped with DPF and SCR technologies was
limited to relatively new engines that were not candidates for rebuild.
101
-------
while the block remained in the frame was typically done at lower mileage than rebuilding an
engine removed from the vehicle. The results of the study by vehicle weight class are presented
in Table 2-10.
Table 2-10 Average Mileage and Age at First Rebuild for Heavy-Duty CI Engines From 2013 EPA Rebuild
Industry Characterization Report
Vehicle Weight Class
In-Frame Rebuild
Out-of-Frame Rebuild
Mileage
Years
Mileage
Years
Class 3
216,900
9.5
256,000
9.5
Class 4
236,800
11.6
346,300
10.3
Class 5
298,300
10.9
344,200
11.9
Class 6
332,200
13.0
407,700
10.6
Class 7
427,500
15.8
509,100
13.2
Class 8
680,200
11.9
909,900
8.9
McKay & Company does not collect information on aftertreatment systems (e.g., diesel
oxidation catalysts, SCR systems, or three-way catalysts), so neither EPA's 2013 report nor
CARB's more recent report for their HD Omnibus rulemaking include aftertreatment system age
information.1^ We consider the mileage at rebuild or replacement of an engine to represent the
operational life of that engine, including any aftertreatment components that were part of its
original certified configuration. We have no data to indicate aftertreatment systems lose
functionality before engines are rebuilt or replaced, and our technology demonstrations Chapter
3 show aftertreatment catalysts are able to maintain performance when bench-aged to beyond the
equivalent of current useful life mileages.
We averaged the mileages for these vehicle classes according to EPA's primary intended
service classes for heavy-duty CI engines as defined in 40 CFR 1036.140. Specifically, we
averaged Classes 3, 4, and 5 to represent Light HDE, Classes 6 and 7 to represent Medium HDE,
and Class 8 to represent Heavy HDE. These values are shown in Table 2-11 with the current
useful life mileages that apply to each intended service class. As seen in the tables, the study
reported typical engine rebuild mileages that are more than double the current useful life
mileages for those classes.
L See Chapter 2.5.2.3 for a summary of the CARB report that reflects engine rebuilds and replacements occurring
between calendar years 2012 and 2018.
102
-------
Table 2-11 Average Mileage at First Rebuild for Heavy-Duty CI Engines Based on EPA Intended Service
Classes
Primary Intended
Service Class
Mileage at First
In-Frame Rebuild
Mileage at First
Out-of-Frame Rebuild
Current Useful
Life Mileage
Light HDE
(Vehicle Classes 3-5)
250,667
315,500
110,000a
Medium HDE
(Vehicle Classes 6-7)
379,850
458,400
185,000
Heavy HDE
(Vehicle Class 8)
680,200
909,900
435,000
a The useful life mileage that applies for Light HDE for GHG emission standards is 150,000 miles. See existing 40
CFR 1036.108(d).
We note that Light HDE intended for smaller vehicle classes are not designed with cylinder
liners to facilitate rebuilding, suggesting they are more likely to be scrapped at the end of their
operational life. The rebuilding report notes that seven percent of the diesel-fueled engines in the
2012 Class 3 vehicle population were removed from the vehicle to be rebuilt, but does not
include data on the corresponding number of engines or vehicles scrapped in that year. We
assume the mileage at which an engine has deteriorated enough to consider rebuilding would be
similar to the mileage of an engine eligible for scrappage and both similarly represent the
operational life of an engine for the purpose of this analysis.
2.4.2.2 Spark-Ignition Engine Service Life Data
The current useful life mileage that applies for GHG emission standards for Spark-ignition
HDE is 150,000 miles, which is longer than the current useful life of 110,000 miles for criteria
pollutant emission standards for those same engines.M For our updates to the useful life for
Spark-ignition HDE criteria pollutant emission standards, we considered available data to
represent the operational life of recent heavy-duty SI engines. A review of market literature for
heavy-duty gasoline engines showed that at least one line of engine-certified products is
advertised with a service life of 200,000 miles.24 Compliance data for MY 2019 indicate that the
advertised engine model represents 20 percent of the Spark-ignition HDE certified for MY 2019.
Additionally, CARB's HD Omnibus rulemaking cited heavy-duty Otto-cycle engines (i.e., Spark-
ignition HDE) for vehicles above 14,000 lb GVWR that were replaced during calendar years
2012 through 2018 as reaching more than 217,000 miles on average.25 The mileages in these two
examples are almost double the current useful life for Spark-ignition HDE, indicating many
miles of operation beyond the current useful life.
2.4.2.3 CARB HD Omnibus Useful Life Values
The CARB HD Omnibus rulemaking, finalized in August 2020, lengthens useful life for
heavy-duty CI and SI engines in two steps.N'26 As part of their rule, CARB analyzed recent
MacKay & Company survey data from calendar years 2012 through 2018 and reported rebuild
mileages for CI engine categories that were similar to those described in the Chapter 2.4.2.1.
M See 40 CFR 1036.108(d) for the GHG useful life, and the definition of "useful life" in 40 CFR 86.004-2 for the
criteria pollutant useful life.
N EPA is reviewing a waiver request under CAA section 209(b) from California for the Omnibus rule. See 87 FR
35765 (June 13, 2022).
103
-------
CARB also included average replacement mileage information for heavy-duty Otto-cycle (HD
SI) engines.27 The CARB/MacKay & Company data is summarized in Table 2-12.
Table 2-12 Summary of CARB/MacKay & Company engine rebuild and replacement mileages for the HD
Omnibus regulation a
Engine Class
Average Mileage at
Rebuild or Replacement
HD Otto (Spark-ignition HDE)
(All Vehicle Classes above 14,000 lb GVWR)
217,283
LHDD (Light HDE)
(Vehicle Classes 4-5)
326,444
MHDD (Medium HDE)
(Vehicle Classes 6-7)
432,652
HHDD (Heavy HDE)
(Vehicle Class 8)
854,616
a CARB's naming conventions for HD engines differ from EPA; corresponding EPA
names are noted in parentheses
Although the CARB HD Omnibus program set standards for MY 2024, the program
maintained the current useful life values through MY 2026. Table 2-13 summarizes the useful
life values finalized in the HD Omnibus rule for heavy-duty Otto-cycle engines (HDO), and light
heavy-duty diesel (LHDD), medium heavy-duty diesel (MHDD), and heavy heavy-duty diesel
(HHDD) engines.
Table 2-13 CARB useful life mileages for heavy-duty engines in the HD Omnibus rulemakinga
Model Year
HDO
LHDD
MHDD
HHDD
(Spark-ignition HDE)
(Light HDE)
(Medium HDE)
(Heavy HDE)b
2024-2026
110,000 miles
110,000 miles
185,000 miles
435,000 miles
10 years
22,000 hours
10 years
10 years
10 years
2027-2030
155,000 miles
190,000 miles
270,000 miles
600,000 miles
11 years
30,000 hours
12 years
12 years
11 years
2031 and later
200,000 miles
15 years
270,000 miles
15 years
350,000 miles
12 years
800,000 miles
12 years
40,000 hours
a CARB's naming conventions for HD engines differ from EPA; corresponding EPA names are noted in
parentheses.
b CARB adopted an intermediate useful life mileage of 435,000 miles for MY 2027 and later HHDD engines.
As seen in the table, CARB's Omnibus increased useful life first in MY 2027 with a second
step in MY 2031. The final useful life mileages in the CARB regulation are the result of
stakeholder engagement throughout the development of CARB's HD Omnibus rulemaking. In
two 2019 public workshops, CARB staff presented useful life mileage values under
consideration that were longer than these final mileages and, in their September 2019
presentation, very close to the engine rebuild mileages.28 In response to feedback from
stakeholders indicating concerns with availability of data for engines and emission controls at
104
-------
those mileages, CARB shortened their final useful life mileages for MY 2031 and later engines
from the values presented in 2019, and the MY 2027 values were chosen to be approximately the
mid-point between the current and final useful life mileages.29 Additionally, CARB finalized an
intermediate useful life mileage for MY 2027 and later HHDD engines that correspond to the
current useful life of 435,000 miles. Consistent with current useful life periods, CARB finalized
hours values for the HHDD engine class based on the useful life mileage and an average vehicle
speed of 20 miles per hour.
Similar to the useful life mileage values, CARB's useful life values in years were also
adjusted from the values presented in their public workshops based on stakeholder feedback. In
particular, emission controls manufacturers recommended CARB consider replacing the 18-year
useful life presented in their September 2019 workshop with a useful life of 12 years for heavy-
duty engines.30 CARB agreed that 12 years was reasonable for MHDD and HHDD, but adopted
a 15 year useful life for HDO and LHDD based on the useful life in years that applies to chassis-
certified engines.
Chapter 2 References
1 81 FR 73478, October 25, 2016.
2 See 40 CFR 1036.630 and 1037.550.
3 86 FR 34321, June 29, 2021.
4 "OEM perspective - Meeting EPA/NHTSA GHG/Efficiency Standards", 7th Integer
Emissions Summit USA 2014, Volvo Group North America.
5 Duran, A., Kotz, A., Thorton, M., and Kelly, K. NREL Low Load Engine Test Cycle
Development Update Presentation. February 16,2018.
6 Heavy-Duty Low NOx Program Workshop - Low Load Cycle Development Presentation.
January 23, 2019 https://www.arb.ca.gov/msprog/hdlownox/hdlownox.htm
7 Heavy-Duty Low NOx Program Workshop - Low Load Cycle Presentation. September 26,
2019 https://www.arb.ca.gov/msprog/hdlownox/hdlownox.htm
8 Sharp, C. Heavy-Duty Engine Low-Load Emission Control Calibration, Low-Load Test Cycle
Development, and Evaluation of Engine Broadcast Torque and Fueling Accuracy During Low-
Load Operation Low-Load Operation Low NOx Demonstration Program - Stage 2, prepared for
the California Air Resources Board, December 20, 2019.
9 Pondicherry, R., Besch, M., Thiruvengadam, A., and Carder, A. A Vehicle Activity-based
Windowing approach to evaluate real-world NOx emissions from Modern Heavy-duty Diesel
Trucks, Atmospheric Environment, Vol. 247, 2021, 118169,
http://www.sciencedirect.com/science/article/pii/S1352231020308992
105
-------
10 Hamady, Fakhri, Duncan, Alan. "A Comprehensive Study of Manufacturers In-Use Testing
Data Collected from Heavy-Duty Diesel Engines Using Portable Emissions Measurement
System (PEMS)". 29th CRC Real World Emissions Workshop, March 10 -13, 2019.
11 Sandhu, Gurdas; Sonntag, Darrell; Sanchez, James. 2018. Identifying Areas of HighNOx
Operation in Heavy-Duty Vehicles, 28th CRC Real-World Emissions Workshop, March 18-21,
2018, Garden Grove, California, USA
12 Sonntag, Darrell. Exhaust Emission Rates for Heavy-Duty Onroad Vehicles in
MOVESCTINPRM. Attachment to Memorandum to Docket EPA-HQ-OAR-2019-0055:
"Updates to MOVES for Emissions Analysis of the HD 2027 NPRM." May 2021.
13 COMMISSION REGULATION (EU) No 582/2011, May 25, 2011. Available online:
https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:02011R0582-
20180118&from=EN.
14 COMMISSION REGULATION (EU) 2018/932, June 29, 2018. Available online: https://eur-
lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:32018R0932&from=EN.
15 Rodriguez, F.; Posada, F. "Future Heavy-Duty Emission Standards An Opportunity for
International Harmonization". The International Council on Clean Transportation. November
2019. Available online:
https://theicct.org/sites/default/files/publications/Future%20_HDV_standards_opportunity_2019
1125.pdf
16 SGS-Aurora, Eastern Research Group, "Light Heavy-Duty Gasoline Vehicle Evaporative
Emissions Testing." EPA-420-R-19-017. December 2019.
17 U.S. Environmental Protection Agency. "Summary of "Light Heavy-Duty Gasoline Vehicle
Evaporative Emissions Test Program"" EPA-420-S-19-002. December 2019.
18 40 CFR 86.1829-01(b)(5).
19 U.S. EPA, "Summary and Analysis of Comments on the Notice of Proposed Rulemaking for
Revised Gaseous Emission Regulations for 1984 and Later Model Year Light-Duty Trucks and
Heavy-Duty Engines", July 1983, p 43.
20 U.S. EPA, "Summary and Analysis of Comments: Control of Emissions of Air Pollution from
Highway Heavy-Duty Engines", EPA-420-R-97-102, September 1997, pp 43-47.
21 76 FR 57181, September 15,2011.
22 See 79 FR 23414, April 28, 2014 and 81 FR 73496, October 25, 2016.
23 ICF International, "Industry Characterization of Heavy Duty Diesel Engine Rebuilds" EPA
Contract No. EP-C-12-011, September 2013.
24 See, e.g., Isuzu Truck webpage. "Isuzu Commercial Vehicles: N-Series Gas Trucks."
Available online: www.isuzucv.com/en/nseries/nseries_gas. Accessed February 28, 2020.
106
-------
25 California Air Resources Board / MacKay & Company, "CARB Summary Report on the
Analysis of the MacKay & Company Data on Heavy-Duty Engine Rebuilds and Replacements",
March 2019.
26 California Air Resources Board. Heavy-Duty Omnibus Regulation. Available online:
https://ww2.arb.ca.gov/rulemaking/2020/hdomnibuslownox.
27 California Air Resources Board / MacKay & Company, "CARB Summary Report on the
Analysis of the MacKay & Company Data on Heavy-Duty Engine Rebuilds and Replacements",
March 2019.
28 Brakora, Jessica. Memorandum to Docket: EPA-HQ-OAR-2019-005 5. CARB 2019 Public
Workshop Presentations Related to Regulatory Useful Life and Emissions Warranty. March 19,
2021.
29 California Air Resources Board. Staff Report: Initial Statement of Reasons-Public Hearing to
Consider the Proposed Heavy-Duty Engine and Vehicle Omnibus Regulation and Associated
Amendments. June 23, 2020. Page 111-57.
30 Manufacturers of Emission Controls Association. "Preliminary Suggestions for Future
Warranty and FUL Requirements." Presentation to CARB. September 5, 2019.
107
-------
Chapter 3 Feasibility Analysis for the Final Standards
3.1 Compression-Ignition Technology Feasibility
3.1.1 Diesel Technology Demonstration Programs
3.1.1.1 CARB Heavy-duty Low NOx Stage 3 Research Program
In 2016, the California Air Resources Board (CARB) funded the CARB Heavy-duty Low
NOx Research Program at Southwest Research Institute (SwRI) in San Antonio, TX to explore
the feasibility of diesel HDEs achieving 0.02 g/bhp-hr NOx composite emissions over the FTP
transient test cycle. Stage 3 of this research program investigated the use of dual-SCR systems,
cylinder deactivation (CDA), heated urea dosing system, intercooler bypass, turbine bypass, and
EGR cooler bypass via engine dynamometer testing of a developmental engine based on the MY
2017 Cummins X15 diesel Heavy HDE. Major specifications for the engine are shown in Table
3-1. Like many other MY 2010 and later diesel HDEs, the X15 is equipped with a variable
geometry turbocharger (VGT), high pressure common rail fuel injection and cooled EGR. The
X15's original equipment exhaust aftertreatment system (EAS) consists of a DOC, DPF, SCR,
and ASC in series. The XI5 engine was modified by SwRI and Eaton to incorporate individual
cylinder deactivation. SwRI also developed an engine calibration to aid catalyst warmup using a
combination of later combustion phasing and increased idle speed. Further details of the
specific, fixed CDA modes evaluated and other details regarding CDA development and engine
instrumentation can be found in two papers by SwRI and Eaton.1>2 Details regarding the EAS,
control systems, and calibration are summarized in three additional papers by SwRI.3'4'5 The
complete summary of the work completed as part of Stage 3 is included in a final report from
SwRI to CARB.6
Table 3-1: Major engine specifications for the MY 2017 Cummins X15 engine used for the CARB Low NOx
Stage 3 Research Program
Engine Displacement
14.95 L
Bore X Stroke
137 X 169 mm
Rated Power (a)y Speed
373 kW @ 1800 rpm
Rated Torque @ Speed
2500 N-m @ 1000 rpm
108
-------
Figure 3-1: Developmental Cummins X15 Engine equipped with individual cylinder deactivation undergoing
engine dynamometer testing as part of the CARB Stage 3 research at SwRI.
A schematic layout of the developmental EAS with light-off SCR is shown in Figure 3-2.
Photos of the EAS showing details of its installation within the engine dynamometer test cell are
shown in Figure 3-3. The heated urea dosing system, mixer, and light-off SCR with ASC zone-
coated onto the rear of the SCR substrate were mounted downstream of the turbocharger. The
remaining components were mounted in an insulated single-box housing to improve heat
retention. Catalyst substrate specifications are summarized in Table 3-2. The total EAS volume
was approximately 4.6 times the engine displacement. The total volume of the SCR, not
including ASC, was approximately 2.8 times engine displacement. The volume of the light-off
SCR was approximately 0.58 times engine displacement and the volume of the downstream SCR
system was approximately 2.2 times engine displacement (both excluding ASC). The volume of
the downstream SCR is comparable to the sales-weighted displacement-specific SCR volume for
MY 2019 Fleavy FIDE, so the increase in total SCR volume relative to MY 2019 Heavy FIDE
applications was due to the addition of light-off SCR.
109
-------
4^
13x6
; DEF Dosing
HC Dosing (7th
Injector)
: NH,Sensor
= Heated DEF Dosing
= Temp Sensor
Insulated Single-box System
SC
O
<0
I
O
2-10.5x4
tc
DC
K
o
—
o
O
o
w
W
V)
m
<
cc
OS
DC
o
1—
o
o
O
V)
w 1
«
w
<
Internal "Switchback" Mixing Tube
4-10.5x10
Figure 3-2: Schematic layout (not to scale) of the dual-SCR EAS tested as part of the CARB Stage 3 research
at SwRI.
Heated Doser
Mount
SCR + SCR/ASC
(Dual Path)
Figure 3-3: Developmental EAS with light-off SCR installed in engine dynamometer test cell at SwRI (upper
left, upper right) and details of the downstream, single "box" unit (lower left, lower right).
110
-------
Table 3-2: Summary of catalyst specifications for developmental EAS with light-off SCR
Component
Dimensions,
Dia. X
Length
(inches)
Substrate
Volume (liters)
Cell Density
(cpsi) / wall
thickness (mil)
Notes
Light-off
SCR/ASC
13X6
13.1
400/4
ASC is zone-coated to the
rearmost 2" of the light-off
SCR substrate
Zone-
coated
CDPF
13X7
15.2
300/7
Zone-coated wall-flow
substrate providing both
DOC and CDPF
functionality
SCR
10.5X4
11.4
600/4.5
Two substrates in parallel
SCR
10.5X5
14.2
600/4.5
Two substrates in parallel
SCR/ASC
10.5X5
14.2
600/4.5
Two substrates in parallel,
both with ASC zone-
coated to the rearmost 2"
of the SCR substrates
Emissions were evaluated using engine dynamometer testing over the cold-start and hot-start
FTP transient cycle, the SET, the LLC, and over specific cycles representing real world
operation that were provided by the Engine Manufacturers Association (EMA). A baseline,
original equipment MY 2017 Cummins XI5 EAS was tested at a low-hour condition with the
EAS in a degreened (broken-in) state. The developmental EAS with light-off SCR was tested in
a degreened state and then was subjected to accelerated aging using the Diesel Aftertreatment
Accelerated Aging Cycle (DAAAC).7 The DAAAC incorporated chemical deSOx at 30-hour
intervals and DPF ash maintenance at 500-hour intervals. Emissions results over the FTP, SET,
and LLC for the baseline and developmental EAS are summarized in Table 3-3 through Table
3-7. FTP Composite and SET NOx emissions results were just over 20 mg/bhp-hr after
accelerated aging equivalent to approximately 435,000 miles. The NOx emissions results over
the LLC were just under 50 mg/bhp-hr. Emissions of N2O were approximately half that of the
current 0.10 g/bhp-hr standards. The infrequent regeneration factor (IRAF) for this engine and
EAS configuration was 2 mg/hp-hr for the FTP and SET and 5 mg/hp-hr for the LLC.
Ill
-------
Table 3-3: Baseline (degreened) emissions results for the OE Cummins EAS. Results do not include
adjustments for IRAF.
Cycle
Results:
FTP
cold
95%
CI
FTP
hot
95%
CI
FTP
composite
95%
CI
SET
95%
CI
LLC
95%
CI
NOx
(mg/bhp-hr)
271
31
132
99
152
87
140
6
1005
335
PM
(mg/bhp-hr)
2.0
0.2
2.0
1.0
2.0
0.9
1.2
0.3
NA
NA
NMHC
(mg/bhp-hr)
3
7
2
2
3
3
1.7
0.2
12
25
CO
(mg/bhp-hr)
48
37
17
31
22
29
7.9
0.9
30
24
C02
(g/bhp-hr)
530
4
508
8
511
8
452
4
609
7
N20
(mg/bhp-hr)
42
2
63
9
61
11
68
8
64
NA*
For results where the 95% CI is greater than the average, the results are not statistically different from zero based
on a 2-sided Student's t-test at a=0.05
Table 3-4: 0-hour (degreened) emissions results for the developmental EAS system with light-off SCR
Results do not include adjustments for IRAF or crankcase emissions.
Cycle
Results:
FTP
cold
95%
CI
FTP
hot
95%
CI
FTP
composite
95%
CI
SET
95%
CI
LLC
95%
CI
NOx
(mg/bhp-hr)
54.7
0.8
11
2
17
1
8.9
0.8
20
8
PM
(mg/bhp-hr)
2
1
1.4
0.1
1.5
0.2
0.9
0.2
4.0
0.7
NMHC
(mg/bhp-hr)
23
56
12
50
14
51
1
2
47
171
CO
(mg/bhp-hr)
110
41
12
3
26
6
7
1
62
51
C02
(g/bhp-hr)
539
4
499
1
505
2
454
3
600
4
N20
(mg/bhp-hr)
39
3
47
1
46
2
53
9
43
9
For results where the 95% CI is greater than the average, the results are not statistically different from zero based
on a 2-sided Student's t-test at a=0.05
Table 3-5: Emissions results for the developmental EAS system with light-off SCR after 334 hours of
accelerated thermal and chemical aging using the DAAAC (equivalent to approximately 145,000 miles of
112
-------
operation). The SET (2021) results represent updated 40 CFR 1036.510 SET procedures. Results do not
include adjustments for IRAF or crankcase emissions.
Cycle
Results:
FTP
cold
95%
CI
FTP
hot
95%
CI
FTP
composite
95%
CI
SET
95%
CI
SET
(2021)
95%
CI
LLC
95%
CI
NOx
(mg/bhp-hr)
56
5
12
2
18
2
15
2
15
2
22
9
PM
(mg/bhp-hr)
1.3
0.2
1.6
0.5
1.6
0.4
0.7
0.0
0.7
0.4
3
6
NMHC
(mg/bhp-hr)
33
58
18
42
20
44
10
15
3.0
0.6
19
18
CO
(mg/bhp-hr)
186
94
25
10
48
22
7
1
8
2
104
45
C02
(g/bhp-hr)
541
4
506
5
511
5
454
1
450
2
602
2
N20
(mg/bhp-hr)
31
4
37
4
36
4
34
2
34
1
32
17
For results where the 95% CI is greater than the average, the results are not statistically different from zero based
on a 2-sided Student's t-test at a=0.05
Table 3-6: Emissions results for the developmental EAS system with light-off SCR after 667 hours of
accelerated thermal and chemical aging using the DAAAC (equivalent to approximately 290,000 miles of
operation). The SET (2021) results represent updated 40 CFR 1036.510 SET procedures. Results do not
include adjustments for IRAF or crankcase emissions.
Cycle
Results:
FTP
cold
95%
CI
FTP
hot
95%
CI
FTP
composite
95%
CI
SET
95%
CI
SET
(2021)
95%
CI
LLC
95%
CI
NOx
(mg/bhp-hr)
63
6
15
2
22
3
19
2
14.8
0.4
50
6
PM
(mg/bhp-hr)
1.8
0.4
2.0
1.3
2.0
1.0
1
2
0.8
0.3
4.5
0.5
NMHC
(mg/bhp-hr)
29
12
12
5
14
3
3
2
2.1
0.7
43
5
CO
(mg/bhp-hr)
227
40
139
25
151
17
10
2
11
3
305
42
C02
(g/bhp-hr)
538
1
512
2
515
2
461
2
454.1
0.7
616
1
N20
(mg/bhp-hr)
28
9
32
7
31
7
23
3
25
1
21
6
Table 3-7: Emissions results for the developmental EAS system with light-off SCR after 1000 hours of
accelerated thermal and chemical aging using the DAAAC (equivalent to approximately 435,000 miles of
113
-------
operation). The SET (2021) results represent updated 40 CFR 1036.510 SET procedures. Results do not
include adjustments for IRAF or crankcase emissions.
Cycle
Results:
FTP
cold
95%
CI
FTP
hot
95%
CI
FTP
composite
95%
CI
SET
95%
CI
SET
(2021)
95%
CI
LLC
95%
CI
NOx
(mg/bhp-hr)
61
8
17
3
23
3
22
4
20
4
47
4
PM
(mg/bhp-hr)
2
0
2
1
2
1
2
0
1
0
6
1
NMHC
(mg/bhp-hr)
41
10
24
13
26
12
5
5
4
0
28
77
CO
(mg/bhp-hr)
257
44
184
38
194
34
12
1
11
1
371
79
C02
(g/bhp-hr)
535
9
512
1
515
2
461
6
456
2
617
1
N20
(mg/bhp-hr)
37
7
42
21
41
18
23
2
23
1
18
12
For results where the 95% CI is greater than the average, the results are not statistically different from zero based
on a 2-sided Student's t-test at a=0.05
In addition to evaluating the feasibility of the new criteria pollutant standards, we also
evaluated how CO2 was impacted using the CARB Stage 3 engine for both the test procedures
used to show compliance with the engine standards in 40 CFR 1036.108 and the vehicle
standards in 40 CFR part 1037, subpart B. To do this we evaluated how CO2 emissions changed
from the base engine on the FTP, SET, and LLC, as well as the fuel mapping test procedures
defined in 40 CFR 1036.535 and 1036.540. For all three cycles the Stage 3 engine emitted CO2
with no statistically significant difference at a 95% level of confidence when compared to the
base MY 2017 Cummins XI5 engine. Comparing the CARB Stage 3 engine using the 0-hour
(degreened) EAS provided the most direct comparison with the MY 2017 Cummins XI5 engine
since the MY 2017 Cummins XI5 original equipment (OE) EAS was degreened but not
hydrothermally or chemically aged. The percent reduction in CO2 for the FTP, SET and LLC,
was 1, 0 and 1% respectively for the Stage 3 configuration relative to the OE configuration.
Because SwRI made changes to the thermal management strategies of the CARB Stage 3 engine
(which increased CO2 emissions from the engine) to improve NOx reduction at low SCR
temperatures after these initial data were taken, there is no direct comparison between the
baseline engine and the CARB Stage 3 engine. For the data at an equivalent of 435,000 miles
that include these changes, the percent increase in CO2 for the FTP, SET and LLC, was 0.6, 0.7
and 1.3% respectively, but since the aftertreatment had been aged to an equivalent of 435,000
miles prior to these tests, which included ash exposure from this aging and which thus increased
the back pressure on the engine (Figure 3-4), this was not a direct comparison with the baseline
engine. To evaluate impacts to CO2 emissions of the CARB Stage 3 engine on the HD GHG
Phase 2 test procedure, the test procedures were executed with both the baseline engine and the
CARB Stage 3 engine with development aged aftertreatment. The fuel maps from these tests
were used as inputs for GEM simulations. The results from this analysis (summarized in the
SwRI Stage 3 report)6 also showed that the CARB Stage 3 engine emitted CO2 at approximately
the same rate as the MY 2017 Cummins XI5.
114
-------
12
10
Sa 8
PL,
TO
13
PL,
PL,
Q
• Olii' average
6671ii" average
500 1000 1500
time (sec)
2000
2500
Figure 3-4: The average pressure drop across the DPF on the SET for the degreened aftertreatment and the
equivalent of 290,000 miles of operation aftertreatment
3.1.1.2 EPA Stage 3 Demonstration
Once the CARB Stage 3 Demonstration was completed, the work was continued into a second
phase of the demonstration by EPA and is referred to as the EPA Stage 3 Demonstration. During
the EPA Stage 3 Demonstration, improvements were made to the aftertreatment by replacing the
zone-coated catalyzed soot filter with a separate 13-inch diameter by 4-inch length DOC and 13-
inch diameter by 7-inch length DPF. These components were separately aged via the DAAAC to
the equivalent of 435,000 miles prior to their integration into the EPA Stage 3 EAS. Changes
were also made to the downstream "One-box" system to further improve urea mixing and
distribution. The entire system (LO-SCR, DOC, DPF, SCR, and SCR/ASC) was then aged over
the DAAAC to the equivalent of 800,000 miles. A schematic of the aftertreatment is shown in
Figure 3-5. The results of testing the EPA Stage 3 engine at the equivalent of 435,000, 600,000
and 800,000 miles are shown in Table 3-8, Table 3-9 and Table 3-10. The testing of the EPA
Stage 3 engine also included testing with the crankcase vent not connected to the CVS tunnel to
determine the crankcase emissions. The results from these tests showed that the NOx emissions
from the EPA Stage 3 engine are at a minimum 6 mg/hp-hr for the FTP, SET, and LLC cycles.
By closing the crankcase, the NOx emissions from each of these duty cycles would be reduced to
0 mg/hp-hr. For the feasibility assessment of the standards, we included the emissions
reductions from closing the crankcase. The complete summary of the work completed is
included in a final report from SwRI.8
For our feasibility analysis the 435,000 miles test point was used for assessing the Light HDE
standards since the final useful life for Light HDE is below 435,000 miles, and because we
believe the 435,000 miles test point adequately represents the deterioration of Light HDE to its
final useful life. The interpolated emissions performance at 650,000 miles was used for Medium
and Heavy HDEs because we believe that test point adequately reflects the deterioration of these
engines to each final useful life mileage (350,000 miles and 650,000 miles, respectively). While
the final useful life for Medium HDE is fewer miles than this test point, EPA considered
115
-------
comments with supporting data that showed, due to the real-world operation of these engines,
that Medium HDE experience more hydrothermal aging than a Heavy HDE at the final useful
life values of 350,000 miles for Medium HDE and 650,000 miles for Heavy HDE (up to 4x).
EPA also considered comments with supporting data that showed, regarding Heavy HDE real
world operation, that the aftertreatment of Medium HDE experience approximately 1/3 less
chemical poisoning than Heavy HDE do at the final useful life values of 350,000 miles and
650,000 miles respectively. Thus, the magnitude of these two aging mechanisms (hydrothermal
and chemical poisoning) is different for Medium HDE compared to Heavy HDE. When
considering the aging carried out on the EPA Stage 3 engine, in our assessment, the greater real-
world aftertreatment hydrothermal aging that Medium HDEs are exposed to when compared to
Heavy HDE is addressed by the additional chemical poisoning the aftertreatment of the EPA
Stage 3 engine is exposed to during aging out to 650,000 miles. The real-world data from
Medium and Heavy HDEs supports the assessment that the EPA Stage 3 data at the equivalent of
650,000 miles is the appropriate data to be used when determining the feasibility of the Medium
HDE standards at the final useful life value of 350,000 miles. The hydrothermal and chemical
poisoning of Medium HDEs versus Heavy HDEs was provided as a late comment on the
proposal, with information claimed as CBI comment. The late comment is included in the
docket, though the information claimed as CBI is not publicly available.
44
= MO.. Senao*
Uxfi.
- DEF Doiiitg | |= NHj 5unsoi HaatAd DEF Diiilng | = Tamp Suiitui
¦ InMilalrd "One-box" System
HL Doling
|7" lnje*dor)
m
BC
u
+
2 10.5x4
13*4 13*7
IL
O
Internal '"Switchback"1 Mixing Tube
I 0C U DC O , .
^ 8 II S|g||l|ll-|
IJ-* 5II Slsll t
I
Figure 3-5: Schematic layout (not to scale) of the dual-SCR EAS tested as part of the EPA Stage 3 research at
SwRI.
Table 3-8: Emissions results for the developmental EPA Stage 3 EAS system with light-off SCR and separate
DOC and DPF after 1000 hours of accelerated thermal and chemical aging using the DAAAC (equivalent to
116
-------
approximately 435,000 miles of operation). The SET (2021) results represent updated 40 CFR 1036.510 SET
procedures. Results do not include adjustments for IRAF or crankcase emissions.
Cycle
Results:
FTP
cold
95% CI
FTP
hot
95% CI
FTP
composite
95% CI
SET
(2021)
95% CI
LLC
95% CI
NOx
(mg/bhp-hr)
55
1
14
1
20
1
17
1
29
11
PM
(mg/bhp-hr)
2
1
2
1
2
1
1
1
3
1
NMHC
(mg/bhp-hr)
25
7
9
2
12
2
1
1
35
51
CO
(mg/bhp-hr)
221
61
128
77
141
75
30
22
245
438
C02
(g/bhp-hr)
534
1
511
2
514
2
455
4
617
11
N20
(mg/bhp-hr)
84
7
74
9
76
9
24
69
132
45
For results where the 95% CI is greater than the average, the results are not statistically different from zero based
on a 2-sided Student's t-test at a=0.05
Table 3-9: Emissions results for the developmental EPA Stage 3 EAS system with light-off SCR and separate
DOC and DPF after 1379 hours of accelerated thermal and chemical aging using the DAAAC (equivalent to
approximately 600,000 miles of operation). The SET (2021) results represent updated 40 CFR 1036.510 SET
procedures. Results do not include adjustments for IRAF or crankcase emissions.
Cycle
Results:
FTP
cold
95%
CI
FTP
hot
95%
CI
FTP
composite
95%
CI
SET
(2021)
95%
CI
LLC
95%
CI
NOx
(mg/bhp-hr)
61
5
21
2
27
2
24
1
33
2
PM
(mg/bhp-hr)
2
0
1
2
1
2
1
0
4
1
NMHC
(mg/bhp-hr)
23
11
7
4
9
5
1
0
16
6
CO
(mg/bhp-hr)
245
31
127
134
144
119
15
0
153
20
C02
(g/bhp-hr)
546
3
515
2
519
2
460
1
623
6
N20
(mg/bhp-hr)
69
9
57
4
58
4
30
6
64
22
For results where the 95% CI is greater than the average, the results are not statistically different from
zero based on a 2-sided Student's t-test at a=0.05
Table 3-10: Emissions results for the developmental EPA Stage 3 EAS system with light-off SCR and separate
DOC and DPF after 1839 hours of accelerated thermal and chemical aging using the DAAAC (equivalent to
approximately 800,000 miles of operation) before ash cleaning of the DPF. The SET (2021) results represent
117
-------
updated 40 CFR 1036.510 SET procedures. Results do not include adjustments for IRAF or crankcase
emissions.
Cycle
Results:
FTP
cold
95%
CI
FTP
hot
95%
CI
FTP
composite
95%
CI
SET
(2021)
95%
CI
LLC
95%
CI
NOx
(mg/bhp-hr)
74
3
32
2
38
2
28
1
28
4
PM
(mg/bhp-hr)
1
0
1
1
1
1
1
0
3
1
NMHC
(mg/bhp-hr)
34
12
14
2
17
3
11
5
49
13
CO
(mg/bhp-hr)
259
54
143
25
160
14
18
1
215
71
C02
(g/bhp-hr)
540
10
514
7
518
7
457
6
620
19
N20
(mg/bhp-hr)
118
13
88
5
93
6
34
3
126
8
For results where the 95% CI is greater than the average, the results are not statistically different from
zero based on a 2-sided Student's t-test at a=0.05
Table 3-11: Emissions results for the developmental EPA Stage 3 EAS system with light-off SCR and separate
DOC and DPF after 1839 hours of accelerated thermal and chemical aging using the DAAAC (equivalent to
approximately 800,000 miles of operation) after ash cleaning of the DPF. The SET (2021) results represent
updated 40 CFR 1036.510 SET procedures. Results do not include adjustments for IRAF or crankcase
emissions.
Cycle
Results:
FTP
cold
95%
CI
FTP
hot
95%
CI
FTP
composite
95%
CI
SET
(2021)
95%
CI
LLC
95%
CI
NOx
(mg/bhp-hr)
73
12
31
3
37
1
30
0
34
8
PM
(mg/bhp-hr)
1
1
1
1
1
1
2
1
1
4
NMHC
(mg/bhp-hr)
32
16
11
6
14
7
1
0
40
20
CO
(mg/bhp-hr)
260
68
130
73
149
68
23
7
205
40
C02
(g/bhp-hr)
544
2
516
4
520
4
458
0
629
2
N20
(mg/bhp-hr)
99
59
91
45
92
47
28
4
125
17
For results where the 95% CI is greater than the average, the results are not statistically different from
zero based on a 2-sided Student's t-test at a=0.05
3.1.1,2.1 EPA Stage 3 Off-cycle Emissions Performance
In addition to the FTP, SET and LLC, the EPA Stage 3 engine (with the DAAAC aged
aftertreatment to an equivalent of 435,000 miles) was run on 5 cycles that that cover a range of
off-cycle operation. These cycles are the CARB Southern Route Cycle, Grocery Delivery Truck
Cycle, Drayage Truck Cycle, Euro-VI ISC Cycle and the ACES 4-hour Cycle. The CARB
Southern Route Cycle is predominantly highway operation with elevation changes resulting in
extended motoring sections followed by high power operation. The Grocery Delivery Truck
Cycle represents goods delivery from regional warehouses to downtown and suburban
118
-------
supermarkets and extended engine-off events characteristic of unloading events at supermarkets.
The Drayage Truck Cycle includes near dock and local operation of drayage trucks, with
extended idle and creep operation. The Euro-VI ISC Cycle is modeled after Euro VI ISC route
requirements with a mix of 30% urban, 25% rural and 45% highway operation. The ACES 4-
hour Cycle includes a 5 mode cycle developed as part of the ACES program. Figure 3-6 through
Figure 3-10 show the engine speed, engine torque and vehicle speed of the cycles. The engine
speed and torque shown in the plots are specific to the EPA Stage 3 engine.
CARE. Southern NTE Hoote
0 30W 6000 9DD0 J20D3 150BJ ISOOO 2.1000
Figure 3-6: CARB Southern Route Cycle
119
-------
WVU Grocery Delivery Cycle
^ 120 .220
0 2000 4 GOO 6000 80 DO 10000 12000 14OD0 16000 18000 20000 22000 24000 24.000 28000 30000
Figure 3-7: Grocery Delivery Truck Cycle
120
-------
WVU Ell ISC RauC
-------
NOx standards after accounting for the emissions reductions from closed crankcase (estimated at
6 mg/hp-hr) are greater than 90% and 43%, for Bin 1 and Bin 2, respectively.
Table 3-12: Off-cycle NOx emissions results for the developmental EAS system with light-off SCR and
separate DOC and DPF after 1000 hours of accelerated thermal and chemical aging using the DAAAC
(equivalent to approximately 435,000 miles of operation) without adjustments for IRAF or crankcase
emissions
Bin
SNTE
Grocery
Cycle
ACES
EU ISC
Drayage
1 (g/hr)
0.7
1.0
0.9
0.4
0.3
2 (mg/hp-hr)
32
21
20
31
19
Table 3-13: Off-cycle NMHC emissions results for the developmental EAS system with light-off SCR and
separate DOC and DPF after 1000 hours of accelerated thermal and chemical aging using the DAAAC
(equivalent to approximately 435,000 miles of operation) without adjustments for IRAF or crankcase
emissions
Bin
SNTE
Grocery
Cycle
ACES
EU ISC
Drayage
2 (mg/hp-hr)
1
19
0
2
19
Table 3-14: Off-cycle CO emissions results for the developmental EAS system with light-off SCR and
separate DOC and DPF after 1000 hours of accelerated thermal and chemical aging using the DAAAC
(equivalent to approximately 435,000 miles of operation) without adjustments for IRAF or crankcase
emissions
Bin
SNTE
Grocery
Cycle
ACES
EU ISC
Drayage
2 (mg/hp-hr)
16
122
31
47
261
Table 3-15 Off-cycle NOx emissions results for the developmental EAS system with light-off SCR and
separate DOC and DPF after 1839 hours of accelerated thermal and chemical aging using the DAAAC
(equivalent to approximately 800,000 miles of operation) without adjustments for IRAF or crankcase
emissions
Bin
SNTE
Grocery
Cycle
ACES
EU ISC
Drayage
1 (g/hr)
0.7
3.3
1.5
0.4
1.1
2 (mg/hp-hr)
47
32
34
32
28
Table 3-16 Off-cycle NMHC emissions results for the developmental EAS system with light-off SCR and
separate DOC and DPF after 1839 hours of accelerated thermal and chemical aging using the DAAAC
122
-------
(equivalent to approximately 800,000 miles of operation) without adjustments for IRAF or crankcase
emissions
Bin
SNTE
Grocery
Cycle
ACES
EU ISC
Drayage
2 (mg/hp-hr)
4
30
12
20
30
Table 3-17 Off-cycle CO emissions results for the developmental EAS system with light-off SCR and separate
DOC and DPF after 1839 hours of accelerated thermal and chemical aging using the DAAAC (equivalent to
approximately 800,000 miles of operation) without adjustments for IRAF or crankcase emissions
Bin
SNTE
Grocery
Cycle
ACES
EU ISC
Drayage
2 (mg/hp-hr)
42
300
123
108
334
3.1.1.2.2 EPA Stage 3 Idle Emissions Performance
To evaluate the idle NOx emissions performance of the EPA Stage 3 engine, we directed our
contractor to test the engine in three different configurations over the Clean Idle test procedure at
an ambient temperature range of 20 to 30 °C. The first configuration tested was the EPA Stage 3
calibration. This calibration requires an EGR cooler bypass to be able to stay at this operational
mode for an extended period of time. The second configuration tested simulated the original idle
calibration that the engine was certified with, which wouldn't require the use of an EGR cooler
bypass at ambient temperatures greater than 0 °C. The original idle calibration was simulated by
matching the HC and soot levels during idle operation of the Stage 3 engine which includes CDA
to the original Cummins XI5 engine. Under this calibration, the engine could operate at idle
without an EGR cooler bypass (even for extended periods of time) until the EGR had to be shut
off, for example, due to low ambient temperature. The third configuration tested represented the
second configuration but was operated with the EGR shut off at idle to approximate the
emissions at idle of the second configuration under low ambient conditions. The results from
these tests are in Table 3-18. The data from this testing shows that NOx emission can be
controlled below 10 g/hr under all conditions above 0 °C, even without the use of an EGR cooler
bypass.
Table 3-18 Idle NOx Emission Results from EPA Stage 3 Engine with EAS aged to the Equivalent of 800,000
miles
Mode 1 NOx (g/hr)
Mode 2 NOx (g/hr)
Engine Out
Tailpipe
Engine Out
Tailpipe
1st configuration:
Normal EPA Stage 3 calibration
2.5
0.13
96
0.4
2nd configuration:
EGR rates that wouldn't require EGR
cooler bypass when ambient
temperature is greater than 0 °C
9.5
0.33
104
0.48
3rd configuration: No EGR
86
10.3
229
0.35
123
-------
3.1.1.2.3 Description of the In-use Testing Allowance
The following section describes how we determined the in-use testing allowance for Medium
and Heavy HDEs.
The basic methodology used leverages the methods outlined in the Guide to Expression of
Uncertainty in Measurement (GUM). The basic approach is to describe the distribution of each
term, where each term might have a different shape depending on the input data. The details for
each input are described below. In the case of errors that have a bias, the total variance is
generally developed as the variance plus the bias squared. For the toolA used, the bias was
incorporated as a separate term from the variance (so two lines are used) where the full range of
the bias was input, and a uniform distribution was chosen (because the bias on a given
measurement could uniformly fall anywhere within that range). Rect X 2 was chosen in the tool
to account for the bias-squared relationship discussed above.
Description of Inputs
1. Sulfation. This term represents influence of sulfation between deSOx events. This
term incorporates a bias term because this impact will only serve to increase
emissions. The variance was determined based on test results conducted on the Stage
3 engine after 150 hours of equivalent sulfation, prior to the LO-SCR deSOx, which is
the deSOx frequency set for the strategy. It should be noted that the downstream
deSOx occurs more frequently as part of DPF regeneration every 100 hours (or less
for lower load duty cycles). The results indicated an increase of 0.003 g/hp-hr in
tailpipe emissions at the end of the sulfation interval. Testing could be expected to
happen with equal probability at any point during the interval, therefore a rectangular
distribution was chosen. A second bias term was also entered (one Line 8) as Rect X
2 to account for the bias, since the range is actually from 0 to + 0.003 g/hp-hr.
Rectangular at 0.003 g/hp-hr + Bias (Rect x2) at 0.003 g/hp-hr.
2. Fuels. This term represents the short-term influence of fuel variation on test results.
Short-term fuel influence means an impact which is observed as soon as a new fuel
has replaced the previous one in the engine fuel system. It does not include the
potential long-term impact of a given fuel (if any) on the durability of the emission
control system (that is accounted for later). It was anticipated that this term would be
defined based on the results of testing funded by CRC, but this result was not available
in time. Therefore, the result here was calculated from examining the impact of fuels
on engine-out NOx emissions that has been documented in previous studies, and
assuming the same NOx conversion efficiency observed for the certification fuel. A
value of +/- 20% was used for fuel impact on engine out emissions (examples include
observations of B20 at up to +15% and highly paraffinic renewable diesels fuels at -
15%>), but this represents the upper end of most studies and therefore this value was
chosen as a 2SD value on a normal distribution, as most fuels are anticipated to have a
A A copy of the spreadsheet tool is in the docket for this rule.
124
-------
smaller impact. The impact of this change on the final tailpipe level was 0.002 g/hp-
hr. 2SD = 0.002 g/hp-hr.
3. Sensors. This term represents an element of production variation associated with
potential sensor drift and errors. It was determined based on the results of the Step 1
sensor testing that was conducted on behalf of the Truck and Engine Manufacturers
Associate (EMA) using the EPA Stage 3 engine. A number of offsets to the sensors
involved in the SCR control were examined, although the potential mitigating effect of
long-term trim was not yet evaluated (short-term feedback trim was active however).
It was observed that both positive and negative offsets were seen. The variance was
determined based on the largest positive offset observed which was +0.005 g/hp-hr for
an error involving a +10% offset on both the engine-out and LO-SCR out NOX
sensors on the RMC-SET cycle. Given that this offset required a combination of
errors to line up, this value was chosen as a 2SD level on a normal distribution as most
of the sensor errors was observed to have a smaller impact. 2SD = 0.005 g/hp-hr.
4. Production Variability (not including Sensors). This term is meant to represent the
impact of other production variations, such as batch-to-batch variation in catalysts,
dosing system variations (after short-term and long-term trim corrections), and
production variations that might impact engine-out NOx. We note that the variance of
this term is more difficult to pin down because many of the factors influencing it are
not yet well documented. These variations could lead to higher emissions or lower
emissions; therefore, a normal distribution was assumed. We understand that for this
particular element we could attempt to leverage current production data, but this
would be a problematic approach due to the fact that the current NOx standard would
permit a wider range of production variation to be tolerated, and thus there is not
currently incentive to control variation any more than is necessary. It is reasonable to
assume however, that this value would likely be at least as large as the value observed
for the sensor term. Based on this assumption and consideration of statements from
various manufacturers, we selected a projected value of 0.006 g/hp-hr, and this was set
as a 2SD value using a normal distribution. This value represents a variation of +/- 0.1
% NOx conversion efficiency. We acknowledged that this term represents more
guesswork than others but think that our approach is an appropriate projection based
on our analysis. 2SD = 0.006 g/hp-hr. We note that if, for example, 0.006 g/hp-hr
represents 1SD, the final tolerance stack-up would be 3 mg/hp-hr larger.
5. Field Aging. This term represents the degree to which in-field aging might be more or
less severe than the aging data that were used to develop the DAAAC aging protocol
used for the EPA Stage 3 program. This could involve things such as the long-term
impact of a more severe duty cycle (that might require more DPF regenerations, for
example), impact of fuel impurities (as noted earlier), or the impact of other engine
related issues (such as an EGR cooler leak that was too small to detect, or the impact
of a short-term high temperature excursion results from a turbocharger failure prior to
125
-------
fault detection, etc.). Collectively these could result in more degradation than was
predicted, or less in the case of a less severe duty cycle, though the latter is considered
less likely. To estimate the full range of this impact, we decided to project this by
looking at a range of aftertreatment degradation from half as much as what was
observed in the EPA Stage 3 program to twice as much as what was observed in the
CARB Stage 3 program. This calculation results in an impact ranging from -0.003
g/hp-hr to + 0.012 g/hp-hr. This is a full range of 0.015 g/hp-hr centered on a positive
bias of 0.007 g/hp-hr. It is assumed that a normal distribution would apply to this
range. Therefore, a 1SD spread was assigned at +/- 0.005 g/hp-hr, and a bias term was
defined at +0.012 g/hp-hr with a Rect x 2 distribution as discussed earlier.
When the tolerance stack-up calculations were completed, a final variance at 1SD of 0.00755
g/hp-hr was determined. For this kind of tolerance, it is standard practice to use a "coverage
factor" of 2 to define the reasonable limits of variation, therefore a 2SD value of 0.0151 g/hp-hr
was determined, which when rounded to the nearest 2 significant figures becomes 15 mg/hp-hr.
3.1.1.3 EPA Heavy-duty Diesel Low NOx Demonstration Program
EPA evaluated two different EAS designs provided to the Agency by the Manufacturers of
Emissions Control Association (MECA). Both EAS designs incorporated LO-SCR and dual
urea injection. One of the systems, System A, used close-coupling of the light-off SCR (Figure
3-11, Table 3-19). The other EAS design mounted the light-off SCR closer to the other EAS
components in an under-cab position (Figure 3-12, Table 3-20).
Both EAS designs utilized conventional urea dosing systems for the downstream SCR
position and were evaluated using a heated urea dosing system in the upstream SCR position. In
addition, both EAS designs were tested as part of an EPA contract at SwRI using the same
developmental version of a MY 2017 Cummins XI5 15-liter Heavy HDE engine equipped with
CD A as was used for the CARB and EPA Stage 3 Demonstrations.8
126
-------
Close-coupled, Dual SCR System
Team 2 functional schematic
Figure 3-11: MFC A "System A" EAS with close-coupled light-off SCR
Table 3-19: Summary of catalyst specifications for developmental EAS with close-coupled light-off SCR,
Component
Substrate
Dimensions, Dia.
X Length (inches)
Substrate
Volume
(liters)
Cell Density (cpsi) /
wall thickness (mil)
Notes
Light-off
SCR/ASC
13 X 7
15.2
400/4
Thin wall/low-mass substrate
with ASC zone-coated to the
rearmost 2"
DOC
13 X 4
8.7
400/4
Thin wall/low-mass substrate
CDPF
13X8
17.4
300/7
SCR
10.5X7
19.9
600/2
Two substrates in parallel
SCR
10.5X5
14.2
600/2
Two substrates In parallel
SCR/ASC
10.5X5
14.2
600/2
Two substrates in parallel, both
with ASC zone-coated to the
rearmost 2" of the SCR substrates
127
-------
Exhaust-in (note: heated
dosing is located
upstream of this elbow)
Slip Joint
LoSCR
Figure 3-12: MECA "Team B" EAS with light-off SCR integrated into an under-cab mounting position. This
system was designed to be installed in a Navistar Daycab which is shown in the upper right.
Table 3-20: Summary of catalyst specifications for developmental "System B" EAS with light-off SCR
mounted under-cab.
Component
Substrate
Dimensions,
Dia. X Length
(inches)
Substrate
Volume (liters)
Cell Density
(cpsi) / wall
thickness (mil)
Notes
Light-off SCR/ASC
10.5X8
11.4
400/4
High porosity/low -mass substrate
with ASC zone-coated to the
rearmost 2"
DOC
10.5X6
8.5
400/4
High porosity/low-mass substrate
CDPF
13 X 7
15.2
300/7
Two substrates in series - volume is
SCR+SCR/ASC
13 X 7
30.5
600/3
for combined total. ASC is zone-
coated to the rearmost 2" of SCR #3
EAS accelerated aging of System A was conducted using a "burner aging" version of the
DAAAC.7-9 This used a burner system fueled with diesel fuel and additized engine lubricant to
expose an EAS to both accelerated thermal aging and accelerated chemical aging. The burner
was operated over a series of controlled burner exhaust flow rates and burner exhaust
temperature setpoints that matched specific engine speed and engine load setpoints during
operation of the targeted engine and EAS application (see Figure 3-13). A higher sulfur diesel
fuel (>100 ppm) was used during DAAAC burner aging in order to accelerate sulfur exposure.
128
-------
The DAAAC aging was designed to accelerate thermal and chemical effects by approximately
10 times normal engine operation (i.e., 1 hour of DAAAC is approximately equivalent to 10
hours of actual engine operation). Operation on the DAAAC for 1,000 hours was approximately
equivalent to Heavy HDE operation in a truck application to the end of UL (435,000 miles).
Emissions testing with System A was conducted using the developmental XI5 engine at
accelerated aging equivalents of 435,000 and 650,000 miles of operation. System B was tested
using the developmental XI5 engine in a "degreened" condition only. While System B
demonstrated high NOx reductions over the regulatory cycles in a degreened condition, CO2
emissions increased by 2% on the SET, due to increased exhaust backpressure compared to the
baseline OE EAS. This was likely due to the decision to of use 10.5" diameter substrates for the
LO-SCR and DOC within the EAS design. Emissions results for System A in a degreened
condition are summarized in Table 3-21, and results after EAS DAAAC aging to an equivalent
of 435,000 and 650,000 miles have been added to the docket.10 Emissions results for System B
in a degreened condition are summarized in Table 3-22.
129
-------
Time,s
Speed Load SCR Inlet Temp
Figure 3-13: Example of engine-speed, engine load, and resulting SCR inlet temperature used over the
DAAAC.
Table 3-21: Emissions results for the MECA System A with the EAS in a "degreened" (near 0-hour)
condition. The SET (2021) results represent updated 40 CFR 1036.510 SET procedures.
Cycle
FTP
95%
FTP
95%.
FTP
95%.
SET
95%.
LLC
95%.
Results:
cold
CI
hot
CI
composite
CI
(2021)
CI
CI
NOx
52
10
10
2
16
2
9
1
13
4
(mg/bhp-hr)
PM
1
1
1
0
1
0
1
0
2
0
(mg/bhp-hr)
NMHC
10
15
4
1
5
2
0
1
8
2
(mg/bhp-hr)
CO
135
101
32
13
47
25
5
2
67
40
(mg/bhp-hr)
C02
544
7
512
3
517
2
454.2
0.5
627
2
(g/bhp-hr)
N20
26
2
29
4
28
4
23
164
35
4
(mg/bhp-hr)
For results where the 95% CI is greater than the average, the results are not statistically different from
zero based on a 2-sided Student's t-test at a=0.05
130
-------
Table 3-22: Emissions results for the MECA System B with the EAS in a "degreened" (near 0-hour)
condition. The SET (2021) results represent updated 40 CFR 1036.510 SET procedures.
Cycle
FTP
95%
FTP
95%
FTP
95%
SET
95%
LLC
95%
Results:
cold
CI
hot
CI
composite
CI
(2021)
CI
CI
NOx
49
1
8
1
14
1
9
0
15
8
(mg/bhp-hr)
PM
2
1
1
0
1
0
1
0
4
0
(mg/bhp-hr)
NMHC
26
32
10
15
12
16
2
5
16
35
(mg/bhp-hr)
CO
178
34
59
45
76
43
42
3
145
148
(mg/bhp-hr)
C02
541
5
514
4
518
3
459
2
621
10
(g/bhp-hr)
N20
25
1
29
6
28
5
29
0
33
8
(mg/bhp-hr)
For results where the 95% CI is greater than the average, the results are not statistically different from
zero based on a 2-sided Student's t-test at a=0.05
3.1.1.4 EPA Heavy-duty Diesel CDA Noise, Vibration, and Harshness Study
As discussion in Chapter 3.1.1, EPA has evaluated the use of CDA to lower NOx
emissions by increasing exhaust temperature without increasing CO2 emission. As show in the
results from the EPA and CARB Stage 3 demonstration significant NOx reductions can be
achieve with CDA in combination with EAS improvements. However, one of the concerns with
CDA is that by reducing the number of firing cylinders can cause concerns with noise, vibration,
and harshness (NVH). To evaluate this concern EPA conducted a study to evaluate the effects of
CDA on NVH. This study investigated the impact of fixed CDA on vibration in an on-highway
tractor and to determine the acceptable bounds of CDA operation for engine calibration. The
conclusions from this study were that there are several ways to reduce NVH through design of
the complete system, including, engine conditions where CDA is used, engine mounts, cab
mounts, and seat calibration.11
3.1.2 Baseline Technology Effectiveness
The basis for our baseline technology assessment is the data provided by manufacturers as
part of the heavy-duty, field testing requirements. This data encompasses real world operation
from nearly 300 LHD, MHD, and HHD vehicles. Chapter 6 of the RIA describes how the data
was used to update the MOVES model emissions rates for HD diesel engines. Chapter 3 of the
RIA summarizes the real world emissions performance of these engines.
To assess emissions levels of current production engines on the regulatory cycles we analyzed
the certification data submitted to the agency. For this analysis we focused on MY 2019 and
newer engines.
Table 3-24 include the data for the certification results of the FTP and SET cycle. The
certification results are the test results adjusted for IRAF and DF.
One observation from the data is the range of margin between the certification results and the
standard. The margin for NOx on the FTP cycle is as small as 0.02 g/hp-hr or 10% and as large
131
-------
as 0.15 g/hp-hr or 75% of the standard. For the SET the average compliance margin is slightly
larger than the average margin for the FTP. For the other criteria pollutants the margin between
the certification results and the applicable standards are much larger than for NOx.
Table 3-23: Summary of certification data for FTP cycle
NOx (g/hp-hr)
PM (g/hp-hr)
NMHC (g/hp-hr)
CO (g/hp-hr)
N20 (g/hp-hr)
Standard
0.2
0.01
0.14
15.5
0.1
Average
0.13
0.00
0.01
0.18
0.07
Minimum
0.05
0.00
0.00
0.00
0.04
Maximum
0.18
0.00
0.04
1.10
0.11
Table 3-24: Summary of certification data for SET cycle
NOx (g/hp-hr)
PM (g/hp-hr)
NMHC (g/hp-hr)
CO (g/hp-hr)
Standard
0.2
0.01
0.14
15.5
Average
0.11
0.00
0.01
0.00
Minimum
0.00
0.00
0.00
0.00
Maximum
0.18
0.00
0.04
0.20
In addition to analyzing the off-cycle real world data submitted by manufacturers, we also
conducted and analyzed engine dynamometer data of three modern HD diesel engines. These
engines include a MY 2018 Cummins B6.7, MY 2018 Detroit DD15 and MY 2018 Navistar
A26. These engines were tested on cycles that ranged in power demand and included the LLC
and the SET cycle defined in 40 CFR 1036.510. These results are summarized in Table 3-25,
Table 3-26, and Table 3-27
For two of these engines, both the SET in 40 CFR 1036.510 and 40 CFR 86.1333 were run.
As can be seen from the data, there was not a measurable difference between the results of these
two cycles. Both of these cycles were also run on the EPA and CARB Stage 3 engine. These
results are summarized in Chapter 3.1.1.1. The LLC cycle was also run for these engines to
understand the performance of current engines on this cycle. The results from this cycle vary
much more than the SET and FTP. The Cummins B6.7, which is the only non-tractor engine that
was tested, had the lowest NOx at 0.35 g/hp-hr. The other two engines including the Cummins
XI5 shown in Table 3-3 had results that were multiple times higher than the current standards for
the FTP and SET.
132
-------
Table 3-25: MY 2018 Detroit DD15 engine emissions in g/hp-hr
Cold
FTP
Hot FTP
FTP Composite
SET in
40 CFR
86.1333
SET in 40
CFR
1036.510
LLC
C02
573
550
554
481
472
642
CO
0.54
0.04
0.11
0.01
0.01
0.07
THC
0.02
0.01
0.01
0.01
0.00
0.03
NOx
0.43
0.05
0.10
0.01
0.01
0.61
NMHC
0.01
0.01
0.01
0.01
0.00
0.02
Table 3-26: MY 2018 Cummins B6.7 engine emissions in g/hp-hr
Cold
FTP
Hot FTP
FTP Composite
SET in
40 CFR
86.1333
SET in 40
CFR
1036.510
LLC
C02
621
569
576
486
480
908
CO
0.09
0.04
0.05
0.01
0.02
0.03
THC
0.02
0.04
0.04
0.02
0.04
0.04
NOx
0.48
0.10
0.15
0.05
0.05
0.35
NMHC
0.02
0.04
0.04
0.02
0.04
0.03
PM
0.00
0.00
0.00
0.00
0.00
Table 3-27: MY 2018 Navistar A26 engine emissions in g/hp-hr
Cold
FTP
Hot FTP
FTP Composite
SET in
40 CFR
86.1333
LLC
C02
546
527
529
459
710
CO
0.03
0.02
0.02
0.01
0.03
THC
0.01
0.00
0.00
0.00
0.05
NOx
0.37
0.10
0.14
0.12
0.81
NMHC
0.01
0.00
0.00
0.00
0.05
PM
0.00
0.00
0.00
0.00
3.1.3 GHG Impacts
The combination of active and passive thermal management anticipated for meeting the final
standards can be designed and developed in a manner that does not pose an additional burden for
meeting the heavy-duty engine GHG standards in 40 CFR 1036.108 or the heavy-duty vehicles
GHG standards in 40 CFR part 1037, subpart B. As described in section 3.1.1, the design and
calibration of the CARB and EPA Stage 3 systems achieved significant NOx reductions that
were GHG neutral. The system design and calibration strategy of that system took advantage of
133
-------
GHG improvements at light-load conditions from CDA and calibration changes at higher load
conditions (e.g., injection timing) that approximately offset the impact of increased backpressure
from the additional SCR catalyst volume. The analyses presented in this and in subsequent
chapters assume adoption of GHG neutral technologies similar to what was tested at SwRI.
3,1.4 Estimated Direct Manufacturing Costs for Technology Packages Evaluated
The final program is based on the need to obtain significant emissions reductions from the
heavy-duty transportation sector, and the recognition that there are technically feasible
technologies to achieve such reductions in the MY 2027 timeframe with no compromise to
vehicle utility or safety. As in many prior mobile source rulemakings, the selected standard is
informed by the effectiveness of the emissions control technology, the cost of that emission
control technology, and the lead time needed for manufacturers to employ the control
technology.
3.1.4.1 Cost Teardown Studies
Publicly available information regarding the engineering cost of new engine and vehicle
technologies is a subject of considerable interest. A number of cost analyses in the past few
years have utilized supplier price quotes on designated bills of materials as a methodology for
estimating the increased cost of vehicle improvements. In general, the actual price quotes
provided by suppliers were claimed as confidential business information and have not been
released. In addition, supplier price quotes are typically provided for near-term (e.g., 3-5 years)
estimation, as these are how contracts between OEMs and suppliers are typically written.
This methodology for estimating technology costs to the consumer has several deficiencies.
The lack of transparency regarding the data provided by suppliers does not provide an
opportunity for a full public evaluation of the information. In addition, these near-term price
quotes may not be appropriate for estimating the long-term costs of a regulatory program
implemented in the future. A large fraction of the near-term fixed costs may be recovered and no
longer part of the costs to consumers. EPA is required to evaluate the near- and long-term costs
to consumers that may result from regulatory requirements. To effectively estimate and
communicate those costs, EPA requires a transparent engineering analysis that separates direct
and indirect costs for each major component in the technologies it projects will be implemented
to meet the new requirements.
EPA previously directed contractual work to develop an analytical methodology that is based
on technical knowledge of the engineering, design, and development of advanced vehicle
technology components, systems, and subsystems. In addition, the previous contractual work
performed pilot studies to demonstrate the methodology on representative vehicle categories.
A key objective for these studies was transparency— methodologies, assumptions, and inputs
well-documented, clearly explained, and releasable to the public, except to the extent that those
essential inputs included information claimed as confidential.
3.1.4.1.1 EPA HDE Cylinder Deactivation and Variable Geometry Valvetrain Teardown Cost
Study
The cost of CDA for Light HDE, Medium HDE, Heavy HDE, and Urban Buses was
estimated based upon a detailed, tear-down study of heavy-duty diesel valvetrains, the Heavy -
134
-------
Duty Engine Valvetrain Technology Cost Assessment (or "FEV Valvetrain Study").12 The study
was conducted by FEV North America, Inc. under a contract with EPAB and was submitted to an
independent peer review.13'14 The FEV Valvetrain Study investigated design modifications to a
production engine cylinder head from the Cummins XI5 engine. These design modifications
allowed the addition of a variable-geometry valvetrain system in one of two different
configurations. One configuration was implementation of individual CDA with an integrated
exhaust brake. The other configuration was implementation of late intake valve closing (LIVC).
The final objective was to evaluate the incremental cost of CDA and LIVC hardware as two
distinct technology packages.
The cost of the CDA and LIVC technology packages were evaluated relative to a baseline
valvetrain technology represented by a 2019 Cummins X15 engine. FEV also investigated other
valvetrain designs used by diesel HDE in order to develop fleet average per-cylinder costs for
these valvetrain technologies. The baseline and new technology packages were required to have
similar overall performance with respect to service life and other functional objectives. Table
3-1 shows estimated direct manufacturing costs for application of CDA to diesel HDE based on
costs derived from the FEV study. For purposes of EPA's cost analysis of CDA applied to diesel
HDE, Light HDE costs were based on application of CDA hardware to 8-cylinder engines while
Medium HDE, Heavy HDE, and Urban Bus costs were based on application of CDA hardware to
6-cylinder engines. Both costs, airflow control, and thermal management appeared to roughly
comparable for both CDA and LIVC within the analysis, with some advantages with respect to
higher BMEP CO2 reduction for LIVC at slightly higher cost. For more details regarding the
FEV Valvetrain Study, please refer to the final report for the study within the docket for this rule.
The costs for CDA were assumed as the cost for active thermal management via valvetrain
system improvements within EPA's costs for diesel HDE for the emissions standards in the final
rule.
Table 3-28: Summary of CDA Direct Manufacturing Costs from Teardown Study
EPA HD Engine Class
Light HDE
Medium HDE
Heavy HDE
Urban Bus
CDA Valvetrain Hardware - Tier 1
Supplier Cost to Manufacturer (2019 $):
204.16
153.12
214.56
153.12
3.1,4.1.2 EPA AdvancedEAS Teardown Cost Study
EPA also sponsored an additional study to examine in detail an advanced heavy-duty diesel
EAS technology package utilizing a dual-SCR system with heated dosing for the light-off SCR
and capable of approximately 90% NOx reduction relative to post-2010 heavy-duty diesel
emission control systems, the Heavy-Duty Vehicles Aftertreatment Systems Cost Assessment
(FEV EAS Study)15. As with the valvetrain study, this was also conducted by FEV North
America, Inc. The FEV EAS Study previously submitted to the docket for the proposal was
updated slightly for the final rule to incorporate updated PGM costs.15'16 The updated FEV EAS
Study costs served as the basis for diesel EAS system costs for the final rule.
Within the FEV EAS Study, direct manufacturing costs associated with the advanced EAS
technology packages were evaluated relative to a baseline OE (MY 2018) EAS technology
B U.S. EPA Contract No. 68HERC19D0008, Task Order No. 68HERH20F0041.
135
-------
representative of the current state of design. The updated costs from the FEV EAS Study are
summarized within Table 3-29. The incremental cost increase for the advanced system capable
of meeting the final standards was approximately $108 to 316 higher relative to the costs used
for the proposal. For details regarding the FEV EAS Study, please refer to the final report for the
updated study within the docket for this rule.
Table 3-29: Summary of Direct Manufacturing Costs from the FEV Exhaust Aftertreatment System Study
with PGM Costs Updated for the Final Rule
EPA HD Engine Class
Light HDE
Medium
HDE
Heavy HDE
Urban Bus
Total "Baseline" MY2018 EAS Cost
$2,693.69
$2,643.39
$3,919.54
$2,723.17
Total "Advanced" EAS Cost (2019$)
$4,490.44
$4,345.84
$6,080.84
$4,459.33
Total EAS Incremental Cost from 2018 System to
Final Rule (2019 $):
$1,796.75
$1,702.45
$2,161.30
$1,736.16
Difference in Incremental Cost Comparing the Final
Rule to the Proposal (2019 $)
$ 316.22
$ 206.96
$ 110.13
$205.37
Note: Costs from the FEV Study and subsequent updates were originally calculated on a 2020$ basis and were
converted to a 2019$ basis for this table to allow direct comparison with other costs within this RIA.
3.1.4.2 Closed Crankcase Systems Technology Costs
We project that manufacturers meeting the final crankcase requirement by closing the crankcase
on turbocharged engines that do not have closed crankcase systems already, will rely on
engineered closed crankcase ventilation systems. These systems filter oil from the blow-by gases
prior to routing them into either the engine intake or the exhaust system downstream of the
turbocharger but upstream of the exhaust aftertreatment system. We have estimated the initial
direct manufacturing cost for manufacturers of these systems to be approximately $41 (2002$).17
To estimate the baseline cost, we multiplied $41 (2002$) by the percentage of engines that
already have closed crankcase systems, which resulted in a baseline cost of $13 (2002$). We
estimated the percentage of engines that already have closed crankcase systems at 32.5%, based
on the certification data. For our cost analysis, we converted these estimates to 2017 dollars,
which resulted in $18 (2017$) for the baseline cost and the same cost of $37 to implement closed
crankcases in the remaining CI engines for our final standards.
3.2 Spark-Ignition Technology Feasibility
3.2.1 Baseline Technology Effectiveness
In 2018, EPA evaluated heavy-duty gasoline Class 3 and 4 vehicles from three different
manufacturers to better understand the state of criteria pollutant control technology incorporated
on gasoline engines used in these applications.18 Evaluations were conducted using laboratory
chassis dynamometer testing and real-world Portable Emissions Measurement System (PEMS)
testing.
Most chassis-certified heavy-duty vehicles are subject to EPA's light-duty Tier 3 program and
these vehicles have adopted many of the emissions technologies from their light-duty
counterparts (79 FR 23414, April 28, 2014). To meet these Tier 3 emission standards,
manufacturers reduced the time needed for the catalyst to reach operational temperature by
implementing cold-start calibration strategies to reduce light-off time. They have also moved the
136
-------
catalyst closer to the engine. Manufacturers have not widely adopted the same strategies for their
heavy-duty engine-certified products. The purpose of this test program was to observe emissions
performance for technologies that are available in the market today and establish a baseline to
evaluate the performance of advanced technologies to further reduce criteria emissions.
3.2.1.1 Baseline Vehicles Tested for Exhaust Emissions
Three vehicles were chosen for evaluation based on majority market share. Two vehicles,
Class 4 (GVWR 14,001 - 16,000 pounds), with powertrains produce by General Motors (GM)
and Ford respectively, utilized engines that were dynamometer certified. The third vehicle,
produced by Fiat Chrysler Automobiles (FCA), was a Class 3 (GVWR 10,001 - 14,000 pounds)
chassis certified truck that meets the Federal HDV2 Tier 3 Bin 570 standards, and has the same
powertrain (engine and transmission) that FCA uses in their Class 4 engine certified trucks. The
FCA test article had comparable gross combined vehicle weight (GCWR) as the Class 4 vehicles
tested but employed aftertreatment technology tailored for Tier 3 chassis certification. Table
3-30 lists the major specifications of the three vehicle/powertrain combinations considered in this
study.
Table 3-30: Heavy-Duty Gasoline Vehicle Emissions Investigation Vehicle Specifications
G.M.
Ford
FCA
Configuration
Box Truck
Box Truck
Pickup
Certification
HDGE
HDGE
Tier 3 Bin 570
Model Year
2015
2016
2017
Odometer
48,000
37,000
43,000
Eng. System
NA/PFI/TWC
NA/PFI/TWC
NA/PFI/EGR/TWC
Displacement
6.0L V8
6.8L V10
6.4L V8
Power
297 hp @4,300 rpm
305 hp (a> 4,250 rpm
410 hp @ 4,600 rpm
Torque
372 lb ft @ 4,000
rpm
420 lb ft @ 3,250
rpm
429 lb ft @ 4,000 rpm
Transmission
6 spd Auto
6 spd Auto
6 spd Auto
GVWR (lbs)
14,500
14,500
13,300
GCWR (lbs)
20,500
22,000
19,900
3.2.1.2 Baseline Tests Performed for Exhaust Emissions
As previously stated, two of the vehicles tested were equipped with dynamometer certified
engines for use in vehicles with a GVWR over 14,000 lbs while the third vehicle was a chassis
certified HDV2. These vehicles were chosen as the engines representing the bulk of the HD SI
vehicle market. The lighter HDV2 vehicle was chosen because of its chassis certification
resulting in its aftertreatment system more closely resembling what is commonly found on Tier
III light-duty vehicles.
The purpose of this particular program was to investigate the current state of HD SI engine
criteria emissions performance. Because cold start emissions are not strongly emphasized in HD
SI engine test, manufacturers generally locate three-way catalysts for exhaust aftertreatment
significantly downstream from the exhaust manifold. Because chassis certification places a
higher weighting on cold start results, the HDV2 vehicle we tested was designed to reach light-
137
-------
off temperature sooner. Its catalysts were significantly closer to the exhaust manifold than in an
engine-certified configuration. Table 3-31 shows the average distance, in meters, from the outlet
of the exhaust manifold to the front face of the catalyst substrate. Where two catalysts are used,
one for each bank of cylinders on a V8 or VI0 engine, the value represents the average of the
two distances.
Table 3-31: Average distance from exhaust manifold to catalyst.
Manufacturer
Average Exhaust Manifold
to Catalyst Distance
(meters)
G.M./Isuzu
2
Ford
1.6
FCA
0.9
Location of the catalyst relative to the exhaust manifold has a significant impact on overall
tailpipe emissions, as a shorter distance enables more rapid heating and catalytic reduction after
cold start. A longer distance reduces the maximum catalyst temperature during high load
operation, protecting the washcoat from thermal degradation. The assumptions investigated in
this test program were as follows:
• Gasoline stoichiometric operation and advanced three-way catalyst can provide a high
level of efficiency and nearly zero warmed-up emissions rates.
• Vehicle weights and loads can drive high exhaust gas temperatures.
• High exhaust gas temperatures can drive the need for fuel enrichment and related
strategies that protect engine components and the catalyst.
• Location of the catalyst is partially dictated by exhaust gas temperature.
• Rearward catalyst locations can hinder catalyst light-off as well as performance under
extended low-load operation.
3.2.1.3 On Road PEMS testing for Exhaust Emissions
Each vehicle was subject to real world emissions testing and driven during the workday on
several routes EPA uses to collect real world drive emissions. During these drive evaluations,
each vehicle was equipped with a portable emissions measurement system (PEMS). These
PEMS units are 40 CFR part 1065 compliant, with independent emissions measurement for CO,
CO2, NO, NO2 and NMHC emissions. Additionally, data is collected for exhaust flow, GPS
location, environmental conditions, and selected CAN signals via the vehicles OBD connector.
Temperatures within the exhaust system were also recorded during these drive cycles: exhaust
gas temperature at the exhaust manifold outlet and catalyst inlet cone, and catalyst substrate
temperature 1-inch reward of the front face of the catalyst.
Where possible, each vehicle was tested across a range of test weights: curb weight plus
instrumentation; 90% GVRW (gross vehicle weight rating); and where possible, 90% GCWR
(gross combined weight rating, which represents truck and trailer weight.) Each real-world drive
schedule was on public roads and subject to varying traffic loads dependent on the time of day as
well as all variation as a result of traffic control. When possible, each vehicle was driven on each
route on three different days. Table 3-32 describes the routes used for collection of real-world
drive emissions.
138
-------
Table 3-32: Description of Real-world PEMS Testing Routes
Route
Distance
(mi)
Avg. Speed
(mph)
Description
A
7
21
Low speed, light load
B
12
24
Medium speed, short duration high load
C
32
45
High speed, short duration high load
D
84
63
High speed, sustained high load
E
30
40
Medium speed, high load
3.2.1.4 Laboratory Chassis Testing for Exhaust Emissions
The vehicles described previously were tested at EPA's Ann Arbor, MI National Vehicle and
Fuel Emissions Laboratory (NVFEL). Laboratory testing was conducted to remove the
variability inherent in real world PEMS testing such as engine coolant start temperature, load,
and traffic conditions. Both the Ford and the GM/Isuzu vehicles engines were, as previously
noted, certified to HD engine standards on an engine dynamometer, while the FCA vehicle was
chassis certified as a complete vehicle to HDV2 standards. By testing engine certified vehicles
on the chassis rolls, it was possible to highlight the emphasis each type of certification places on
the vehicle, engine, and catalyst system design considerations. For testing purposes, each vehicle
was subjected to both cold start and hot start testing using Tier 3 (10% ethanol) certification fuel.
Each vehicle was tested at an estimated test weight (ETW) condition that represented the
vehicle's curb weight plus the weight of instrumentation, as well as 90% GVWR condition.
Because the FCA vehicle was certified as a Tier 3 HDV2 it was tested at curb weight plus 1/2
payload. Table 3-33 are the test weight and dynamometer coefficients used for this testing.
Table 3-33: Test weights and dynamometer coefficients used for NVFEL HD gasoline testing.
Test #1
Test #2
Test #3 (FCA
only)
ETW (lbs)
9,320
14,000
11,000
Target A (lbf)
123.23
87.54
87.54
Target B (lbf/mph)
0
1.399
1.399
Target C (lbf/mphA2)
0.0917
0.1215
0.1215
Each vehicle-engine combination was subject to four distinct test cycles, two of which were cold starts, with the
remaining two warm starts. Each test cycle is described below.
Cold Start, Federal Test Procedure (FTP-75). The FTP-75 is used for emissions certification
and fuel economy testing of light-duty vehicles in the United States. The FTP-75 consists of
three phases: a cold transient phase (ambient temperature 20-30 deg. C), a stabilized phase and a
hot start transient phase. The FTP-75 was used to understand the differences in cold start
strategies and catalyst architectures resulting from the differences in certification testing between
HD dynamometer and HD chassis. Figure 3-14 illustrates the FTP-75 three-phase test cycle.
139
-------
70
60
_ 50
o.
£
v
4i
CI.
IS)
Ji 30
:>
20
10
tPA federal i^st Procedure
Lentgh=1874 seconds, Distance=l 1,04 mi, Avg. Speed 21.2 mph
Cold Start Phase
0 -1
A
Stabilized Phase
11ot Start Phase
A
k
200 400 S00 SOO 1000 1200 1400 1600 1SOO 2000
Time (sec)
Figure 3-14: FTP-75, Cold start three phase test cycle
The Highway Fuel Economy Driving Schedule (HWFE) was chosen to compare emissions
performance under simulated urban driving conditions. Figure 3-15 illustrates the FIFE test cycle
that was used. This test was run as double FTWFE cycles, where the first cycle is used as a
warmup, or prep, and there is no emissions sampling or recording conducted
140
-------
70
60
-50
Q.
E
-o 40
ai
aj
Q.
aj 30
u
20
10
0 J
c
EPA Highway Fuel Economy Cycle
Length=765 seconds, Distance=10.26 mi, Avg. Speed=48.3 mph
) 100 200 300 400 500 600 700
Time (sec)
800 900
Figure 3-15: EPA Highway Fuel Economy Cycle
Phase LA92 drive cycle. The LA92 drive cycle is a CARB-developed dynamometer schedule.
The LA92 was originally developed as an inventory improvement tool, and compared to the
FTP, it has a higher top speed, a higher average speed, less idle time, fewer stops per mile, and
higher rates of acceleration. For the purposes of this comparison testing, two back-to-back LA 92
cycles were utilized, doubling the distance and creating a 4-phase test. Phase 1, a warm start, was
followed by a warmed-up Phase 2, followed by a 30-minute engine off soak. Phase 3 is a warm
start after the engine off soak and is a repeat of the earlier Phase 1 drive cycle, and Phase 4 is a
repeat of Phase 2.
141
-------
4 Phase LA-92 Drive Schedule
Length=2870 seconds, Distance=19.64 mi, Avg. Speed=24.6 mph
80
70
60
_c
Q_
50
~o
CD
01
Q_
40
00
Q)
(J
30
In
d)
>
20
10
0
Phase 1
500
Phase 2
h
1000
Phase 3
1500
Time (sec)
2000
Phase 4
h
t,
2500
Figure 3-16: EPA 4 phase LA92 test cycle.
HD GEM Cycle (i.e., Super Cycle). This cycle is a composite of the many cycles used in the
process of certifying trucks to HD GHG vehicle standards. The super cycle drive cycle consists
of a combination of low speed, low load cycles followed by a 10-minute idle, and a return-to-
service portion consisting of 55 and 65 mph cruise conditions. Phases 1 and 2 are consecutive
ARB Heavy Heavy-Duty Diesel Truck transient modes from the ARB HHDDT schedule, are
14.3 miles in combined length, and represent an average speed 15.4 mph. Phases 1 and 2 are
followed immediately by Phase 3, a 10-minute idle. Phase 4 is a return-to-service cycle
consisting of an acceleration from idle to a 55-mph cruise, followed by another acceleration to a
65 mph cruise, and a return to idle. Phase 4 for has an average speed of 55.8 mph, is 29.2 miles
in length, and both acceleration and deceleration rates are 0.5 mph/sec. The purpose of this cycle
was to investigate how the lower exhaust gas temperature resulting from low-load operation
affect catalyst activity and the emissions generated during a high-load, return-to-service event.
142
-------
90
80
70
_c
60
Q.
£
TP
V
50
V
CL
l/l
40
20
10
0
Heavy-duty GHG Cycle (Super Cycle)
Length=58Q0 seconds, Distance=43.4 mi, Avg. Speed=30 mph
Phase 1
nJ
a
n
i
1000
Phase 2
2000 3000
Time (sec)
Idle
Phase 4
4000
5000
Figure 3-17: Super cycle, GEM greenhouse gas cycle
3.2.1.5 Baseline Exhaust Emissions Results
Due to the inherent variability of real-world driving conditions, as well as the absence of
defined test cycles or off-cycle emission standards for HD gasoline engines, direct comparisons
of the onroad PEMS and chassis testing results cannot be used to classify the emissions
performance of any one vehicle as above or below an applicable HD standard. However,
comparisons of emission rates observed under similar conditions, time to catalyst light-off, and
overall performance of each configuration relative to the others, can be made.
Most illustrative are the test results and data acquired in the laboratory employing the FTP-75
cold start test procedure. Figure 3-18 and Figure 3-19 show cumulative hydrocarbons and NOx
respectively for each vehicle at each test weight. Each figure clearly shows that once catalyst
light-off is achieved, the sharp knee in the curve between 20 seconds and 140 seconds, emissions
rates decline significantly and remain so for the remainder of the test. The top lines in the figures
also illustrate how quickly the emissions can accumulate if catalyst light-off is delayed, allowing
most of the emissions totals to be achieved in the first minutes of operation.
143
-------
100
90
tto
/O i
FiO
50
•10
2D
> 10
II
n
_L_
Pi
7 DO
400
60O BOO
Timi? (see)
1ODO
1700
-V.5. Ford 6.8L (9320 lbs) Ford 6.81 (1-1,000 lbs}
— GM/lsuzu G.OL (9320 lbs) GM/lsuzu G.OL (14,000 lbs) FCA G.4L <9320 lbs)
FCA 6.41 (14,000 lbs)
Figure 3-19: FTP-75 Cold start cumulative NOx comparison
144
-------
Figure 3-20 sharpens the focus on catalyst architecture as well as possible calibration
techniques driven by the particular certification tests. Figure 3-20 shows the effect that engine
load has on the exhaust temperature of the Ford 6.8L and the GM/Isuzu 6.0L, both of which are
dyno certified with little emphasis on cold start emissions. In contrast the FCA 6.4L which is a
Tier 3 Bin 570 chassis certified vehicle shows no impact on catalyst light-off time, exhaust gas
temperature, due to increased load. The FCA 6.4L was certified to the FTP-75 which
emphasizes cold start emissions. This emphasis results in catalyst architecture, shorter exhaust
manifold to catalyst distance, as well as cold start controls and calibrations more closely related
to light-duty trucks and passenger cars.
h i...
Ford 6.8L GM/Isuzu 6.0L FCA 6.4L
Dyno Cert Dyno Cert Chassis Cert
¦ 9320# ¦ 14,000#
Figure 3-20: FTP-75 Catalyst light-off time comparison
Catalyst location can not only affect light-off but can also affect catalyst temperature during
extended periods of idle. If idle conditions are long enough, catalyst temperatures may fall below
350 C, a temperature associated with reduced conversion efficiency. Given this reduced catalyst
conversion efficiency, an engine out emissions spike caused by an applied load can become a
tailpipe emissions spike. Figure 3-21 illustrates this condition during the 10-minute idle portion
of the Super Cycle (Figure 3-17). Each vehicle enters the idle with a catalyst temperature of
approximately 500 C. Over the course of the idle, catalyst temperatures decline. Those vehicles
with the largest distances from manifold to catalyst (Table 3-31) fall below 300 C.
160
140
__ 120
O
(D
— 100
o
$ 80
o
+-»
*2 60
i—
40
20
145
-------
Empty Test Weight (9320 lbs)
10 Min. Idle
50 52 54 £6 SS 60 62 64 66 6& 70
Time (Mm)
Ford Cat Temp GM/lsuzu Cat Temp FCACatTemp Speed (km/hr)
Figure 3-21: Extended idle catalyst cool down comparison
3.2.1.6 Baseline Refueling Emissions
As mentioned in Chapter 2.3.2.4 of this RIA, these vehicles are subject to evaporative
emission standards, but have no refueling requirements. We are unaware of any HD SI engines
certified for incomplete vehicles that implement ORVR technologies today. For our feasibility
analysis, we believe these HD SI engines would not implement ORVR without a regulatory
driver and assumed zero adoption of ORVR technology in our baseline.
3.2.2 Projected Technology Effectiveness
The emissions performance of the advanced catalyst technologies was evaluated in EPA's HD
SI demonstration program. We also evaluated a combination of additional data sources including
MY 2019 compliance data and engine mapping data to project the effectiveness of these
technologies and inform the level of stringency in our standards. A description of these data and
our analysis of them is presented in this section.
We project the effectiveness of implementing ORVR for incomplete HD SI vehicles based on
the performance of complete vehicles subject to the Tier 3 evaporative and refueling
requirements applying assumptions to account for increased fuel tank sizes.
3.2.2.1 MY 2019 HD SI Compliance Data for FTP Exhaust Emissions Performance
Four engine manufacturers certified HD SI engines in MY 2019. These manufacturers
certified six engine families ranging in displacement from 6.0 to 8.8 liters.19 20 Table 3-34
presents the MY 2019 FTP-based emission levels reported for the three pollutants addressed by a
TWC: NOx, NMHC and CO. We organized the engines by descending NOx level. One engine,
labeled "Cert Engine 6", is below the final NOx standard for MY 2027 while maintaining
146
-------
relatively lowNMHC and CO emissions. While this high performing engine, which is available
today, demonstrates that it is possible to meet the new NOx standard, we acknowledge that these
certification results are representative of a shorter useful life period than we are implementing.
PM emissions for most of these engines were undetectable and reported as zero for certification,
suggesting the 5 mg/hp-hr standard is feasible for HD SI.C
Table 3-34: Family Emission Limits Reported for the Six Certified HD SI Engines in MY 2019; NOx and
NMHC values are converted from g/hp-hr to mg/hp-hr to match the units of our new standards
Cert Engine
1
Cert Engine
2
Cert Engine
3
Cert Engine
4
Cert Engine
5
Cert Engine
6
NOx (mg/hp-hr)
240
160
120
104
70
40
NMHC (mg/hp-
hr)
50
50
60
80
80
80
CO (g/hp-hr)
1.5
3.7
6.6
8.6
12.7
3.7
Fraction of MY
2019 HD SI Sales
20%
2%
20%
4%
48%
5%
In order to evaluate the NMHC and CO emissions, we calculated an overall average emission
rate for each pollutant that includes all engines. Table 3-35 compares this average with the EPA
2010 standards, the new 2027 standards, and results from the engine family with the best NOx
emission performance of the MY 2019 compliance data.
Table 3-35: Average emission performance for Certified HD SI Engines in MY 2019
Pollutant
EPA 2010
Standard
EPA 2027
Standard
Overall Average
Best NOx
Performance
NOx (mg/hp-hr)
200
35
122
40
HC (mg/hp-hr)
140
60
67
80
CO (g/hp-hr)
14.4
6
6.1
3.7
Table 3-35 compares the average NOx, NMHC, and CO emission performance of the six
engines and displays the EPA 2010 standard, the EPA 2027 standard, as well as the 2019 cert
engine family with the best NOx. When calibrating their engines, SI manufacturers experience a
tradeoff in emissions performance for the three pollutants in their TWCs and each manufacturer
will optimize their emission controls differently. As expected, the table shows no clear trend in
NMHC and CO emissions related to the reduced NOx. The new 2027 standard levels are aligned
with the six certification engines' average emissions for NMHC and CO. The new standards are
likely achievable by minor calibration changes, such as incorporation of cold start catalyst light-
off strategies and refinement of the catalyst protection fuel enrichment and related strategies.
These results support FTP standards of 60 mg NMHC/hp-hr and 6 g CO/hp-hr, consistent with
the overall average NMHC and CO levels achieved for 2019. We describe the feasibility of the
final standards based on NMHC, CO, NOx, and PM emissions levels achieved in our
demonstration program in detail in section 3.2.2.3.
c One engine reported a 0.005 g/hp-hr PM FEL.
147
-------
Four engine manufacturers certified alternative fuels HD SI engines in MY 2019 - 2021 that
performed favorably as compared to the MY 2027 standards. These manufacturers certified
eight engine families ranging in displacement from 6.0 to 8.8 liters. Four of the engine families
selected for comparison were certified with compressed natural gas (CNG) and four engine
families were certified with liquified petroleum gas (LPG). Table 3-36 presents the FTP-based
emission levels reported for the three pollutants: NOx, NMHC and CO. Three of the CNG and
all of the LPG engine families are below the final NOx standards for MY 2027. All of the
NMHC and CO emissions reported are below the new MY 2027 standards. We acknowledge
that these certification results are representative of a shorter useful life period than we are
finalizing in this rule; but, PM emissions levels certified for these selected CNG and LPG
engines were 1 mg/hp-hr or lower, which supports the5 mg/hp-hr standard for PM.D
Table 3-36 Family Emission Limits Reported for Four CNG and Four LPG HD Engines in MY 2019 - 2021;
NOx and NMHC values are converted from g/hp-hr to mg/hp-hr to match the units of our new standards
Cert El
Cert E2
Cert E3
Cert E4
Average
MY 2027
Std
Fuel
CNG
CNG
CNG
CNG
NOx (mg/hphr)
6
20
10
70
27
35
NMHC (mg/hphr)
3
9
22
7
10
60
CO (g/hphr)
4.0
1.3
2.5
4.4
3
6
Cert E5
Cert E6
Cert E7
Cert E8
Average
MY 2027
Std
Fuel
LPG
LPG
LPG
LPG
NOx (mg/hphr)
20
20
10
20
18
35
NMHC (mg/hphr)
15
48
42
51
39
60
CO (g/hphr)
2.8
2.7
5
5.6
4
6
3.2.2.2 EPA Engine Mapping Test Program for SET Exhaust Emissions Estimation
To assess the potential for emission reductions in HD gasoline engines over sustained loads,
EPA evaluated engine fuel mapping data from a testing program previously performed by EPA
as part the of the HD GHG Phase 2 rule. EPA contracted SwRI to test a production MY 2015
Ford 6.8L VI0 gasoline engine to assess CO2 emissions and to evaluate the new fuel mapping
test procedures developed for that rulemaking. As part of that work, the engine was run on an
early version of 40 CFR 1036.535, which is the steady state fuel mapping procedure that requires
the engine to be run at nearly 100 speed and torque points for 90 seconds. The first 60 seconds
are for the engine and fuel consumption to reach stability and the last 30 seconds are averaged to
create the fuel map.
Since continuous dilute criteria emissions were also collected for the test, we recently directed
SwRI to reevaluate those results and create three versions of the data that summarized fuel
consumption and emissions (NOx, CO, NMHC and CO2) versus engine speed and torque. The
first version analyzed conditions where the engine went into power enrichment, consistent with
D One engine reported a 0.005 g/hp-hr PM FEL.
148
-------
strategy used in the production application of the engine. The second version analyzed the
conditions where the engine controller activated a catalyst protection fuel enrichment strategy
but did so before a power enrichment strategy was activated (this is due to a programmed delay
for power enrichment of approximately one minute in the production engine controller). The
third version analyzed only conditions where the engine maintained stoichiometric air-fuel ratio,
achieved by limiting engine load to keep exhaust temperatures slightly below the level that
would activate the thermal protection strategy programmed into the production software.
These three analyses of the data differed only in the peak torque portions of the map. As in
other portions, the engines maintained stoichiometric air-fuel ratio control for a majority of the
points (below about 90 percent throttle). For each of the maps, the peak torque points were used
to calculate the A, B and C speeds as well as the torque values, so there were three unique sets of
surrogate SET test points. Emission mass rates for CO, NMHC, NOx, and CO2 and fuel
consumption were calculated from each map by interpolation of the maps at each of the SET test
points. Finally, the results were weighted according to the existing Cl-based weighting factors
outlined in 40 CFR 1036.512. The engine and emission control components were not aged to the
useful life requirements in this analysis.
The data analysis below includes operation at three distinct engine speeds described above
and at several different loads, consistent with the approximate test points that would be required
to perform the SET test procedure and then calculate a composite emissions level for the engine.
The data presented include emission levels, fuel consumption rates, and engine power observed
at the required SET test points and while operating in three distinct modes as allowed by
production software controls (i.e., power enrichment mode, catalyst protection enrichment
mode, and stoichiometric operation).
While not typically observed during the transient FTP test or torque mapping procedure, the
engine controller activated a power enrichment mode after approximately one minute when
throttle openings were above 90%. The extra fuel resulted in a slight increase in power. Power
enrichment is sometimes used on gasoline engines to produce approximately 5% additional
power beyond what is made when the air to fuel ratio is maintained at stoichiometry.
Stoichiometric operation is the fundamental operating mode needed for three-way catalyst
systems to simultaneously reduce HC, CO and NOx emissions, but as described above, it is not
the mode that produces peak power.
Another operating mode observed in the data is catalyst protection fuel enrichment. When the
catalyst or other critical engine components are exposed to high exhaust gas temperatures,
damage can occur that affects the durability of these components, and manufacturers typically
implement control strategies that use a limited amount of fuel enrichment to cool the exhaust gas
and protect critical components. The fuel enrichment reduces the amount of excess oxygen that
supports the exothermic (heat releasing) reaction in the catalyst and also reduces the temperature
of the combustion gases exiting the engine. The combination of these two temperature-reducing
strategies effectively provides control of exhaust gas temperatures and protects critical exhaust
components from irreversible damage. Other strategies that maintain effective emission control,
expand the area of stoichiometric operation, and still provide protection of critical engine and
catalyst components are discussed in Chapter 2 of the RIA.
As observed in the composite SET test data below, any enrichment mode, whether for power
or catalyst protection purposes, can result in substantial emission increases and higher fuel
149
-------
consumption. As seen in Table 3-37, when the engine is commanded into power enrichment
mode and is no longer maintaining stoichiometric operation, the NMHC and CO increase
substantially, and the engine consumes more fuel. The NMHC emissions are more than 10 times
higher while the CO emissions are almost 50 times higher than the stoichiometric operating
mode. NOx emissions are reduced about 60% in power enrichment mode as expected because of
the rich operation, however stoichiometric NOx emissions can be improved with catalyst design
and calibration. Since this is a MY 2015 production engine, it was not designed or calibrated for
optimum emissions for sustained high load operation at mid operating speeds such as
demonstrated over the SET cycle. Improved NOx emission control required over the FTP test
cycle with this rule is expected to also result in improvements in the NOx levels over the SET
cycle. It is important to note that this power enrichment mode is not typically observed during
the transient FTP test due to the short periods of time spent at high and full power loads. The
short FTP time at load limits any power related enrichment features from activating like
observed in the sustained full power test points in the SET testing described above.
Table 3-37: Comparison of Simulated 6.8L V10 SET Composite Emissions to MY 2027 Standards
NOx
HCa
CO
CO2
BSFC
(mg/hp-hr)
(mg/hp-hr)
(g/hp-hr)
(g/hp-hr)
(lb/hp-hr)
Power Enrichment Allowed
11
110
45.2
587.3
0.479
Catalyst Protection with
No Power Enrichment
19
30
11.4
617.7
0.463
Stoichiometric Operation
28
10
0.97
626.6
0.457
Spark-Ignition Exhaust
Emission Standards for SET
35
60
14.4
Duty Cycle
MY2027 and later
a Hydrocarbons measured in the dataset were NMHC.
As discussed above and illustrated in Figure 3-22, NOx emissions remain reasonably
controlled under all operating modes; however, NMHC and CO emissions increases are closely
tied to enrichment events. The MY 2027 HC and CO standards for the FTP cycle are achieved in
stoichiometric operation, but CO begins to approach today's FTP standard when catalyst
protection is enabled. Power enrichment causes drastic spikes in both NMHC and CO. We are
including the SET duty cycle to incentivize manufacturers to expand the stoichiometric operation
under heavy load conditions of their HD SI engines and maintain the maximum TWC
effectiveness. The SET standards for HC and CO will require manufacturers to significantly
reduce the frequency of fuel enrichment events, yet allow for some necessary catalyst protection
and power enrichment operation. We are applying the same numeric values for FTP and SET
duty cycles for HC and NOx standards. We are remaining generally consistent with a fuel neutral
approach in the final FTP and SET standards, with the exception of CO for Spark-ignition HDE
over the new SET duty cycle. These new SET standards are summarized in Table 3-38.
150
-------
120
100
CL
O)
u_
o
u*>
E.
-------
washcoats and other related exhaust components have progressed in recent light-duty
applications and are able to tolerate significantly higher exhaust gas temperatures while still
achieving acceptable component durability and catalytic deterioration targets. The use of these
improved materials, along with the more robust temperature models or temperature measurement
devices discussed in Chapter should result in significant reductions in CO emissions and allow
engines to meet the new emission standards. Also discussed in Chapter 2.2.1.7 is the use of
engine down speeding, which can avoid the high speed, high exhaust gas temperature conditions
that typically result in fuel enrichment due to engine component durability and catalyst thermal
concerns. With the integration of modern multi-speed electronically controlled transmissions,
this down speeding approach is extremely feasible and likely to also reduce engine wear and
improve fuel consumption with little perceived effect on performance under commercial and
vocational operation. Note that in order to meet GHG and fuel consumption goals, this engine
has already implemented some degree of down speeding as evident in the reduced maximum test
speeds reported by one manufacturer. The agency believes that the more recent introduction of
10-speed transmissions provides additional opportunities for down speeding that have not yet
been explored.
Finally, the third mode of operation, where the ECM maintains a stoichiometric fuel-to-air
ratio throughout all of the SET cycle test points, and potentially, under high load real world
operation as well, results in the greatest degree of emission control. Under stoichiometric
operation, NMHC, CO, and NOx emissions are simultaneously reduced, as the three-way
catalyst can be optimized to reduce all three pollutants. This strategy is discussed in chapter
2.2.1.6, and our analysis of MY 2019 certification data indicates that NMHC and CO emissions
are well below the new SET standards and NOX emissions meet the new SET MY 2027 and
later standards. This level of NOX control was achieved without any improvements or
refinements to the calibration and control strategies that we believe manufacturers will utilize to
meet the new FTP standards. As observed in the analysis in Table 3-39, a slight drop in power is
observed at the three SET test points as fuel enrichment decreases, however, this slight power
loss is also accompanied by a noticeable decrease in fuel consumption, which can be a
potentially important operational cost benefit in a commercial vehicle application. Similar to the
previous discussion, the agency believes that several engine hardware and control technologies,
as well as the additional gear ratios in current transmission designs, will provide the opportunity
for maintaining stoichiometric fuel-air control under all load and speed conditions.
Table 3-39 SET Operation Mode Power Comparison
Power (kW)
Torque (Nm)
SET Set Points
SET Set Points
A
B
C
A
B
C
Power Enrichment Allowed
211
187
145
546
572
547
Catalyst Protection with
No Power Enrichment
211
182
141
542
554
524
Stoichiometric Operation
201
179
137
522
551
526
3.2.2.3 Spark-Ignition Technology Demonstration Program
EPA initiated a program with Southwest Research Institute to better understand the emissions
performance limitations of current heavy-duty SI engines as well as investigate the feasibility of
152
-------
advanced three-way catalyst aftertreatment and technologies and strategies to meet new exhaust
emission standards21. In addition to investigating emission performance on the FTP duty cycle,
the test program evaluated the SET duty cycle that is now required for certification. This section
describes the results of the SI demonstration program. See Chapter 1.2 for an expanded
description of these and other technologies and strategies to address exhaust emissions for HD SI
engines.
A MY 2019-certified heavy-duty gasoline engine was used for this evaluation. This particular
engine was chosen because it represented the newest design among the three most-common
engines in the market and included technologies not normally found in HD SI engines, such as
variable valve timing (VVT) and cooled EGR. Additional considerations for selecting this engine
were the availability of chassis-certified trucks with the options and driveline configuration
desired, as well as the ability to install and operate the engine in a dynamometer test cell. Table
3-40 describes the HD SI engine that was used for this evaluation.
Table 3-40 Major engine specifications of the MY2019 HD SI gasoline engine used for the EPA demonstration
program
Engine Component
Specification
Engine Displacement (L)
6.4
Configuration / Type
90° PushrodV-8
Bore (mm)
103.9
Stroke (mm)
94.8
Aspiration
Naturally aspirated
Injection
Sequential multi-port fuel injected
Compression Ratio
10.0:1
Engine Block Material
Cast Iron
Cylinder Head Material
Cast Aluminum, Hemispherical combustion chamber
Valve Train
2 valve per cylinder, Cam-in-block, VVT, Hydraulic roller lifters
Ignition
8 individual coils, 16 spark plugs, 2 per cylinder
Exhaust Gas Recirculation
Cooled EGR
Fuel Requirement
89 Octane recommended
Peak Horsepower
360 HP (a), 4715 rpm
Peak Torque
408 lb-ft (S)7 4000 rpm
This program includes a baseline evaluation of emissions performance as well as a
demonstration of the reductions possible through the application of advanced catalyst designs
that included decreased substrate wall thickness and increased cell density, a washcoat
formulation that is more tolerant of high exhaust gas temperatures, and forward placement of the
catalyst substrate (i.e., moving a portion of the total catalyst volume closer to the engine). The
catalysts were artificially aged to represent performance equivalent of 250,000 miles of real-
world operation in a manner approved by the engine manufacturer.
We also investigated the impact of engine down-speeding and calibration changes to
demonstrate further emission reduction potential of both the baseline and advanced catalyst
configurations on the FTP and SET. As noted in Chapter 1.2, this engine down-speeding strategy
is currently used by at least one HD gasoline engine manufacturer and this lower speed is made
possible by transmission strategies preventing over-speeding, which allows the emission controls
to operate in a much more desirable and lower emitting area of engine operation. For the down-
153
-------
speed testing in our demonstration program, the maximum test speed (MTS) was lowered from
the manufacturer's stated MTS of 4715 rpm to 4000 rpm.
Finally, engine calibration parameters affecting air-fuel enrichments and biasing (lambda <
1.0) were manipulated to further reduce CO emissions on the SET. Because of the limited
abilities of aftermarket vehicle engine control module programmers and the complexity of OEM
engine calibrations and control strategies, this effort was met with limited success. We do
however believe that manufacturers with access to the latest tools for calibration and complete
access to engine control strategies and calibrations will be able to optimize lambda biasing as
well as any necessary air-fuel enrichments for catalyst and engine protection. With the
capabilities previously mentioned we believe that through a combination of engine down-
speeding and calibration optimization the final emissions standards are achievable.
Installation of the engine in the test cell included instrumenting the engine's aftertreatment
with thermocouples at exhaust manifold exit, catalyst inlet, and 1-inch rearward of the catalyst
front face. For all engine tests, Controller Area Network (CAN) data from the engine control
module, including, but not limited to engine speed, short and long-term fuel correction, and spark
advance were recorded.
We evaluated the following test procedures, performing three repeats of each cycle:
• HD SET (40 CFR 1036.510)
• HD SI FTP cycle (40 CFR 1036.512(a)(1))
• Engine mapping (40 CFR 1036.535 and 1036.540)
In all tests, we measured NOx, CO, PM, and NMHC, as well as the GHG-related parameters
of brake-specific fuel consumption (BSFC), CH4, and CO2. Emissions were measured from two
locations throughout each test cycle: before the catalytic aftertreatment and at the tailpipe.
Table 3-41 and Table 3-42 present results representing three technology packages over the
FTP and SET duty cycles, respectively: advanced catalyst technology, advanced catalyst with
engine down-speeding, and a combination of advanced catalyst, engine down-speeding, and
calibration. Over both duty cycles, the results show NOx and NMHC to be at or below the final
MY 2027 standards. For the FTP results in Table 3-41, CO is below the final MY 2027 standards
with the advanced catalyst alone and further reduced by downspeeding to 4000 rpm MTS. Figure
3-23 illustrates the CO breakthrough associated with the 4715 MTS. The lambda excursions seen
in Figure 3-23 are a direct result of catalyst protection lambda enrichment specifically associated
with higher engine speed operation as observed in Table 3-41 below. Please see the discussion in
Chapter 1.2 regarding engine operating modes and possible calibration philosophy to address
excess CO emissions. We applied one set of calibration changes to create a richer lambda bias
and avoid throttle-based enrichment that led to higher SET CO levels. Those calibration changes
resulted in slight increases in NOx and CO over the FTP compared to the unmodified calibration,
but the emission levels remained below the final standards.
154
-------
Table 3-41 Spark-Ignition Demonstration Program FTP Results
NOx
CO
NMHC
PM
BSFC
(mg/hp-hr)
(g/hp-hr)
(mg/hp-hr)
(mg/hp-hr)
(lb/hp-hr)
New Standards
35
6.0
60
MY 2027 and later
3
250k Catalysts
4715 RPMMTS
19
4.9
32
4.8
0.456
250k Catalysts
4000 RPMMTS
18
0.25
35
4.5
0.448
250k Catalysts
4000 RPMMTS
21
0.99
1
4.4
0.448
Modified Cal
1.5
~a
_Q
0.5
4715 rpm Vs. 4000 rpm MTS Comparison
FTP CO Reduction
Nffili 1
5
.
1
L ...L
30000
20000
200
400
600
Time (seconds)
800
1000
1200
DO
CD
oc
ioooo IS
o
u
•4715 rpm Lambda 4000 rpm Lambda
•4175 rpm CO 4000 rpm CO
Figure 3-23: Engine RPM Down-Speeding FTP CO Comparison
Table 3-42 compares the emissions results over the SET cycle at MTSs of 4715 rpm and 4000
rpm. Like the FTP results NMHC and NOx remained low for all three technology packages.
Unlike the FTP results, the advanced catalyst alone did not reduce CO below the new standard.
Although CO at 4000 rpm MTS meets the new standard, calibration changes, described above,
were applied to attempt further reducing the CO. The calibration changes effectively reduced
NOx, but did not have the desired effect of reducing CO. We note that some of the catalyst
separated from the mat during the 4715 rpm test. The catalyst continued to control NOx, CO, and
NMHC, but some of the separated material was captured by the particulate filter resulting in the
7 mg/hp-hr PM measurement shown in Table 3-42. In the absence of the separated catalyst
material, we expect the PM level would be below the 5 mg/hp-hr standard we are finalizing.
155
-------
Table 3-42 Spark-Ignition Demonstration Program SET Results
NOx
CO
NMHC
PM
BSFC
(m g/hp-hr)
(g/hp-hr)
(mg/hp-hr)
(mg/hp-hr)
(lb/hp-hr)
New Standards
35
14.4
60
MY 2027and later
3
250k Catalysts
4715 RPM MTS
8
36.7
6
7
0.462
250k Catalysts
4000 RPM MTS
5
7.21
1
3
0.437
250k Catalysts
4000 RPM MTS
1
9.65
1
3
0.438
Modified Cal
1.4
1.2
1
¦a
-Q
0.8
E
m
0.6
1
0.4
0.2
0
4715 rpm Vs. 4000 rpm MTS Comparison
SET CO Reduction
09—"-jhs-
n.
li
i
ii
100000
80000
60000 iB
1000 2000 3000 4000
Time (seconds)
5000
6000
DO
40000
20000
ro
oc
10
10
ro
O
U
¦4715 rpm Lambda 4000 rpm Lambda
4175 rpm CO 4000 rpm CO
Figure 3-24: Engine RPM Down-Speeding SET CO Comparison
As mentioned previously, SwRI undertook a calibration effort to address the level of CO
witnessed, 7.21 g/hp-hr, during SET testing and lower NOx for both the FTP and SET. Because
of the limited abilities of aftermarket vehicle engine control module programmers and the
complexity of OEM engine calibrations and control strategies, this effort was met with limited
success. We do however believe that manufacturers with access to the latest tools for calibration
and complete access to engine control strategies and calibrations will be able to optimize lambda
biasing as well as any necessary air-fuel enrichments for catalyst and engine protection. With the
capabilities previously mentioned we believe that through a combination of engine down-
speeding and calibration optimization the SET CO emission standard of 14.4 g/hp-hr is
achievable.
156
-------
3.2.2.4 Refueling Emissions Technology Effectiveness
As described Chapter 2.3.2.4 of this RIA, HD SI engines certified as incomplete heavy-duty
vehicles are not currently required to meet ORVR. The technology package we considered for
these engines is based on the technologies implemented by chassis-certified complete vehicles to
meet the evaporative and refueling requirements of Tier 3. The technology package includes four
main equipment components and strategies that incomplete heavy-duty vehicles may need to
update to implement ORVR: increased working capacity of the carbon canister to handle
additional vapors volumes, flow control valves to manage vapor flow pathway during refueling,
filler pipe and seal to prevent vapors from escaping, and the purge system and management of
the additional stored fuel vapors. Chapter 1.2.3 includes descriptions of these technologies. The
assumptions we applied to account for the larger fuel tanks and other considerations for larger
incomplete vehicles are summarized in Chapter 3.2.3.2 where we present our projected direct
manufacturing costs.
The final refueling standards are projected to result in 27.8% lower VOC and Benzene by
2030, 80.2% lower by 2040 and 88.5% lower by 2045 for heavy duty gasoline vehicles over
14,000 lbs. See the discussion and results in Chapter 5.3.
3.2.3 Estimated Direct Manufacturing Costs for Technology Packages Evaluated
For this analysis of the aftertreatment costs, heavy-duty spark-ignition (HD SI) engines are
categorized by the type of fuel they use: liquid fuels (i.e., gasoline, gasoline-ethanol blends, and
ethanol) or gaseous fuels (i.e., compressed natural gas and liquified petroleum gas). The gaseous-
fueled category includes engines derived from SI platforms and engines converted to SI from
heavy-duty CI engines. The heavy-duty SI engine category is further divided into heavy heavy-
duty (HHD) and urban bus. We projected the costs of achieving the HD SI engine exhaust
emission standards based on the technologies we evaluated in our demonstration program (see
Chapter 3.2.2.3).
3.2.3.1 Spark Ignition Exhaust Aftertreatment System Direct Manufacturing Cost Analysis
Manufacturers will optimize the design of their aftertreatment systems specific to their
different vehicles. Manufacturers' 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 heavy-duty emission standards and meet the
durability criteria for all vehicle classes.
Similar to the CI engine cost analysis, costs for baseline HD SI engine aftertreatment systems
were estimated using cost data published by Dallman et al.22, Pasoda et al.23 as well as data from
manufacturer's technical descriptions of aftertreatment catalyst components submitted as part of
engine certification packages for MY 2019. Manufacturer's data were then combined into
projected sales-weighted averages by type of fuel (liquid and gaseous fuels), including two
categories for gaseous-fueled engines identified as heavy heavy-duty and urban bus that have
157
-------
distinctly different aftertreatment demands. Costs in this study are driven by the catalyst and
precious metal loading. No significant labor costs were identified. The direct manufacturing
cost for these technology packages is equal to the catalyst piece cost.
Baseline projected sales-weighted average engine displacements, catalyst volumes, PGM
loadings and costs are shown in Table 3-43 for both liquid and gaseous fueled SI engines.E As
mentioned previously, these are based on certification data from MY 2019. These MY 2019
engine and aftertreatment costs estimates are used as the MY 2027 baseline cost presented in
RIA Chapter 3.2.3.1 after conversion to 2021 dollars.
Table 3-43: 2019 MY Sales-Weighted Baseline SI Engine Direct Manufacturing Costs (2021$)
Liquid Fueled
SI Engine
Gaseous Fueled
SI Engine
Engine Displacement (L)
6.6
7.2
Total TWC Volume (L)
4.3
5.1
CATv/ENGd Ratio (L/L)
0.65
0.72
Total Pt ($)
$0
$0
Total Pd ($)
$140
$178
Total Rh ($)
$2,371
$2,214
Substrate cost ($)
$56
$68
Washcoat cost ($)
$26
$31
Canning cost ($)
$15
$19
Total direct manufacturing cost
($2019)
$2,371
$2,510
We separately evaluated two distinct categories for the three gaseous-fueled HD SI engines
that certified to California's optional and more stringent 0.02 g/hp-hr NOx standard in MY 2019:
HHD and urban bus. One engine is derived from a traditional SI gasoline-fueled engine and two
are converted from CI diesel-fueled engines. Given the small number of engines in each of these
categories, we are not publicly releasing the component-level details for MY 2019 HD SI, HHD
and urban bus engines, but summarize the total costs for these categories. Table 3-44 shows the
baseline aftertreatment cost for HHD and urban bus gaseous-fueled engines.
Table 3-44: 2019 MY HHD and Urban Bus Gaseous-fueled Direct Manufacturing Baseline Costs (2021$)
Gaseous Fueled
SI Engine
Gaseous Fueled
HHD Engine
Gaseous Fueled
Urban Bus Engine
Engine Displacement (L)
6.8
11.9
8.9
Total direct manufacturing
cost ($2019)
$5,770
$8,474
$6,356
As mentioned previously, the three MY2019 gaseous-fueled HD SI, HHD and urban bus
engines currently meet a 0.02 g/hp-hr NOx standard, and we assumed that additional technology
would not be needed for these engines to meet the standards in future model years. However, it is
E PGM costs were determined by taking the average price across all regions from 8/31/2020 to 8/31/2022 from
Johnson Matthey's PGM management website, https://matthev.com/products-and-markets/pgms-and-
circularitv/pgm-management/
158
-------
reasonable to believe that improvements in the materials and design of the catalyst substrate
support structure (e.g., can material, mat, seals, etc.) will be needed to achieve durability over the
longer useful life and we estimated a nominal addition per-engine cost for these engine
categories.
For the other gaseous fueled engines category, we assumed the same technologies would be
used to meet the MY2027 and later standards. Table 3-45 shows the MY 2027 and later
gaseous-fueled engine direct manufacturing costs adjusted for improved catalyst and component
durability.
Table 3-45: Projected Gaseous Fueled Engine Direct Manufacturing Cost (2021$)
Gaseous Fueled
SI Engine
Gaseous Fueled
HHD Engine
Gaseous Fueled
Urban Bus Engine
MY 2027 and later total
direct manufacturing cost
($2019)
$5,770
$8,474
$6,356
The MY 2027 technology cost for the liquid fueled SI engines are based on the demonstration
engine described in Chapter 3.2.2.3. Costs were estimated using the same Dallman et al.24 and
Pasoda et al.25 data as our baseline estimates and data from the specific aftertreatment catalyst
components used for the HD SI demonstration program. We did not make any specific cost
adjustments to account for the lengthened useful life, since the aftertreatment system used in the
demonstration program represented catalysts aged to 250,000 miles. Table 3-46 contains the
details of this analysis.
159
-------
Table 3-46: Projected Liquid Fueled SI Engine Piece Cost (2021$)
Technology Description
MY 2027 and later
Liquid Fueled SI Engine
Total TWC Volume (L)
5.8
CATv/ENGd Ratio (L/L)
0.91
Light-off Catalyst
Number of Catalysts
2
L.O. Catalyst Volume (L)
0.82
Total Pt ($)
$0.0
Total Pd ($)
$707
Total Rh ($)
$496
Substrate Cost ($)
$21
Washcoat Cost ($)
$10
Canning Cost ($)
$6
Underfloor Catalyst
Number of Catalyst
2
U.F. Catalyst Volume (L)
2.1
Total Pt ($)
$0
Total Pd ($)
$494
Total Rh ($)
$1,270
Substrate Cost ($)
$55
Washcoat Cost ($)
$25
Canning Cost ($)
$18
Total Demonstration TWC Cost ($2019)
S3,101
Table 3-47 summarizes the costs for each of the HD SI engine categories evaluated in this
analysis. These direct manufacturing costs are used in the analysis to determine the overall costs
of the program, as detailed in Chapter 7 of this RIA.
Table 3-47: Summary of HD SI Engine Direct Manufacturing Cost Comparison
Cost Packages (2021$)
Liquid Fueled
SI Engine
Gaseous Fueled
SI Engine
SI HHP
SI Urban Bus
Baseline Technology
$2,371
$2,510
$8,447
$6,335
MY 2027 Technology
$3,101
$5,770
$8,474
$6,336
MY 2027 Incremental
$730
$3,260
$28
$21
3.2.3.2 Onboard Refueling Vapor Recovery Anticipated Costs
As described in Chapter 2.3.2.4 of this RIA, HD SI engines certified as incomplete heavy-
duty vehicles are not currently required to meet ORVR. There are four main equipment
components and strategies incomplete heavy-duty vehicles need to update to implement ORVR:
increased working capacity of the carbon canister to handle additional vapors volumes, flow
control valves to manage vapor flow pathway during refueling, filler pipe and seal to prevent
vapors from escaping, and the purge system and management of the additional stored fuel
vapors. Chapter 1.2.3 includes more information on these technologies. The associated direct
manufacturing costs for these updates are summarized below. No labor cost was identified so the
direct manufacturing cost is equal to the piece cost plus tooling cost (per piece). ORVR
requirements will be extended to heavy-duty gasoline engines in incomplete vehicles starting in
160
-------
model year 2027. For our cost analysis, we assumed all heavy-duty gasoline engines that are
identified as LHD, MHD and HHD in MOVES will have an average of a 70-gallon fuel tank.
Capturing the increased vapor volume from the vapor displaced during a refueling event will
require canisters to increase vapor or "working" capacity approximately 15%-40% depending on
the individual vehicle systems (i.e., fuel tank size). This can be achieved by increasing the
canister volume using conventional carbon, the fundamental material used to store fuel vapors. A
typical Tier 3 canister has approximately 5.1 liters of conventional carbon to capture overnight
diurnal evaporative emissions for a 70-gallon fuel tank. An increase in required capacity to allow
refueling vapors to be captured results in the need for an additional 1.9 liters of conventional
carbon. A change in canister volume to accommodate additional carbon includes increased costs
for retooling and additional canister plastic material, as well as design considerations to fit the
larger canister on the vehicle.
An alternative to retooling for a larger single canister would be to add a second canister for
the extra canister volume to avoid the re-tooling costs. Several smaller volume canisters are
available on the market today. Another approach, based on discussions with canister and carbon
manufacturers, can be achieved by using a higher adsorption carbon along with modifications to
compartmentalization within the existing canister plastic shell that will increase the canister
working capacity without requiring a larger canister size.
Additionally, there are two primary technologies used to prevent vapors from escaping into
the atmosphere through the filler neck and around the fuel nozzle area when the vehicle is
refueling that can affect the canister vapor capacity design requirements: a mechanical seal
which makes direct physical contact with the refueling nozzle to create a nozzle to filler neck
seal; or a liquid seal further down in the filler pipe which uses the liquid fuel mass flowing down
the filler pipe and entering the tank to hydraulically prevent vapors from migrating back up the
fill pipe. There is approximately a 20% reduction in carbon volume required if a mechanical seal
is used at the filler neck versus a liquid seal approach. While mechanical seals are not currently
the preferred technology, manufacturers facing the choices available for the larger volume fuel
tanks and the need for a larger matching carbon containing canister to handle these large
quantities of fuel vapors, may opt for more a mechanical seal design to avoid excess canister
carbon requirements and possible retooling charges. We share our assumptions and cost
estimates for both seal options in Table 3-48 and Table 3-49. A mechanical seal approach costs
approximately $10.00 per seal. A dual tank may require two seals if dual filler necks are used
instead of a single filler neck and transfer pump to move fuel between the two tanks.
The second required equipment update would be to install flow control valves, which may be
integrated into existing roll-over/vapor lines. The flow control valves are needed to manage the
vapors during the refueling event by providing a low restriction pathway for vapors to enter the
canister for adsorption and storage on the carbon materials. We anticipate vehicles would require
on average one valve per vehicle which would be approximately $6.50 per valve. A dual tank
system may require a flow control valve system per tank depending on the design approach.
Thirdly, as mentioned above, a filler pipe and seal system would be needed for each filler
nozzle to keep the vapors contained during refueling. Manufacturers have the option of a
mechanical seal that costs approximately $10.00 per seal, or a liquid seal which in itself costs
nothing but requires approximately $15 of new hardware modifications to provide enough back
pressure to stop the refueling nozzle fuel flow when tank reaches full capacity.
161
-------
Lastly, the engine control of the canister purge rates would need to be addressed. This update
will include calibration improvements and potentially additional hardware to ensure adequate
purge volumes are achieved as required to maintain an appropriate canister state to manage
vapors generated during diurnal and subsequent refueling events. If required for a dual tank
system, an extra purge valve may be needed if the two-tank system maintains independent
canisters instead of a single common cannister as observed in dual-tank, single canister light-
duty applications.
Table 3-48 shows our calculations estimating the amount of extra canister size for
conventional carbon for a 70-gallon tank, using Tier 3 requirements as a baseline. Currently
under Tier 3 requirements the canister and purge strategy are sized for the diurnal test and
designed to meet the Bleed Emissions Test Procedure (BETP) requirements. During the diurnal
test, the canister is loaded with hydrocarbons over two or three days, allowing the hydrocarbons
to load a conventional carbon canister (1500 GWC, gasoline working capacity) at a 70 g/L
efficacy. During a refueling event, which takes place over a few minutes, the vapor from the gas
tank is quickly loaded onto the carbon in the canister with an ORVR system, causing the
efficiency of the canister loading to drop to 50 g/L efficacy mainly because of the high volume
of fuel vapors and the composition of those vapors required to be adsorbed in the short period of
a refueling event. Typically, a design safety margin adds an extra 10% carbon to ensure adequate
performance over the life of the system. Therefore, even though there is typically less fuel vapor
mass generated and managed during a refueling event than is generated over a three-day diurnal
time period, the amount of carbon that is necessary to contain the vapor is higher for a refueling
event.
In order for carbon in the canisters to be effective at managing vapors for diurnals and
refueling events, the vehicle engine must sufficiently purge the canister during engine operation
in preparation for future events that will require vapor adsorbing capacity. The purge
requirements are shown in Table 3-48. The diurnal drive cycle is only 30 minutes and targets 200
bed volumes of purge to clean the canister before the evaporative emissions test. When the bed
volumes of purge are multiplied by the canister volume, the total purge volume can be
calculated. The total purge volume divided by the number of minutes driving gives us the
average purge rate. An ORVR test requires proper conditioning for a very clean canister in order
to pass the ORVR test. To clean out the canister over the 97 minutes of driving cycles for the
ORVR prep, a much higher amount of bed volumes is necessary; therefore, the purge rate
required is also higher. Table 3-49 shows cost estimations for the different approaches. For our
direct manufacturing cost we used $25 (2019 dollars), which is the average of all approaches
considered, as the cost estimate for the additional canister capacity and hardware to meet the
refueling standard. These costs were converted to 2021 dollars as described in the cost analysis
of Chapter 7.
162
-------
Table 3-48: Assumptions for gasoline-fueled heavy-duty spark-ignition vehicles for conventional carbon
requirements to meet the refueling standard
Tier 3 Baseline
for Evaporative
Standards
Updates for Refueling Standards
Mechanical Seal
Liquid Seal
Diurnal
ORVR
Diurnal Heat Build
72-96°F
80°F
RVP
9 psi
Nominal Tank Volume
70 gallons
Fill Volume
40%
10% to 100%
Air Ingestion Rate
0%
13.50%
Mass Vented per heat build, g/day
120
Mass Vented per refueling event
255
315
Hot Soak Vapor Load
5
Mass Vented over 48-hour test
227.2
Mass Vented over 72-hour test
323.3
1500 GWC, g/L a
70
50
50
Excess Capacity
10%
10%
10%
Canister Volume, liters'3
4 8-hour
3.6
72-hour
5.1
ORVR0
5.6
6.9
Limiting Drive Cycle, minutes
30
97
97
Bed Volumes Purge
200
646
646
Total Purge Volume, liters'1
1020
3618
4457
Average Purge Rate, LPMe
34
37
46
BETP Purge
37
46
a Storage capability of conventional carbon
b Canister Volume = l.l(mass vented)/ 1500 GWC (Efficiency)
0 ORVR adds .5 liters and 1.8 liters for Mechanical Seal and Liquid Seal respectively
d Total Purge Volume, liters = canister volume, liters * Bed Volumes Purge represent the potential volume of
purge for the 97 minute drive cycle used for the ORVR test procedure. Required purge volume to clean out the
canister of fuel vapors for the larger ORVR canisters is likely much lower, approximately the ratio of new canister
volume to the previous canister volume multiplied by the target bed volumes (220 bed volumes for a mechanical
seal and 271 bed volumes for a liquid seal)
e Average Purge Rate, LPM = Total Purge Volume, liters / Limiting Drive Cycle, minutes however as noted in (d),
this is not necessarily the required purge volumes or rates
163
-------
Table 3-49: Estimated Direct Manufacturing Costs for ORVR Over Tier 3 as Baseline
Liquid Seal
Mechanical Seal
New Canister
Dual Existing
Canisters in Series
New Canister
Dual Existing
Canisters in Series
Additional Canister
Costs
$20
$15
$8
$8
Additional Tooling3
$0.50
$0.50
Flow Control Valves
$6.50
$6.50
Seal
$0
$0
$10
Totalb
$27
$22
$25
a Assumes the retooling costs will be spread over a five-year period
b Possible additional hardware for spitback requirements
Chapter 3 References
1 Neely, G., Sharp, C., Pieczko, M., and McCarthy, J., (2020) Simultaneous NOX and C02
Reduction for Meeting Future CARB Standards Using a Heavy-Duty Diesel CDA-NVH
Strategy. SAEInt. J. Engines 13(2).
2 Morris, A., & McCarthy, J. (2020). The Effect of Heavy-Duty Diesel Cylinder Deactivation on
Exhaust Temperature, Fuel Consumption, and Turbocharger Performance up to 3 bar BMEP.
SAE Technical Paper No. 2020-01-1407.
3 Zavala, B., Sharp, C., Neely, G., & Rao, S. (2020). CARB Low NO X Stage 3 Program-
Aftertreatment Evaluation and Down Selection. SAE Technical Paper No. 2020-01-1402.
4 Rao, S., Sarlashkar, J., Rengarajan, S., Sharp, C., & Neely, G. (2020). A Controls Overview on
Achieving Ultra-Low NOx SAE Technical Paper No. 2020-01-1404.
5 Neely, G. D., Sharp, C., & Rao, S. (2020). CARB Low NOx Stage 3 Program-Modified Engine
Calibration and Hardware Evaluations. SAE Technical Paper No. 2020-01-0318.
6 Sharp, C. (2020). Further Development and Validation of Technologies to Lower Oxides of
Nitrogen Emissions from Heavy-Duty Vehicles. Southwest Research Institute. Final Report
CARB Contract 16MSC-010.
7 Eakle, S., & Bartley, G. (2014). The DAAAC Protocol™ for Diesel Aftertreatment System
Accelerated Aging. In Proceedings of the Emissions 2014 Conference.
8 Sharp, C. (2022). Demonstration of Low NOx Technologies and Assessment of Low NOx
Measurements in Support of EPA's 2027 Heavy Duty Rulemaking. Southwest Research
Institute. Final Report EPA Contract 68HERC20D0014.
9 Nagar, P., Szailer, T., & Webb, C. (2013). Comparison of SCR Catalyst Performance on RMC
SET Emission Cycle between an Engine and a High Flow Burner Rig. SAE Technical Paper No.
2013-01-1070.
164
-------
10 Sanchez, James. Memorandum to docket EPA-HQ-OAR-2019-0055. Test Results from
System A Demonstration. October 31, 2022.
11 Schenk, Charles. Memorandum to docket EPA-HQ-OAR-2019-0055. Evaluation of Vibration
in a Heavy-Duty Tractor Induced by Cylinder Deactivation. September 30, 2022.
12 Mamidanna, S. 2021. Heavy-Duty Engine Valvetrain Technology Cost Assessment. U.S.
EPA Contract with FEV North America, Inc., Contract No. 68HERC19D0008, Task Order No.
68HERH20F0041 .Submitted to the Docket.
13 Mamidanna, S. 2021. Heavy-Duty Engine Valvetrain Technology Cost Assessment - Peer
Review Responses. Contract No. 68HERC19D0008, Task Order No. 68HERH20F0041,
Submitted to the Docket.
14 Eastern Research Group, Inc. 2021. "External Peer Review of Heavy-Duty Engine Valvetrain
Technology Cost Assessment". Contract No. 68HE0C18C0001, Work Assignment 2-05.
Submitted to the Docket
15 Mamidanna, S. 2021. Heavy-Duty Vehicles Aftertreatment Systems Cost Assessment.
Submitted to the Docket.
16 McDonald, J. 2022. Heavy-Duty Vehicle Exhaust Aftertreatment Systems Costs, Calculations,
and Updates. Memo to the Docket.
17 69 Federal Register at 39126. June 29, 2004. See https://www.govinfo.gov/content/pkg/FR-
2004-06-29/pdf/04-11293.pdf.
18 Mitchell, George, "EPA's Medium Heavy-Duty Gasoline Vehicle Emissions Investigation".
February 2019.
19 U.S. EPA. "Heavy-Duty Highway Gasoline and Diesel Certification Data (Model Years:
2015 - Present)". Available online: https://www.epa.gov/system/files/documents/2022-02/heavy-
duty-gas-and-diesel-engines-2015-present.xlsx Accessed October 2022.
20 U.S. EPA. Annual Certification Data for Vehicles, Engines, and Equipment. Available online:
https://www.epa.gov/compliance-and-fuel-economy-data/annual-certification-data-vehicles-
engines-and-equipment. Accessed October 2022.
21 Ross, M. (2022). Heavy-Duty Gasoline Engine Low NOx Demonstration. Southwest Research
Institute. Final Report EPA Contract 68HERC20D0014.
22 Dallman, T., Posada, F., Bandivadekar, A. 2018. "Costs of Emission Reduction Technology
for Diesel Engines Used in Non-Road Vehicles and Equipment." International Council on Clean
Transportation.
23 Pasoda Sanchez, F., Bandivadekar, A., German, J. 2012. "Estimated Cost of Emissions
Reduction Technolgies for Liight-Duty Vehicles." International Council on Clean
Transportation.
165
-------
24 Dallman, T., Posada, F., Bandivadekar, A. 2018. "Costs of Emission Reduction Technology
for Diesel Engines Used in Non-Road Vehicles and Equipment." International Council on Clean
Transportation.
25 Pasoda Sanchez, F., Bandivadekar, A., German, J. 2012. "Estimated Cost of Emissions
Reduction Technolgies for Liight-Duty Vehicles." International Council on Clean
Transportation.
166
-------
167
-------
Chapter 4 Health and Environmental Impacts
4.1 Health Effects Associated with Exposure to Pollutants
Heavy duty vehicles emit pollutants that contribute to ambient concentrations of ozone, PM,
NO2, CO, and air toxics. A discussion of the health effects associated with exposure to these
pollutants is presented in this section of the RIA. The following discussion of health impacts is
mainly focused on describing the effects of air pollution on the population in general.
Additionally, because children have increased vulnerability and susceptibility for adverse
health effects related to air pollution exposures, EPA's findings regarding adverse effects for
children related to exposure to pollutants that are impacted by this rule are noted in this section.
The increased vulnerability and susceptibility of children to air pollution exposures may arise
because infants and children generally breathe more relative to their size than adults do, and
consequently may be exposed to relatively higher amounts of air pollution.1 Children also tend to
breathe through their mouths more than adults and their nasal passages are less effective at
removing pollutants, which leads to greater lung deposition of some pollutants, such as PM.2'3
Furthermore, air pollutants may pose health risks specific to children because children's bodies
are still developing.A For example, during periods of rapid growth such as fetal development,
infancy, and puberty, their developing systems and organs may be more easily harmed.4'5 EPA's
America's Children and the Environment is a tool which presents national trends on air
pollutants and other contaminants and environmental health of children.6
4,1.1 Ozone
4.1.1.1 Background on Ozone
Ground-level ozone pollution forms in areas with high concentrations of ambient nitrogen
oxides (NOx) and volatile organic compounds (VOCs) when solar radiation is high. Major U.S.
sources of NOx are highway and nonroad motor vehicles and engines, power plants, and other
industrial sources, with natural sources, such as soil, vegetation, and lightning, serving as smaller
sources. Vegetation is the dominant source of VOCs in the U.S. Volatile consumer and
commercial products, such as propellants and solvents, highway and nonroad vehicles, engines,
fires, and industrial sources also contribute to the atmospheric burden of VOCs at ground-level.
The processes underlying ozone formation, transport, and accumulation are complex.
Ground-level ozone is produced and destroyed by an interwoven network of free radical
reactions involving the hydroxyl radical (OH), NO, NO2, and complex reaction intermediates
derived from VOCs. Many of these reactions are sensitive to temperature and available sunlight.
High ozone events most often occur when ambient temperatures and sunlight intensities remain
high for several days under stagnant conditions. Ozone and its precursors can also be transported
hundreds of miles downwind, which can lead to elevated ozone levels in areas with otherwise
low VOC or NOx emissions. As an air mass moves and is exposed to changing ambient
A Children's environmental health includes conception, infancy, early childhood and through adolescence until 21
years of age as described in the EPA Memorandum: Issuance of EPA's 2021 Policy on Children's Health. October
5, 2021. Available at https://www.epa.gov/system/files/documents/2021-10/2021-policy-on-childrens-health.pdf.
168
-------
concentrations of NOx and VOCs, the ozone photochemical regime (relative sensitivity of ozone
formation to NOx and VOC emissions) can change.
When ambient VOC concentrations are high, comparatively small amounts of NOx catalyze
rapid ozone formation. Without available NOx, ground-level ozone production is severely
limited, and VOC reductions would have little impact on ozone concentrations. Photochemistry
under these conditions is said to be "NOx-limited." When NOx levels are sufficiently high, faster
NO2 oxidation consumes more radicals, dampening ozone production. Under these "VOC-
limited" conditions (also referred to as "NOx-saturated" conditions), VOC reductions are
effective in reducing ozone, and NOx can react directly with ozone resulting in suppressed ozone
concentrations near NOx emission sources. Under these NOx-saturated conditions, NOx
reductions can actually increase local ozone under certain circumstances, but overall ozone
production (considering downwind formation) decreases. Even in VOC-limited areas, NOx
reductions are not expected to increase ozone levels if the NOx reductions are sufficiently large -
large enough to become NOx-limited.
4.1.1.2 Health Effects Associated with Exposure to Ozone
This section provides a summary of the health effects associated with exposure to ambient
concentrations of ozone.B The information in this section is based on the information and
conclusions in the April 2020 Integrated Science Assessment for Ozone (Ozone ISA).7 The
Ozone ISA concludes that human exposures to ambient concentrations of ozone are associated
with a number of adverse health effects and characterizes the weight of evidence for these health
effects. c 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 exposure. It also
concludes that metabolic effects, including metabolic syndrome (i.e., changes in insulin or
glucose levels, cholesterol levels, obesity and blood pressure) and complications due to diabetes
are likely to be causally associated with short-term exposure to ozone. The evidence is also
suggestive of a causal relationship between short-term exposure to ozone and cardiovascular
effects, central nervous system effects, and total mortality.
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 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, metabolic effects,
reproductive and developmental effects, central nervous system effects and total mortality. The
B 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 breathing route and rate.
c 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 ISA.
169
-------
evidence is inadequate to infer a causal relationship between chronic ozone exposure and
increased risk of 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. In addition, some
groups are at increased risk of exposure due to their activities, such as outdoor workers and
children. 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.D Children also have a higher asthma prevalence compared to
adults. Recent epidemiologic studies provide generally consistent evidence that long-term ozone
exposure is associated with the development of asthma in children. Studies comparing age
groups reported higher magnitude associations for short-term ozone exposure and respiratory
hospital admissions and emergency room visits among children than among adults. Panel studies
also provide support for experimental studies with consistent associations between short-term
ozone exposure and lung function and pulmonary inflammation in healthy children. Additional
children's vulnerability and susceptibility factors are listed in Section XIII.B of the Preamble.
4,1.2 Particulate Matter
4.1.2.1 Background on Particulate Matter
Particulate matter (PM) is a complex mixture of solid particles and liquid droplets distributed
among numerous atmospheric gases which interact with solid and liquid phases. Particles in the
atmosphere range in size from less than 0.01 to more than 10 micrometers (|im) in diameter.8
Atmospheric particles can be grouped into several classes according to their aerodynamic
diameter and physical sizes. Generally, the three broad classes of particles include ultrafine
particles (UFPs, generally considered as particles with a diameter less than or equal to 0.1 |im
[typically based on physical size, thermal diffusivity or electrical mobility]), "fine" particles
(PM2.5; particles with a nominal mean aerodynamic diameter less than or equal to 2.5 |im), and
"thoracic" particles (PM10; particles with a nominal mean aerodynamic diameter less than or
equal to 10 |im). Particles that fall within the size range between PM2.5 and PM10, are referred to
as "thoracic coarse particles" (PM10-2.5, particles with a nominal mean aerodynamic diameter
D Children are more susceptible than adults to many air pollutants because of differences in physiology, higher per
body weight breathing rates and consumption, rapid development of the brain and bodily systems, and behaviors
that increase chances for exposure. Even before birth, the developing fetus may be exposed to air pollutants through
the mother that affect development and permanently harm the individual.
Infants and children breathe at much higher rates per body weight than adults, with infants under one year of age
having a breathing rate up to five times that of adults. In addition, children breathe through their mouths more than
adults and their nasal passages are less effective at removing pollutants, which leads to a higher deposition fraction
in their lungs.
170
-------
greater than 2.5 jam and less than or equal to 10 |im). EPA currently has standards that regulate
PM2.5 and PMio.e
Most particles are found in the lower troposphere, where they can have residence times
ranging from a few hours to weeks. Particles are removed from the atmosphere by wet
deposition, such as when they are carried by rain or snow, or by dry deposition, when particles
settle out of suspension due to gravity. Atmospheric lifetimes are generally longest for PM2.5,
which often remains in the atmosphere for days to weeks before being removed by wet or dry
deposition.9 In contrast, atmospheric lifetimes for UFP and PM10-2.5 are shorter. Within hours,
UFP can undergo coagulation and condensation that lead to formation of larger particles, or can
be removed from the atmosphere by evaporation, deposition, or reactions with other atmospheric
components. PM10-2.5 are also generally removed from the atmosphere within hours, through wet
or dry deposition.10
Particulate matter consists of both primary and secondary particles. Primary particles are
emitted directly from sources, such as combustion-related activities (e.g., industrial activities,
motor vehicle operation, biomass burning), while secondary particles are formed through
atmospheric chemical reactions of gaseous precursors (e.g., sulfur oxides (SOx), NOx and
VOCs).
4.1.2.2 Health Effects Associated with Exposure to Particulate Matter
Scientific evidence spanning animal toxicological, controlled human exposure, and
epidemiologic studies shows that exposure to ambient PM is associated with a broad range of
health effects. These health effects are discussed in detail in the Integrated Science Assessment
for Particulate Matter, which was finalized in December 2019 (PM ISA).11 In addition, there is a
more targeted evaluation of studies published since the literature cutoff date of the 2019 PM ISA
in the Supplement to the Integrated Science Assessment for PM (Supplement).12 The PM ISA
characterizes the causal nature of relationships between PM exposure and broad health categories
(e.g., cardiovascular effects, respiratory effects, etc.) using a weight-of-evidence approach.17
Within this characterization, the PM ISA summarizes the health effects evidence for short-term
(i.e., hours up to one month) and long-term (i.e., one month to years) exposures to PM2.5, PM10-
2.5, and ultrafine particles, and concludes that exposures to ambient PM2.5 are associated with a
number of adverse health effects. The discussion below highlights the PM ISA's conclusions,
E 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).
F The causal framework draws upon the assessment and integration of evidence from across scientific disciplines,
spanning atmospheric chemistry, exposure, dosimetry and health effects studies (i.e., epidemiologic, controlled
human exposure, and animal toxicological studies), and assess the related uncertainties and limitations that
ultimately influence our understanding of the evidence. This framework employs a five-level hierarchy that
classifies the overall weight-of-evidence with respect to the causal nature of relationships between criteria pollutant
exposures and health and welfare effects using the following categorizations: causal relationship; likely to be causal
relationship; suggestive of, but not sufficient to infer, a causal relationship; inadequate to infer the presence or
absence of a causal relationship; and not likely to be a causal relationship (U.S. EPA. (2019). Integrated Science
Assessment for Particulate Matter (Final Report). U.S. Environmental Protection Agency, Washington, DC,
EPA/600/R-19/188, SectionP. 3.2.3).
171
-------
and summarizes additional information from the Supplement where appropriate, pertaining to the
health effects evidence for both short- and long-term PM exposures. Further discussion of PM-
related health effects can also be found in the 2022 Policy Assessment for the review of the PM
NAAQS.13
EPA has concluded that recent evidence in combination with evidence evaluated in the 2009
PM ISA supports a "causal relationship" between both long- and short-term exposures to PM2.5
and premature mortality and cardiovascular effects and a "likely to be causal relationship"
between long- and short-term PM2.5 exposures and respiratory effects.14 Additionally, recent
experimental and epidemiologic studies provide evidence supporting a "likely to be causal
relationship" between long-term PM2.5 exposure and nervous system effects, and long-term
PM2.5 exposure and cancer. Because of remaining uncertainties and limitations in the evidence
base, EPA determined the evidence is "suggestive of, but not sufficient to infer, a causal
relationship" for long-term PM2.5 exposure and reproductive and developmental effects (i.e.,
male/female reproduction and fertility; pregnancy and birth outcomes), long- and short-term
exposures and metabolic effects, and short-term exposure and nervous system effects.
As discussed extensively in the 2019 PM ISA and the Supplement, recent studies continue to
support a "causal relationship" between short- and long-term PM2.5 exposures and mortality.15
For short-term PM2.5 exposure, multi-city studies evaluated in the PM ISA, in combination with
single- and multi-city studies evaluated in the 2009 PM ISA, provide evidence of consistent,
positive associations across studies conducted in different geographic locations, populations with
different demographic characteristics, and studies using different exposure assignment
techniques. Additionally, the consistent and coherent evidence across scientific disciplines for
cardiovascular morbidity, particularly ischemic events and heart failure, and to a lesser degree
for respiratory morbidity, including exacerbations of chronic obstructive pulmonary disease
(COPD) and asthma, provide biological plausibility for cause-specific mortality and ultimately
total mortality. Recent epidemiologic studies evaluated in the Supplement, including studies that
employed alternative methods for confounder control, provide additional support to the evidence
base that contributed to the 2019 PM ISA conclusion for short-term PM2.5 exposure and
mortality.
The 2019 PM ISA concluded a "causal relationship" between long-term PM2.5 exposure and
mortality. In addition to reanalyses and extensions of the American Cancer Society (ACS) and
Harvard Six Cities (HSC) cohorts, multiple new cohort studies conducted in the U.S. and
Canada, consisting of people employed in a specific job (e.g., teacher, nurse) and that apply
different exposure assignment techniques, provide evidence of positive associations between
long-term PM2.5 exposure and mortality. Biological plausibility for mortality due to long-term
PM2.5 exposure is provided by the coherence of effects across scientific disciplines for
cardiovascular morbidity, particularly for coronary heart disease, stroke and atherosclerosis, and
for respiratory morbidity, particularly for the development of COPD. Additionally, recent studies
provide evidence indicating that as long-term PM2.5 concentrations decrease there is an increase
in life expectancy. Recent cohort studies evaluated in the Supplement, as well as epidemiologic
studies that conducted accountability analyses or employed alternative methods for confounder
controls, support and extend the evidence base that contributed to the 2019 PM ISA conclusion
for long-term PM2.5 exposure and mortality.
172
-------
A large body of studies examining both short- and long-term PM2.5 exposure and
cardiovascular effects builds on the evidence base evaluated in the 2009 PM ISA. The strongest
evidence for cardiovascular effects in response to short-term PM2.5 exposures is for ischemic
heart disease and heart failure. The evidence for short-term PM2.5 exposure and cardiovascular
effects is coherent across scientific disciplines and supports a continuum of effects ranging from
subtle changes in indicators of cardiovascular health to serious clinical events, such as increased
emergency department visits and hospital admissions due to cardiovascular disease and
cardiovascular mortality. For long-term PM2.5 exposure, there is strong and consistent
epidemiologic evidence of a relationship with cardiovascular mortality. This evidence is
supported by epidemiologic and animal toxicological studies demonstrating a range of
cardiovascular effects including coronary heart disease, stroke, impaired heart function, and
subclinical markers (e.g., coronary artery calcification, atherosclerotic plaque progression),
which collectively provide coherence and biological plausibility. Recent epidemiologic studies
evaluated in the Supplement, as well as studies that conducted accountability analyses or
employed alternative methods for confounder control, support and extend the evidence base that
contributed to the 2019 PM ISA conclusion for both short- and long-term PM2.5 exposure and
cardiovascular effects.
Studies evaluated in the 2019 PM ISA continue to provide evidence of a "likely to be causal
relationship" between both short- and long-term PM2.5 exposure and respiratory effects.
Epidemiologic studies provide consistent evidence of a relationship between short-term PM2.5
exposure and asthma exacerbation in children and COPD exacerbation in adults, as indicated by
increases in emergency department visits and hospital admissions, which is supported by animal
toxicological studies indicating worsening allergic airways disease and subclinical effects related
to COPD. Epidemiologic studies also provide evidence of a relationship between short-term
PM2.5 exposure and respiratory mortality. However, there is inconsistent evidence for respiratory
effects, specifically lung function declines and pulmonary inflammation, in controlled human
exposure studies. With respect to long term PM2.5 exposure, epidemiologic studies conducted in
the U.S. and abroad provide evidence of a relationship with respiratory effects, including
consistent changes in lung function and lung function growth rate, increased asthma incidence,
asthma prevalence, and wheeze in children; acceleration of lung function decline in adults; and
respiratory mortality. The epidemiologic evidence is supported by animal toxicological studies,
which provide coherence and biological plausibility for a range of effects including impaired
lung development, decrements in lung function growth, and asthma development.
Since the 2009 PM ISA, a growing body of scientific evidence examined the relationship
between long-term PM2.5 exposure and nervous system effects, resulting for the first time in a
causality determination for this health effects category of a "likely to be causal relationship". The
strongest evidence for effects on the nervous system come from epidemiologic studies that
consistently report cognitive decrements and reductions in brain volume in adults. The effects
observed in epidemiologic studies in adults are supported by animal toxicological studies
demonstrating effects on the brain of adult animals including inflammation, morphologic
changes, and neurodegeneration of specific regions of the brain. There is more limited evidence
for neurodevelopmental effects in children with some studies reporting positive associations with
autism spectrum disorder and others providing limited evidence of an association with cognitive
function. While there is some evidence from animal toxicological studies indicating effects on
the brain (i.e., inflammatory and morphological changes) to support a biologically plausible
173
-------
pathway for neurodevelopmental effects, epidemiologic studies are limited due to their lack of
control for potential confounding by copollutants, the small number of studies conducted, and
uncertainty regarding critical exposure windows.
Building off the decades of research demonstrating mutagenicity, DNA damage, and other
endpoints related to genotoxicity due to whole PM exposures, recent experimental and
epidemiologic studies focusing specifically on PM2.5 provide evidence of a relationship between
long-term PM2.5 exposure and cancer. Epidemiologic studies examining long-term PM2.5
exposure and lung cancer incidence and mortality provide evidence of generally positive
associations in cohort studies spanning different populations, locations, and exposure assignment
techniques. Additionally, there is evidence of positive associations with lung cancer incidence
and mortality in analyses limited to never smokers. In addition, experimental and epidemiologic
studies of genotoxicity, epigenetic effects, carcinogenic potential, and that PM2.5 exhibits several
characteristics of carcinogens, provide biological plausibility for cancer development. This
collective body of evidence contributed to the conclusion of a "likely to be causal relationship."
For the additional health effects categories evaluated for PM2.5 in the 2019 PM ISA,
experimental and epidemiologic studies provide limited and/or inconsistent evidence of a
relationship with PM2.5 exposure. As a result, the 2019 PM ISA concluded that the evidence is
"suggestive of, but not sufficient to infer a causal relationship" for short-term PM2.5 exposure
and metabolic effects and nervous system effects, and long-term PM2.5 exposures and metabolic
effects as well as reproductive and developmental effects.
In addition to evaluating the health effects attributed to short- and long-term exposure to
PM2.5, the 2019 PM ISA also conducted an extensive evaluation as to whether specific
components or sources of PM2.5 are more strongly related with specific health effects than PM2.5
mass. An evaluation of those studies resulted in the 2019 PM ISA concluding that "many PM2.5
components and sources are associated with many health effects, and the evidence does not
indicate that any one source or component is consistently more strongly related to health effects
than PM2.5 mass."16
For both PM10-2.5 and UFPs, for all health effects categories evaluated, the 2019 PM ISA
concluded that the evidence was "suggestive of, but not sufficient to infer, a causal relationship"
or "inadequate to determine the presence or absence of a causal relationship." For PM10-2.5,
although a Federal Reference Method (FRM) was instituted in 2011 to measure PM10-2.5
concentrations nationally, the causality determinations reflect that the same uncertainty identified
in the 2009 PM ISA persists with respect to the method used to estimate PM10-2.5 concentrations
in epidemiologic studies. Specifically, across epidemiologic studies, different approaches are
used to estimate PM10-2.5 concentrations (e.g., direct measurement of PM10-2.5, difference between
PM10 and PM2.5 concentrations), and it remains unclear how well correlated PM10-2.5
concentrations are both spatially and temporally across the different methods used.
For UFPs, which have often been defined as particles <0.1 |im, the uncertainty in the evidence
for the health effect categories evaluated across experimental and epidemiologic studies reflects
the inconsistency in the exposure metric used (i.e., particle number concentration, surface area
concentration, mass concentration) as well as the size fractions examined. In epidemiologic
studies the size fraction examined can vary depending on the monitor used and exposure metric,
with some studies examining number count over the entire particle size range, while
174
-------
experimental studies that use a particle concentrator often examine particles up to 0.3 |im.
Additionally, due to the lack of a monitoring network, there is limited information on the spatial
and temporal variability of UFPs within the U.S., as well as population exposures to UFPs,
which adds uncertainty to epidemiologic study results.
The 2019 PM ISA cites extensive evidence indicating that "both the general population as
well as specific populations and lifestages are at risk for PIVh.s-related health effects."17 For
example, in support of its "causal" and "likely to be causal" determinations, the ISA cites
substantial evidence for (1) PM-related mortality and cardiovascular effects in older adults; (2)
PM-related cardiovascular effects in people with pre-existing cardiovascular disease; (3) PM-
related respiratory effects in people with pre-existing respiratory disease, particularly asthma
exacerbations in children; and (4) PM-related impairments in lung function growth and asthma
development in children. The ISA additionally notes that stratified analyses (i.e., analyses that
directly compare PM-related health effects across groups) provide strong evidence for racial and
ethnic differences in PM2.5 exposures and in the risk of PM2.5-related health effects, specifically
within Hispanic and non-Hispanic Black populations with some evidence of increased risk for
populations of low socioeconomic status. Recent studies evaluated in the Supplement support the
conclusion of the 2019 PM ISA with respect to disparities in both PM2.5 exposure and health risk
by race and ethnicity and provide additional support for disparities for populations of lower
socioeconomic status. Additionally, evidence spanning epidemiologic studies that conducted
stratified analyses, experimental studies focusing on animal models of disease or individuals
with pre-existing disease, dosimetry studies, as well as studies focusing on differential exposure
suggest that populations with pre-existing cardiovascular or respiratory disease, populations that
are overweight or obese, populations that have particular genetic variants, and current/former
smokers could be at increased risk for adverse PM2.5-related health effects. The 2022 Policy
Assessment for the review of the PM NAAQS also highlights that factors that may contribute to
increased risk of PM2.5-related health effects include lifestage (children and older adults), pre-
existing diseases (cardiovascular disease and respiratory disease), race/ethnicity, and
socioeconomic status.18
4.1.3 Nitrogen Oxides
4.1.3.1 Background on Nitrogen Oxides
Oxides of nitrogen (NOx) refers to nitric oxide (NO) and nitrogen dioxide (NO2). Most NO2
is formed in the air through the oxidation of nitric oxide (NO) that is emitted when fuel is burned
at a high temperature. NOx is a major contributor to secondary PM2.5 formation, and NOx along
with VOCs are the two major precursors of ozone. The health effects of PM and ozone are
discussed in Sections 4.1.1 and 4.1.2 respectively.
4.1.3.2 Health Effects Associated with Exposure to Nitrogen Oxides
The most recent review of the health effects of oxides of nitrogen completed by EPA can be
found in the 2016 Integrated Science Assessment for Oxides of Nitrogen - Health Criteria (ISA
for Oxides of Nitrogen).19 The primary source of NO2 is motor vehicle emissions, and ambient
NO2 concentrations tend to be highly correlated with other traffic-related pollutants. Thus, a key
issue in characterizing the causality of N02-health effect relationships consists of evaluating the
extent to which studies supported an effect of NO2 that is independent of other traffic-related
175
-------
pollutants. EPA concluded that the findings for asthma exacerbation integrated from
epidemiologic and controlled human exposure studies provided evidence that is sufficient to
infer a causal relationship between respiratory effects and short-term NO2 exposure. The
strongest evidence supporting an independent effect of NO2 exposure comes from controlled
human exposure studies demonstrating increased airway responsiveness in individuals with
asthma following ambient-relevant NO2 exposures. The coherence of this evidence with
epidemiologic findings for asthma hospital admissions and emergency department visits as well
as lung function decrements and increased pulmonary inflammation in children with asthma
describe a plausible pathway by which NO2 exposure can cause an asthma exacerbation. The
2016 ISA for Oxides of Nitrogen also concluded that there is likely to be a causal relationship
between long-term NO2 exposure and respiratory effects. This conclusion is based on new
epidemiologic evidence for associations of NO2 with asthma development in children combined
with biological plausibility from experimental studies.
In evaluating a broader range of health effects, the 2016 ISA for Oxides of Nitrogen
concluded that evidence is "suggestive of, but not sufficient to infer, a causal relationship"
between short-term NO2 exposure and cardiovascular effects and mortality and between long-
term NO2 exposure and cardiovascular effects and diabetes, birth outcomes, and cancer. In
addition, the scientific evidence is inadequate (insufficient consistency of epidemiologic and
toxicological evidence) to infer a causal relationship for long-term NO2 exposure with fertility,
reproduction, and pregnancy, as well as with postnatal development. The ISA states that a key
uncertainty in understanding the relationship between these non-respiratory health effects and
short- or long-term exposure to NO2 is copollutant confounding, particularly by other roadway
pollutants. The available evidence for non-respiratory health effects does not adequately address
whether NO2 has an independent effect or whether it primarily represents effects related to other
or a mixture of traffic-related pollutants.
The 2016 ISA for Oxides of Nitrogen concluded that people with asthma, children, and older
adults are at increased risk for MVrelated health effects. In these groups and lifestages, NO2 is
consistently related to larger effects on outcomes related to asthma exacerbation, for which there
is confidence in the relationship with NO2 exposure.
4.1.4 Carbon Monoxide
4.1.4.1 Background on Carbon Monoxide
Carbon monoxide (CO) is a colorless, odorless gas emitted from combustion processes.
Nationally, particularly in urban areas, the majority of CO emissions to ambient air come from
mobile sources.
4.1.4.2 Health Effects Associated with Exposure to 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).20 The CO ISA presents
conclusions regarding the presence of causal relationships between CO exposure and categories
176
-------
of adverse health effects.0 This section provides a summary of the health effects associated with
exposure to ambient concentrations of CO, along with the CO ISA conclusions.11
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 observed 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 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 that 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. There is limited epidemiologic 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 that 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 short-term 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,
G 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.
H Personal exposure includes contributions from many sources, and in many different environments. Total personal
exposure to CO includes both ambient and non-ambient components; and both components may contribute to
adverse health effects.
177
-------
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 evidence
suggests an association exists 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 that 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.
4,1.5 Diesel Exhaust
4.1.5.1 Background on Diesel Exhaust
Diesel exhaust is a complex mixture composed of particulate matter, carbon dioxide, oxygen,
nitrogen, water vapor, carbon monoxide, nitrogen compounds, sulfur compounds and numerous
low-molecular-weight hydrocarbons. A number of these gaseous hydrocarbon components are
individually known to be toxic, including aldehydes, benzene and 1,3-butadiene. The diesel
particulate matter present in diesel exhaust consists mostly of fine particles (< 2.5 |im), of which
a significant fraction is ultrafine particles (< 0.1 |im). These particles have a large surface area
which makes them an excellent medium for adsorbing organics and their small size makes them
highly respirable. Many of the organic compounds present in the gases and on the particles, such
as polycyclic organic matter, are individually known to have mutagenic and carcinogenic
properties.
Diesel exhaust varies significantly in chemical composition and particle sizes between
different engine types (heavy-duty, light-duty), engine operating conditions (idle, acceleration,
deceleration), and fuel formulations (high/low sulfur fuel). Also, there are emissions differences
between on-road and nonroad engines because the nonroad engines are generally of older
technology. After being emitted in the engine exhaust, diesel exhaust undergoes dilution as well
as chemical and physical changes in the atmosphere. The lifetimes of the components present in
diesel exhaust range from seconds to days.
4.1.5.2 Health Effects Associated with Exposure to Diesel Exhaust
In EPA's 2002 Diesel Health Assessment Document (Diesel HAD), exposure to diesel
exhaust was classified as likely to be carcinogenic to humans by inhalation from environmental
exposures, in accordance with the revised draft 1996/1999 EPA cancer guidelines.21'22 A number
of other agencies (National Institute for Occupational Safety and Health, the International
Agency for Research on Cancer, the World Health Organization, California EPA, and the U.S.
Department of Health and Human Services) made similar hazard classifications prior to 2002.
EPA also concluded in the 2002 Diesel HAD that it was not possible to calculate a cancer unit
risk for diesel exhaust due to limitations in the exposure data for the occupational groups or the
absence of a dose-response relationship.
In the absence of a cancer unit risk, the Diesel HAD sought to provide additional insight into
the significance of the diesel exhaust cancer hazard by estimating possible ranges of risk that
178
-------
might be present in the population. An exploratory analysis was used to characterize a range of
possible lung cancer risk. The outcome was that environmental risks of cancer from long-term
diesel exhaust exposures could plausibly range from as low as 10"5 to as high as 10"3. Because of
uncertainties, the analysis acknowledged that the risks could be lower than 10"5, and a zero risk
from diesel exhaust exposure could not be ruled out.
Noncancer health effects of acute and chronic exposure to diesel exhaust emissions are also of
concern to EPA. EPA derived a diesel exhaust reference concentration (RfC) from consideration
of four well-conducted chronic rat inhalation studies showing adverse pulmonary effects. The
RfC is 5 |ig/m3 for diesel exhaust measured as diesel particulate matter. This RfC does not
consider allergenic effects such as those associated with asthma or immunologic or the potential
for cardiac effects. There was emerging evidence in 2002, discussed in the Diesel HAD, that
exposure to diesel exhaust can exacerbate these effects, but the exposure-response data were
lacking at that time to derive an RfC based on these then-emerging considerations. The Diesel
HAD states, "With [diesel particulate matter] being a ubiquitous component of ambient PM,
there is an uncertainty about the adequacy of the existing [diesel exhaust] noncancer database to
identify all of the pertinent [diesel exhaust]-caused noncancer health hazards." The Diesel HAD
also notes "that acute exposure to [diesel exhaust] has been associated with irritation of the eye,
nose, and throat, respiratory symptoms (cough and phlegm), and neurophysiological symptoms
such as headache, lightheadedness, nausea, vomiting, and numbness or tingling of the
extremities." The Diesel HAD notes that the cancer and noncancer hazard conclusions applied to
the general use of diesel engines then on the market and as cleaner engines replace a substantial
number of existing ones, the applicability of the conclusions would need to be reevaluated.
It is important to note that the Diesel HAD also briefly summarizes health effects associated
with ambient PM and discusses EPA's then-annual PM2.5 NAAQS of 15 |ig/m3.' There is a large
and extensive body of human data showing a wide spectrum of adverse health effects associated
with exposure to ambient PM, of which diesel exhaust is an important component. The PM2.5
NAAQS is designed to provide protection from the noncancer health effects and premature
mortality attributed to exposure to PM2.5. The contribution of diesel PM to total ambient PM
varies in different regions of the country and also, within a region, from one area to another. The
contribution can be high in near-roadway environments, for example, or in other locations where
diesel engine use is concentrated.
Since 2002, several new studies have been published which continue to report increased lung
cancer risk associated with occupational exposure to diesel exhaust from older engines. Of
particular note since 2011 are three new epidemiology studies that have examined lung cancer in
occupational populations, including, truck drivers, underground nonmetal miners, and other
diesel motor-related occupations. These studies reported increased risk of lung cancer related to
exposure to diesel exhaust, with evidence of positive exposure-response relationships to varying
degrees.23'24'25 These newer studies (along with others that have appeared in the scientific
literature) add to the evidence EPA evaluated in the 2002 Diesel HAD and further reinforce the
concern that diesel exhaust exposure likely poses a lung cancer hazard. The findings from these
newer studies do not necessarily apply to newer technology diesel engines (i.e., heavy-duty
1 See Section 6.1.2 for discussion of the current PM2.5 NAAQS standard.
179
-------
highway engines from 2007 and later model years) since the newer engines have large reductions
in the emission constituents compared to older technology diesel engines.
In light of the growing body of scientific literature evaluating the health effects of exposure to
diesel exhaust, in June 2012 the World Health Organization's International Agency for Research
on Cancer (IARC), a recognized international authority on the carcinogenic potential of
chemicals and other agents, evaluated the full range of cancer-related health effects data for
diesel engine exhaust. IARC concluded that diesel exhaust should be regarded as "carcinogenic
to humans."26 This designation was an update from its 1988 evaluation that considered the
evidence to be indicative of a "probable human carcinogen."
4,1.6 Air Toxics
Heavy-duty engine emissions contribute to ambient levels of air toxics that are known or
suspected human or animal carcinogens, or that have noncancer health effects. These compounds
include, but are not limited to, benzene, formaldehyde, acetaldehyde, and naphthalene. These
compounds were identified as national or regional cancer risk drivers or contributors in the 2018
AirToxScreen Assessment and have significant inventory contributions from mobile sources.27'28
4.1.6.1 Health Effects Associated with Exposure to Benzene
EPA's Integrated Risk Information System (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.29'30'31 EPA states in its IRIS database
that data indicate a causal relationship between benzene exposure and acute lymphocytic
leukemia and suggest a relationship between benzene exposure and chronic non-lymphocytic
leukemia and chronic lymphocytic leukemia. EPA's IRIS documentation for benzene also lists a
range of 2.2 x 10"6 to 7.8 x 10"6 per |ig/m3 as the unit risk estimate (URE) for benzene/'32 The
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 human carcinogen.33 34
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.35'36 The
most sensitive noncancer effect observed in humans, based on current data, is the depression of
the absolute lymphocyte count in blood.37'38 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, studies sponsored by the Health
Effects Institute (HEI) provide evidence that biochemical responses occur at lower levels of
benzene exposure than previously known.39'40'41'42 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.43 K
1A unit risk estimate is defined as the increase in the lifetime risk of cancer of an individual who is exposed for a
lifetime to 1 |ig/m3 benzene in air.
K 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.
180
-------
There is limited information from two studies regarding an increased risk of adverse effects to
children whose parents have been occupationally exposed to benzene.44'45 Data from animal
studies have shown benzene exposures result in damage to the hematopoietic (blood cell
formation) system during development.46'47'48 Also, key changes related to the development of
childhood leukemia occur in the developing fetus.49 Several studies have reported that genetic
changes related to eventual leukemia development occur before birth. For example, there is one
study of genetic changes in twins who developed T cell leukemia at nine years of age.50
4.1.6.2 Health Effects Associated with Exposure to Formaldehyde
In 1991, EPA concluded that formaldehyde is a Class B1 probable human carcinogen based
on limited evidence in humans and sufficient evidence in animals.51 An Inhalation URE for
cancer and a Reference Dose for oral noncancer effects were developed by EPA and posted on
the IRIS database. Since that time, the NTP and IARC have concluded that formaldehyde is a
known human carcinogen.52'53'54
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.55'56'57
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.58 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.59 Finally, a study of embalmers reported formaldehyde exposures
to be associated with an increased risk of myeloid leukemia but not brain cancer.60
Health effects of formaldehyde in addition to cancer were reviewed by the ATSDR in 1999,
supplemented in 2010, and by the World Health Organization. 61>62>63 These organizations
reviewed the scientific literature concerning health effects linked to formaldehyde exposure to
evaluate hazards and dose response relationships and defined exposure concentrations for
minimal risk levels (MRLs). The health endpoints reviewed included sensory irritation of eyes
and respiratory tract, reduced pulmonary function, nasal histopathology, and immune system
effects. In addition, research on reproductive and developmental effects and neurological effects
were discussed along with several studies that suggest that formaldehyde may increase the risk of
asthma - particularly in the young.
In June 2010, 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.64 That 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
2011.65 EPA's draft assessment, which addresses NRC recommendations, was suspended in
2018.66 The draft assessment was unsuspended in March 2021, and an external review draft was
released in April 2022.67 This draft assessment is now undergoing review by the National
Academy of Sciences.68
181
-------
4.1.6.3 Health Effects Associated with Exposure to 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.69 The
URE in IRIS for acetaldehyde is 2.2 x 10"6 per |ig/m3.70 Acetaldehyde is reasonably anticipated
to be a human carcinogen by the NTP in the 14th Report on Carcinogens and is classified as
possibly carcinogenic to humans (Group 2B) by the IARC.71'72
The primary noncancer effects of exposure to acetaldehyde vapors include irritation of the
eyes, skin, and respiratory tract.73 In short-term (4 week) rat studies, degeneration of olfactory
epithelium was observed at various concentration levels of acetaldehyde exposure.74'75 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.76
Children, especially those with diagnosed asthma, may be more likely to show impaired
pulmonary function and symptoms of asthma than are adults following exposure to
acetaldehyde.77
4.1.6.4 Health Effects Associated with Exposure to 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 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.78
Chronic (long term) exposure of workers and rodents to naphthalene has been reported to cause
cataracts and retinal damage.79 Children, especially neonates, appear to be more susceptible to
acute naphthalene poisoning based on the number of reports of lethal cases in children and
infants (hypothesized to be due to immature naphthalene detoxification pathways).80 EPA
released an external review draft of a reassessment of the inhalation carcinogenicity of
naphthalene based on a number of recent animal carcinogenicity studies.81 The draft
reassessment completed external peer review.82 Based on external peer review comments
received, EPA is developing a revised draft assessment that considers inhalation and oral routes
of exposure, as well as cancer and noncancer effects.83 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 NTP 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.84 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.85
Naphthalene also causes a number of non-cancer effects in animals following chronic and
less-than-chronic exposure, including abnormal cell changes and growth in respiratory and nasal
tissues.86 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/m3.87 The ATSDR
MRL for acute and intermediate duration oral exposure to naphthalene is 0.6 mg/kg/day based
on maternal toxicity in a developmental toxicology study in rats.88 ATSDR also derived an ad
182
-------
hoc reference value of 6 x 10~2 mg/m3 for acute (<24-hour) inhalation exposure to naphthalene in
a Letter Health Consultation dated March 24, 2014 to address a potential exposure concern in
Illinois.89 The ATSDR acute inhalation reference value was based on a qualitative identification
of an exposure level interpreted not to cause pulmonary lesions in mice. More recently, EPA
developed acute RfCs for 1-, 8-, and 24-hour exposure scenarios; the <24-hour reference value is
2 x 10"2 mg/m3.90 EPA's acute RfCs are based on a systematic review of the literature,
benchmark dose modeling of naphthalene-induced nasal lesions in rats, and application of a
PBPK (physiologically based pharmacokinetic) model.
4.1.6.5 Health Effects Associated with Exposure to Other Air Toxics
In addition to the compounds described above, other compounds found in gaseous
hydrocarbon and PM emissions from engines will be affected by this rule. Mobile source air
toxics that will potentially be affected include acrolein, ethylbenzene, propionaldehyde, toluene,
and xylene. Information regarding the health effects of these compounds can be found in EPA's
IRIS database.91
4,1,7 Exposure and Health Effects Associated with Traffic
Locations in close proximity to major roadways generally have elevated concentrations of
many air pollutants emitted from motor vehicles. Hundreds of studies have been published in
peer-reviewed journals, concluding that concentrations of CO, CO2, NO, NO2, benzene,
aldehydes, PM, black carbon, and many other compounds are elevated in ambient air within
approximately 300-600 meters (about 1,000-2,000 feet) of major roadways. The highest
concentrations of most pollutants emitted directly by motor vehicles are found at locations within
50 meters (about 165 feet) of the edge of a roadway's traffic lanes.
A large-scale review of air quality measurements in the vicinity of major roadways between
1978 and 2008 concluded that the pollutants with the steepest concentration gradients in
vicinities of roadways were CO, UFPs, metals, elemental carbon (EC), NO, NOx, and several
VOCs.92 These pollutants showed a large reduction in concentrations within 100 meters
downwind of the roadway. Pollutants that showed more gradual reductions with distance from
roadways included benzene, NO2, PM2.5, and PM10. In reviewing the literature, Karner et al.,
(2010) reported that results varied based on the method of statistical analysis used to determine
the gradient in pollutant concentration. More recent studies continue to show significant
concentration gradients of traffic-related air pollution around major roads.93'94'95'96'97' 98>">100
There is evidence that EPA's regulations for vehicles have lowered the near-road concentrations
and gradients.101 Starting in 2010, EPA required through the NAAQS process that air quality
monitors be placed near high-traffic roadways for determining concentrations of CO, NO2, and
PM2.5 (in addition to those existing monitors located in neighborhoods and other locations farther
away from pollution sources). The monitoring data for NO2 indicate that in urban areas,
monitors near roadways often report the highest concentrations of NO2.102 More recent studies
of traffic-related air pollutants continue to report sharp gradients around roadways, particularly
within several hundred meters.103'104
For pollutants with relatively high background concentrations relative to near-road
concentrations, detecting concentration gradients can be difficult. For example, many carbonyls
have high background concentrations as a result of photochemical breakdown of precursors from
183
-------
many different organic compounds. However, several studies have measured carbonyls in
multiple weather conditions and found higher concentrations of many carbonyls downwind of
roadways.105'106 These findings suggest a substantial roadway source of these carbonyls.
In the past 30 years, many studies have been published with results reporting that populations
who live, work, or go to school near high-traffic roadways experience higher rates of numerous
adverse health effects, compared to populations far away from major roads.L In addition,
numerous studies have found adverse health effects associated with spending time in traffic, such
as commuting or walking along high-traffic roadways, including studies among
children.107'108'109'110 The health outcomes with the strongest evidence linking them with traffic-
associated air pollutants are respiratory effects, particularly in asthmatic children, and
cardiovascular effects. Commenters on the NPRM stressed the importance of consideration of
the impacts of traffic-related air pollution, especially NOx, on children's health.
Numerous reviews of this body of health literature have been published. In a 2022 final
report, an expert panel of the Health Effects Institute (HEI) employed a systematic review
focusing on selected health endpoints related to exposure to traffic-related air pollution.111 The
HEI panel concluded that there was a high level of confidence in evidence between long-term
exposure to traffic-related air pollution and health effects in adults, including all-cause,
circulatory, and ischemic heart disease mortality.112 The panel also found that there is a
moderate-to-high level of confidence in evidence of associations with asthma onset and acute
respiratory infections in children and lung cancer and asthma onset in adults. This report follows
on an earlier expert review published by HEI in 2010, where it found strongest evidence for
asthma-related traffic impacts. Other literature reviews have been published with conclusions
generally similar to the HEI panels'.113'114'115'116 Additionally, in 2014, researchers from the U.S.
Centers for Disease Control and Prevention (CDC) published a systematic review and meta-
analysis of studies evaluating the risk of childhood leukemia associated with traffic exposure and
reported positive associations between "postnatal" proximity to traffic and leukemia risks, but no
such association for "prenatal" exposures.117 The U.S. Department of Health and Human
Services' National Toxicology Program (NTP) published a monograph including a systematic
review of traffic-related air pollution and its impacts on hypertensive disorders of pregnancy.
The NTP concluded that exposure to traffic-related air pollution is "presumed to be a hazard to
pregnant women" for developing hypertensive disorders of pregnancy.118
Health outcomes with few publications suggest the possibility of other effects still lacking
sufficient evidence to draw definitive conclusions. Among these outcomes with a small number
of positive studies are neurological impacts (e.g., autism and reduced cognitive function) and
reproductive outcomes (e.g., preterm birth, low birth weight). "9-120-121-122-123
In addition to health outcomes, particularly cardiopulmonary effects, conclusions of numerous
studies suggest mechanisms by which traffic-related air pollution affects health. For example,
numerous studies indicate that near-roadway exposures may increase systemic inflammation,
affecting organ systems, including blood vessels and lungs,124>125>126>127 Additionally, long-term
L In the widely used PubMed database of health publications, between January 1, 1990 and December 31, 2021,
1,979 publications contained the keywords "traffic, pollution, epidemiology," with approximately half the studies
published after 2015.
184
-------
exposures in near-road environments have been associated with inflammation-associated
conditions, such as atherosclerosis and asthma.128'129'130
Several studies suggest that some factors may increase susceptibility to the effects of traffic-
associated air pollution. Several studies have found stronger adverse health associations in
children experiencing chronic social stress, such as in violent neighborhoods or in homes with
low incomes or high family stress.13i-132-133.134
The risks associated with residence, workplace, or schools near major roads are of potentially
high public health significance due to the large population in such locations. The 2013 U.S.
Census Bureau's American Housing Survey (AHS) was the last AHS that included whether
housing units were within 300 feet of an "airport, railroad, or highway with four or more
lanes."M The 2013 survey reports that 17.3 million housing units, or 13 percent of all housing
units in the U.S., were in such areas. Assuming that populations and housing units are in the
same locations, this corresponds to a population of more than 41 million U.S. residents in close
proximity to high-traffic roadways or other transportation sources. According to the Central
Intelligence Agency's World Factbook, based on data collected between 2012-2014, the United
States had 6,586,610 km of roadways, 293,564 km of railways, and 13,513 airports. As such,
highways represent the overwhelming majority of transportation facilities described by this
factor in the AHS.
EPA also conducted a study to estimate the number of people living near truck freight routes
in the United States.135 Based on a population analysis using the U.S. Department of
Transportation's (USDOT) Freight Analysis Framework 4 (FAF4) and population data from the
2010 decennial census, an estimated 72 million people live within 200 meters of these freight
routes.N'° In addition, relative to the rest of the population, people of color and those with lower
incomes are more likely to live near FAF4 truck routes. They are also more likely to live in
metropolitan areas. The EPA's Exposure Factor Handbook also indicates that, on average,
Americans spend more than an hour traveling each day, bringing nearly all residents into a high-
exposure microenvironment for part of the day.136
As described in Section 4.3, we estimate that about 10 million students attend schools
within 200 meters of major roads.137 Research into the impact of traffic-related air pollution on
school performance is tentative. A review of this literature found some evidence that children
exposed to higher levels of traffic-related air pollution show poorer academic performance than
those exposed to lower levels of traffic-related air pollution.138 However, this evidence was
judged to be weak due to limitations in the assessment methods.
While near-roadway studies focus on residents near roads or others spending
considerable time near major roads, the duration of commuting results in another important
M The variable was known as "ETRANS" in the questions about the neighborhood.
N FAF4 is a model from the USDOT's Bureau of Transportation Statistics (BTS) and Federal Highway
Administration (FHWA), which provides data associated with freight movement in the U.S. It includes data from
the 2012 Commodity Flow Survey (CFS), the Census Bureau on international trade, as well as data associated with
construction, agriculture, utilities, warehouses, and other industries. FAF4 estimates the modal choices for moving
goods by trucks, trains, boats, and other types of freight modes. It includes traffic assignments, including truck
flows on a network of truck routes, https://ops.fhwa.dot.gov/freight/freight_analysis/faf/.
0 The same analysis estimated the population living within 100 meters of a FAF4 truck route is 41 million.
185
-------
contributor to overall exposure to traffic-related air pollution. Studies of health that address time
spent in transit have found evidence of elevated risk of cardiac impacts.139'140'141 Studies have
also found that school bus emissions can increase student exposures to diesel-related air
pollutants, and that programs that reduce school bus emissions may improve health and reduce
school absenteeism.142> 143> 144> 145
4.2 Environmental Effects Associated with Exposure to Pollutants
This section discusses the environmental effects associated with pollutants affected by this
rule, specifically PM, ozone, NOx and air toxics.
4,2.1 Visibility Degradation
Visibility can be defined as the degree to which the atmosphere is transparent to visible
light.146 Visibility impairment is caused by light scattering and absorption by suspended particles
and gases. It is dominated by contributions from suspended particles except under pristine
conditions. Fine particles with significant light-extinction efficiencies include sulfates, nitrates,
organic carbon, elemental carbon, sea salt, and soil. 147>148 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 2019
PM ISA.149
The extent to which any amount of light extinction affects a person's ability to view a scene
depends on both scene and light characteristics. For example, the appearance of a nearby object
(e.g., a building) is generally less sensitive to a change in light extinction than the appearance of
a similar object at a greater distance. See Figure 4-1 for an illustration of the important factors
affecting visibility.150
186
-------
Image-forming
light scattered
out of sight path
Image-forming
light absorbed
Optical Characteristics of Illumination
Characteristics of Observer
• Sunlight (Sun Angle)
• Cloud Cover (Overcast, Puffy, etc.)
• Sky
tlcai Characteristics of
Atmosphere
• Light Added to Sight Path by
Particles and Cases
• Image-Forming Light Subtracted
from Sight Path by Scattering
and Absorption
• Detection Thresholds
• Psychological Response to
Incoming Light
• Value Judgements
Light from clouds
scattered into
sight path V
Sunlight ^
seattered Light reflected
from ground
scattered Into
• Color
• Contrast Detail (Texture)
• Form
• Brightness
Figure 4-1: Important Factors Involved in Seeing a Scenic Vista (Malm, 2016)
EPA is working to address visibility impairment. Reductions in air pollution from
implementation of various programs called for in the Clean Air Act Amendments of 1990
(CAAA) have resulted in substantial improvements in visibility and will continue to do so in the
future. Nationally, because trends in haze are closely associated with trends in particulate sulfate
and nitrate emissions due to the relationship between their concentration and light extinction,
visibility trends have improved as emissions of SO2 and NOx have decreased over time due to air
pollution regulations such as the Acid Rain Program.151 However, in the western part of the
country, changes in total light extinction were smaller, and the contribution of particulate organic
matter to atmospheric light extinction was increasing due to increasing wildfire emissions.152
In the Clean Air Act Amendments of 1977, Congress recognized visibility's value to society
by establishing a national goal to protect national parks and wilderness areas from visibility
impairment caused by manmade pollution.153 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 that were in existence on August 7, 1977. Figure 4-2 shows the location of the
156 Mandatory Class I Federal areas.
187
-------
Figure 4-2: Mandatory Class I Federal Areas in the U.S.
EPA has also concluded that PM2.5 causes adverse effects on visibility in other areas that are
not targeted by the Regional Haze Rule, such as urban areas, depending on PM2.5 concentrations
and other factors such as dry chemical composition and relative humidity (i .e., an indicator of the
water composition of the particles). The secondary (welfare-based) PM NAAQS provide
protection against visibility effects. In recent PM NAAQS reviews, EPA evaluated a target level
of protection for visibility impairment that is expected to be met through attainment of the
existing secondary PM standards.
4.2.1.1 Visibility 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., 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 152 sites that represent all but one of the
156 Mandatory Federal Class I areas across the country (see Figure 4-2). This long-term
visibility monitoring network is known as IMPROVE (Interagency Monitoring of Protected
Visual Environments).
IMPROVE provides direct measurement of 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 PM10 and PM2.5 mass,
and for key constituents of PM2.5, such as sulfate, nitrate, organic and elemental carbon (OC and
EC), and other elements that can be used to estimate soil dust and sea salt contributions.
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. The IMPROVE program
utilizes both an "original" and a "revised" reconstruction formula for this purpose, with the latter
188
-------
explicitly accounting for sea salt concentrations. Knowledge of the main constituents of a site's
light extinction "budget" is critical for source apportionment and control strategy development.
In 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 a nephelometer to measure light scattering; some sites also include an
aethalometer for light absorption; and a few sites use 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 total light extinction calculated from the IMPROVE reconstruction
formula are consistent with directly measured extinction. Aerosol-derived light extinction from
the IMPROVE equation 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. Figures 13-1 through 13-14 in the
PM ISA detail the percent contributions to particulate light extinction for ammonium nitrate and
sulfate, EC and OC, and coarse mass and fine soil, by month.154
4,2,2 Plant and Ecosystem Effects of Ozone
The welfare effects of ozone include effects on ecosystems, which can be observed across a
variety of scales, i.e., subcellular, cellular, leaf, whole plant, population and ecosystem. When
ozone effects that begin at small spatial scales, such as the leaf of an individual plant, occur at
sufficient magnitudes (or to a sufficient degree), they can result in effects being propagated along
a continuum to higher and higher levels of biological organization. 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 plant species depending on the
concentration level and the duration of the exposure.155 In those sensitive speciesp, effects from
repeated exposure to ozone throughout the growing season of the plant can tend to accumulate,
so that even relatively low concentrations experienced for a longer duration have the potential to
create chronic stress on vegetation.156>Q Ozone damage to sensitive plant 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.157 These latter impacts
include increased susceptibility of plants to insect attack, disease, harsh weather, interspecies
P Only a small percentage of all the plant species growing within the U.S. (over 43,000 species have been
catalogued in the USDA PLANTS database) have been studied with respect to ozone sensitivity.
Q 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.
189
-------
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 ecosystemsR,
resulting in a loss or reduction in associated ecosystem goods and services.158 Additionally,
visible ozone injury to leaves can result in a loss of aesthetic value in areas of special scenic
significance like national parks and wilderness areas and reduced use of sensitive ornamentals in
landscaping.159 In addition to ozone effects on vegetation, newer evidence suggests that ozone
affects interactions between plants and insects by altering chemical signals (e.g., floral scents)
that plants use to communicate to other community members, such as attraction of pollinators.
The Ozone ISA presents more detailed information on how ozone affects vegetation and
ecosystems.160 The Ozone ISA reports causal and likely causal relationships between ozone
exposure and a number of welfare effects and characterizes the weight of evidence for different
effects associated with ozone.s The Ozone ISA concludes that visible foliar injury effects on
vegetation, reduced vegetation growth, reduced plant reproduction, reduced productivity in
terrestrial ecosystems, reduced yield and quality of agricultural crops, alteration of below-ground
biogeochemical cycles, and altered terrestrial community composition are causally associated
with exposure to ozone. It also concludes that increased tree mortality, altered herbivore growth
and reproduction, altered plant-insect signaling, reduced carbon sequestration in terrestrial
ecosystems, and alteration of terrestrial ecosystem water cycling are likely to be causally
associated with exposure to ozone.
4.2,3 Deposition
Deposited airborne pollutants contribute to adverse effects on ecosystems, and to soiling and
materials damage. These welfare effects result mainly from exposure to excess amounts of
specific chemical species, regardless of their source or predominant form (particle, gas or liquid).
Nitrogen and sulfur tend to comprise a large portion of PM in many locations; however, gas-
phase forms of oxidized nitrogen and sulfur also cause adverse ecological effects. The following
characterizations of the nature of these environmental effects are based on information contained
in the 2019 PM ISA, and the 2020 Integrated Science Assessment for Oxides of Nitrogen,
Oxides of Sulfur, and Particulate Matter - Ecological Criteria.161 -162 This rule will reduce
emissions of nitrogen and PM but will not change emissions of sulfur.
4.2.3.1 Deposition of Nitrogen and Sulfur
Nitrogen and sulfur interactions in the environment are highly complex, as shown in Figure
4-3.163 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.161
In addition, in aquatic ecosystems, sulfur deposition can increase mercury methylation.
R Per footnote above, ozone impacts could be occurring in areas where plant species sensitive to ozone have not yet
been studied or identified.
s 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.
190
-------
Ambient Air
Concentration
Sunlight
Dissolution
~ 2H* ~SO«*"
* H'+NOj'
Oxidation
so2 *¦ H2SO«
NO, ~HNOj
Wet Deposition
H\ NH«', NOj\ SO«2
Dry deposition
NO,, NHX. SO,
Deposition
effects
Acidification of water ~ Eutrophication
Ecological
Effect
Figure 4-3: Nitrogen and Sulfur Cycling, and Interactions in the Environment
4.2.3.1.1 Ecological Effects of Acidification
Deposition of nitrogen and sulfur can cause acidification, which alters biogeochemistry and
affects animal and plant life in terrestrial and aquatic ecosystems across the U.S. Soil
acidification is a natural process, but is often accelerated by acidifying deposition, which can
decrease concentrations of exchangeable base cations in soils.161 Biological effects of
acidification in terrestrial ecosystems are generally linked to aluminum toxicity and decreased
ability of plant roots to take up base cations.161 Decreases in the acid neutralizing capacity and
increases in inorganic aluminum concentration contribute to declines in zooplankton, macro
invertebrates, and fish species richness in aquatic ecosystems.161
Geology (particularly surficial geology) is the principal factor governing the sensitivity of
terrestrial and aquatic ecosystems to acidification from nitrogen and sulfur deposition.161
Geologic formations having low base cation supply generally underlie the watersheds of acid-
sensitive lakes and streams. Other factors contribute to the sensitivity of soils and surface waters
to acidifying deposition, including topography, soil chemistry, land use, and hydrologic flow
path.161
4.2.3.1.1.1 Aquatic Acidification
Aquatic effects of acidification have been well studied in the U.S. and elsewhere at various
trophic levels. These studies indicate that aquatic biota have been affected by acidification at
virtually all levels of the food web in acid sensitive aquatic ecosystems. Effects have been most
191
-------
clearly documented for fish, aquatic insects, other invertebrates, and algae. Biological effects are
primarily attributable to a combination of low pH and high inorganic aluminum concentrations.
Such conditions occur more frequently during rainfall and snowmelt that cause high flows of
water, and less commonly during low-flow conditions, except where chronic acidity conditions
are severe. Biological effects of episodes include reduced fish condition factor, changes in
species composition and declines in aquatic species richness across multiple taxa, ecosystems
and regions.
Because acidification primarily affects the diversity and abundance of aquatic biota, it also
affects the ecosystem services, e.g., recreational and subsistence fishing, that are derived from
the fish and other aquatic life found in these surface waters. In the northeastern United States,
the surface waters affected by acidification are a source of food for some recreational and
subsistence fishermen and for other consumers with particularly high rates of self-caught fish
consumption, such as the Hmong and Chippewa ethnic groups.164'165
4.2.3.1.1.2 Terrestrial Acidification
Acidifying deposition has altered major biogeochemical processes in the U.S. by increasing
the nitrogen and sulfur content of soils, accelerating nitrate and sulfate leaching from soil to
drainage waters, depleting base cations (especially calcium and magnesium) from soils, and
increasing the mobility of aluminum. Inorganic aluminum is toxic to some tree roots. Plants
affected by high levels of aluminum from the soil often have reduced root growth, which restricts
the ability of the plant to take up water and nutrients, especially calcium.161 These direct effects
can, in turn, influence the response of these plants to climatic stresses such as droughts and cold
temperatures. They can also influence the sensitivity of plants to other stresses, including insect
pests and disease leading to increased mortality of canopy trees.166 In the U.S., terrestrial effects
of acidification are best described for forested ecosystems (especially red spruce and sugar maple
ecosystems) with additional information on other plant communities, including shrubs and
lichen.161
Both coniferous and deciduous forests throughout the eastern U.S. are experiencing gradual
losses of base cation nutrients from the soil due to accelerated leaching from acidifying
deposition. This change in nutrient availability may reduce the quality of forest nutrition over
the long term. Evidence suggests that red spruce and sugar maple in some areas in the eastern
U.S. have experienced declining health because of this deposition. For red spruce (Picea
rubens), dieback or decline has been observed across high elevation landscapes of the
northeastern U.S. and, to a lesser extent, the southeastern U.S., and acidifying deposition has
been implicated as a causal factor.167
4.2.3.1.2 Ecological Effects from Nitrogen Enrichment
4.2.3.1.2.1 Aquatic Enrichment
Eutrophication in estuaries is associated with a range of adverse ecological effects including
low dissolved oxygen (DO), harmful algal blooms (HABs), loss of submerged aquatic vegetation
(SAV), and low water clarity. Low DO disrupts aquatic habitats, causing stress to fish and
shellfish, which, in the short-term, can lead to episodic fish kills and, in the long-term, can
damage overall growth in fish and shellfish populations. Low DO also degrades the aesthetic
qualities of surface water. In addition to often being toxic to fish and shellfish and leading to fish
192
-------
kills and aesthetic impairments of estuaries, HABs can, in some instances, also be harmful to
human health. SAV provides critical habitat for many aquatic species in estuaries and, in some
instances, can also protect shorelines by reducing wave strength; therefore, declines in SAV due
to nutrient enrichment are an important source of concern. Low water clarity is in part the result
of accumulations of both algae and sediments in estuarine waters. In addition to contributing to
declines in SAV, high levels of turbidity also degrade the aesthetic qualities of the estuarine
environment.
An assessment of estuaries nationwide by the National Oceanic and Atmospheric
Administration (NOAA) concluded that 64 estuaries (out of 99 with available data) suffered
from moderate or high levels of eutrophication due to excessive inputs of both nitrogen (N) and
phosphorus.168 For estuaries in the Mid-Atlantic region, the contribution of atmospheric
deposition to total N loads is estimated to range between 10 percent and 58 percent.169 Estuaries
in the eastern United States are an important source of food production, in particular for fish and
shellfish production. The estuaries are capable of supporting large stocks of resident commercial
species, and they serve as the breeding grounds and interim habitat for several migratory species.
Eutrophication in estuaries may also affect the demand for seafood (after well-publicized toxic
blooms), water-based recreation, and erosion protection provided by SAV.
4.2.3.1,2.2 Terrestrial Enrichment
Terrestrial enrichment occurs when terrestrial ecosystems receive N loadings in excess of
natural background levels, through either atmospheric deposition or direct application.
Atmospheric N deposition is associated with changes in the types and number of species and
biodiversity in terrestrial systems. Nitrogen enrichment occurs over a long time period; as a
result, it may take as many as 50 years or more to see changes in ecosystem conditions and
indicators. One of the main provisioning services potentially affected by N deposition is grazing
opportunities offered by grasslands for livestock production in the Central U.S. Although N
deposition on these grasslands can offer supplementary nutritive value and promote overall grass
production, there are concerns that fertilization may favor invasive grasses and shift the species
composition away from native grasses. This process may ultimately reduce the productivity of
grasslands for livestock production.
Terrestrial enrichment also affects habitats, for example the Coastal Sage Scrub (CSS) and
Mixed Conifer Forest (MCF) habitats which are an integral part of the California landscape.
Together the ranges of these habitats include the densely populated and valuable coastline and
the mountain areas. Numerous threatened and endangered species at both the state and federal
levels reside in CSS and MCF. Nutrient enrichment of the CSS and MCF also affects the
regulating service of fire, by encouraging the growth of more flammable grasses and thus
increasing fuel loads and altering the fire cycle.
4,2,3.1.3 Vegetation Effects Associated with Gaseous Sulfur Dioxide, Nitric Oxide, Nitrogen
Dioxide, Peroxyacetyl Nitrate, and Nitric Acid
Uptake of gaseous pollutants in a plant canopy is a complex process involving adsorption to
surfaces (leaves, stems, and soil) and absorption into leaves. These pollutants penetrate into
leaves through the stomata, although there is evidence for limited pathways via the cuticle.161
Pollutants must be transported from the bulk air to the leaf boundary layer in order to reach the
193
-------
stomata. When the stomata are closed, as occurs under dark or drought conditions, resistance to
gas uptake is very high and the plant has a very low degree of susceptibility to injury. In
contrast, mosses and lichens do not have a protective cuticle barrier to gaseous pollutants or
stomates and are generally more sensitive to gaseous sulfur and nitrogen than vascular plants.161
Acute foliar injury from SO2 usually happens within hours of exposure, involves a rapid
absorption of a toxic dose, and involves collapse or necrosis of plant tissues. Another type of
visible injury is termed chronic injury and is usually a result of variable SO2 exposures over the
growing season. Besides foliar injury, chronic exposure to low SO2 concentrations can result in
reduced photosynthesis, growth, and yield of plants.161 These effects are cumulative over the
season and are often not associated with visible foliar injury. As with foliar injury, these effects
vary among species and growing environment. SO2 is also considered the primary factor causing
the death of lichens in many urban and industrial areas.170
Similarly, in sufficient concentrations, nitric oxide (NO), nitrogen dioxide (NO2),
peroxyacetyl nitrate (PAN), and nitric acid (HNO3) can have phytotoxic effects on plants such as
decreasing photosynthesis and inducing visible foliar injury. It is also known that these gases can
alter the N cycle in some ecosystems, especially in the western U.S., and contribute to N
saturation. Further, there are several lines of evidence that past and current HNO3 concentrations
may be contributing to the decline in lichen species in the Los Angeles basin.171
4.2.3.1.4 Mercury Methylation
Mercury is a persistent, bioaccumulative toxic metal that is emitted in three forms: gaseous
elemental Hg (Hg°), oxidized Hg compounds (Hg+2), and particle-bound Hg (HgP).
Methylmercury (MeHg) is formed by microbial action in the top layers of sediment and soils
after Hg has precipitated from the air and deposited into waterbodies or land. Once formed,
MeHg is taken up by aquatic organisms and bioaccumulates up the aquatic food web. Larger
predatory fish may have MeHg concentrations many times higher, typically on the order of one
million times, than the concentrations in the freshwater body in which they live. The NOx SOx
ISA—Ecological Criteria concluded that evidence is sufficient to infer a causal relationship
between sulfur deposition and increased mercury methylation in wetlands and aquatic
environments.161 Specifically, there appears to be a relationship between SO42" deposition and
mercury methylation; however, the rate of mercury methylation varies according to several
spatial and biogeochemical factors whose influence has not been fully quantified. Therefore, the
correlation between SO42" deposition and MeHg cannot yet be quantified for the purpose of
interpolating the association across waterbodies or regions. Nevertheless, because changes in
MeHg in ecosystems represent changes in significant human and ecological health risks, the
association between sulfur and mercury cannot be neglected.161
4.2.3.2 Deposition of Metallic and Organic Constituents of PM
Several significant ecological effects are associated with the deposition of chemical
constituents of ambient PM such as metals and organics.161 The trace metal constituents of PM
include cadmium, copper, chromium, mercury, nickel, zinc, and lead. The organics include
persistent organic pollutants (POPs), polyaromatic hydrocarbons (PAHs) and polybromiated
diphenyl ethers (PBDEs). Direct effect exposures to PM occur via deposition (e.g., wet, dry or
occult) to vegetation surfaces, while indirect effects occur via deposition to ecosystem soils or
194
-------
surface waters where the deposited constituents of PM then interact with biological organisms.
While both fine and coarse-mode particles may affect plants and other organisms, more often the
chemical constituents drive the ecosystem response to PM.172 Ecological effects of PM include
direct effects to metabolic processes of plant foliage; contribution to total metal loading resulting
in alteration of soil biogeochemistry and microbiology, plant and animal growth and
reproduction; and contribution to total organics loading resulting in bioaccumulation and
biomagnification.
Particulate matter can adversely impact plants and ecosystem services provided by plants by
deposition to vegetative surfaces.161 Particulates deposited on the surfaces of leaves and needles
can block light, altering the radiation received by the plant. PM deposition near sources of heavy
deposition can obstruct stomata (limiting gas exchange), damage leaf cuticles and increase plant
temperatures.161 Plants growing on roadsides exhibit impact damage from near-road PM
deposition, having higher levels of organics and heavy metals, and accumulating salt from road
de-icing during winter months.161 In addition, atmospheric PM can convert direct solar radiation
to diffuse radiation, which is more uniformly distributed in a tree canopy, allowing radiation to
reach lower leaves.161 Decreases in crop yields (a provisioning service) due to reductions in
solar radiation have been attributed to regional scale air pollution in counties with especially
severe regional haze.173
In addition to damage to plant surfaces, deposited PM can be taken up by plants from soil or
foliage. Copper, zinc, and nickel have been shown to be directly toxic to vegetation under field
conditions.161 The ability of vegetation to take up heavy metals is dependent upon the amount,
solubility and chemical composition of the deposited PM. Uptake of PM by plants from soils
and vegetative surfaces can disrupt photosynthesis, alter pigments and mineral content, reduce
plant vigor, decrease frost hardiness and impair root development.
Particulate matter can also contain organic air toxic pollutants, including PAHs, which are a
class of polycyclic organic matter (POM). PAHs can accumulate in sediments and
bioaccumulate in freshwater, flora and fauna. The uptake of organic air toxic pollutants depends
on the plant species, site of deposition, physical and chemical properties of the organic
compound and prevailing environmental conditions.161 Different species can have different
uptake rates of PAHs. PAHs can accumulate to high enough concentrations in some coastal
environments to pose an environmental health threat that includes cancer in fish populations,
toxicity to organisms living in the sediment and risks to those (e.g., migratory birds) that
consume these organisms.174'175 Atmospheric deposition of particles is thought to be the major
source of PAHs in the sediments of Lake Michigan, Chesapeake Bay, Tampa Bay and other
coastal areas of the U.S.176
Contamination of plant leaves by heavy metals can lead to elevated concentrations in the soil.
Trace metals absorbed into the plant, frequently by binding to the leaf tissue, and then are shed
when the leaf drops. As the fallen leaves decompose, the heavy metals are transferred into the
soil. 177>178 Many of the major indirect plant responses to PM deposition are chiefly soil-mediated
and depend on the chemical composition of individual components of deposited PM. Upon
entering the soil environment, PM pollutants can alter ecological processes of energy flow and
nutrient cycling, inhibit nutrient uptake to plants, change microbial community structure, and
affect biodiversity. Accumulation of heavy metals in soils depends on factors such as local soil
characteristics, geologic origin of parent soils, and metal bioavailability. Heavy metals such as
195
-------
zinc, copper, and cadmium, and some pesticides can interfere with microorganisms that are
responsible for decomposition of soil litter, an important regulating ecosystem service that serves
as a source of soil nutrients.161 Surface litter decomposition is reduced in soils having high metal
concentrations. Soil communities have associated bacteria, fungi, and invertebrates that are
essential to soil nutrient cycling processes. Changes to the relative species abundance and
community composition are associated with deposited PMto soil biota.161
Atmospheric deposition can be the primary source of some organics and metals to watersheds.
Deposition of PM to surfaces in urban settings increases the metal and organic component of
storm water runoff161 This atmospherically-associated pollutant burden can then be toxic to
aquatic biota. The contribution of atmospherically deposited PAHs to aquatic food webs was
demonstrated in high elevation mountain lakes with no other anthropogenic contaminant
sources.161 Metals associated with PM deposition limit phytoplankton growth, affecting aquatic
trophic structure. Long-range atmospheric transport of 47 pesticides and degradation products to
the snowpack in seven national parks in the Western U.S. was recently quantified indicating PM-
associated contaminant inputs in receiving waters during spring snowmelt. The recently
completed Western Airborne Contaminants Assessment Project (WACAP) is the most
comprehensive database on contaminant transport and PM depositional effects on sensitive
ecosystems in the Western U.S.179 In this project, the transport, fate, and ecological impacts of
anthropogenic contaminants from atmospheric sources were assessed from 2002 to 2007 in seven
ecosystem components (air, snow, water, sediment, lichen, conifer needles and fish) in eight core
national parks. The study concluded that bioaccumulation of semi-volatile organic compounds
occurred throughout park ecosystems, an elevational gradient in PM deposition exists with
greater accumulation in higher altitude areas, and contaminants accumulate in proximity to
individual agriculture and industry sources, which is counter to the original working hypothesis
that most of the contaminants would originate from Eastern Europe and Asia.
4.2.3.3 Materials Damage and Soiling
Building materials including metals, stones, cements, and paints undergo natural weathering
processes from exposure to environmental elements (e.g., wind, moisture, temperature
fluctuations, sunlight, etc.). Pollution can worsen and accelerate these effects. Deposition of PM
is associated with both physical damage (materials damage effects) and impaired aesthetic
qualities (soiling effects). Wet and dry deposition of PM can physically affect materials, adding
to the effects of natural weathering processes, by potentially promoting or accelerating the
corrosion of metals, degrading paints and deteriorating building materials such as stone, concrete
and marble.180 The effects of PM are exacerbated by the presence of acidic gases and can be
additive or synergistic depending on the complex mixture of pollutants in the air and surface
characteristics of the material. Acidic deposition has been shown to have an effect on materials
including zinc/galvanized steel and other metal, carbonate stone (such as monuments and
building facings), and surface coatings (paints).181 The effects on historic buildings and outdoor
works of art are of particular concern because of the uniqueness and irreplaceability of many of
these objects. In addition to aesthetic and functional effects on metals, stone and glass, altered
196
-------
energy efficiency of photovoltaic panels by PM deposition is also becoming an important
consideration for impacts of air pollutants on materials.
4.2.4 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. VOCs, some of which are considered
air toxics, have long been suspected to play a role in vegetation damage.182 In laboratory
experiments, a wide range of tolerance to VOCs has been observed.183 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.184
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 NOx. 185>186>187 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.
4.3 Environmental Justice
EPA's 2016 "Technical Guidance for Assessing Environmental Justice in Regulatory
Analysis" provides recommendations on conducting the highest quality analysis feasible,
recognizing that data limitations, time and resource constraints, and analytic challenges will vary
by media and regulatory context.188 When assessing the potential for disproportionately high and
adverse health or environmental impacts of regulatory actions on people of color, low-income
populations, Tribes, and/or indigenous peoples, the EPA strives to answer three broad questions:
(1) Is there evidence of potential environmental justice (EJ) concerns in the baseline (the state of
the world absent the regulatory action)? Assessing the baseline will allow the EPA to determine
whether pre-existing disparities are associated with the pollutant(s) under consideration (e.g., if
the effects of the pollutant(s) are more concentrated in some population groups). (2) Is there
evidence of potential EJ concerns for the regulatory option(s) under consideration? Specifically,
how are the pollutant(s) and its effects distributed for the regulatory options under consideration?
And, (3) do the regulatory option(s) under consideration exacerbate or mitigate EJ concerns
relative to the baseline? It is not always possible to quantitatively assess these questions.
EPA's 2016 Technical Guidance does not prescribe or recommend a specific approach or
methodology for conducting an environmental justice analysis, though a key consideration is
consistency with the assumptions underlying other parts of the regulatory analysis when
evaluating the baseline and regulatory options. Where applicable and practicable, the Agency
endeavors to conduct such an analysis. EPA is committed to conducting environmental justice
analysis for rulemakings based on a framework similar to what is outlined in EPA's Technical
Guidance, in addition to investigating ways to further weave environmental justice into the fabric
of the rulemaking process.
197
-------
There is evidence that communities with EJ concerns are disproportionately impacted by the
emissions sources controlled in this final rule.189 Numerous studies have found that
environmental hazards such as air pollution are more prevalent in areas where people of color
and low-income populations represent a higher fraction of the population compared with the
general population.190'191'192 Consistent with this evidence, a recent study found that most
anthropogenic sources of PM2.5, including industrial sources and light- and heavy-duty vehicle
sources, disproportionately affect people of color.193 In addition, compared to non-Hispanic
Whites, some other racial groups experience greater levels of health problems during some life
stages. For example, in 2018-2020, about 12 percent of non-Hispanic Black; 9 percent of non-
Hispanic American Indian/Alaska Native; and 7 percent of Hispanic children were estimated to
currently have asthma, compared with 6 percent of non-Hispanic White children.194 Nationally,
on average, non-Hispanic Black and Non-Hispanic American Indian or Alaska Native people
also have lower than average life expectancy based on 2019 data, the latest year for which CDC
estimates are available.195
As discussed in Chapter 4.1.7 of this document, concentrations of many air pollutants are
elevated near high-traffic roadways, and populations who live, work, or go to school near high-
traffic roadways experience higher rates of numerous adverse health effects, compared to
populations far away from major roads.
We conducted an analysis of the populations living in close proximity to truck freight routes
as identified in USDOT's FAF4.196 FAF4 is a model from the USDOT's Bureau of
Transportation Statistics (BTS) and Federal Highway Administration (FHWA), which provides
data associated with freight movement in the U.S.T Relative to the rest of the population, people
living near FAF4 truck routes are more likely to be people of color and have lower incomes than
the general population. People living near FAF4 truck routes are also more likely to live in
metropolitan areas. Even controlling for region of the country, county characteristics, population
density, and household structure, race, ethnicity, and income are significant determinants of
whether someone lives near a FAF4 truck route. We note that we did not analyze the population
living near warehousing, distribution centers, transshipment, or intermodal freight facilities.
We additionally analyzed national databases that allowed us to evaluate whether homes and
schools were located near a major road and whether disparities in exposure may be occurring in
these environments. Until 2009, the U.S. Census Bureau's American Housing Survey (AHS)
included descriptive statistics of over 70,000 housing units across the nation and asked about
transportation infrastructure near respondents' homes.197'u We also analyzed the U.S. Department
of Education's Common Core of Data (CCD), which includes enrollment and location
information for schools across the U.S.198
T FAF4 includes data from the 2012 Commodity Flow Survey (CFS), the Census Bureau on international trade, as
well as data associated with construction, agriculture, utilities, warehouses, and other industries. FAF4 estimates the
modal choices for moving goods by trucks, trains, boats, and other types of freight modes. It includes traffic
assignments, including truck flows on a network of truck routes.
https://ops.flrw a. dot.gov/freight/freight_analysis/faf/.
u The 2013 AHS again included the "etrans" question about highways, airports, and railroads within half a block of
the housing unit but has not maintained the question since then.
198
-------
In analyzing the 2009 AHS, we focused on whether a housing unit was located within 300
feet of a "4-or-more lane highway, railroad, or airport" (this distance was used in the AHS
analysis).v We analyzed whether there were differences between households in such locations
compared with those in locations farther from these transportation facilities.199 We included
other variables, such as land use category, region of country, and housing type. We found that
homes with a non-White householder were 22-34 percent more likely to be located within 300
feet of these large transportation facilities than homes with White householders. Homes with a
Hispanic householder were 17-33 percent more likely to be located within 300 feet of these large
transportation facilities than homes with non-Hispanic householders. Households near large
transportation facilities were, on average, lower in income and educational attainment and more
likely to be a rental property and located in an urban area compared with households more
distant from transportation facilities.
In examining schools near major roadways, we used the CCD from the U.S. Department of
Education, which includes information on all public elementary and secondary schools and
school districts nationwide.200 To determine school proximities to major roadways, we used a
geographic information system (GIS) to map each school and roadways based on the U.S.
Census's TIGER roadway file.201 We estimated that about 10 million students attend schools
within 200 meters of major roads, about 20 percent of the total number of public school students
in the U.S.W About 800,000 students attend public schools within 200 meters of primary roads,
or about 2 percent of the total. We found that students of color were overrepresented at schools
within 200 meters of primary roadways, and schools within 200 meters of primary roadways had
a disproportionate population of students eligible for free or reduced-price lunches x Black
students represent 22 percent of students at schools located within 200 meters of a primary road,
compared to 17 percent of students in all U.S. schools. Hispanic students represent 30 percent of
students at schools located within 200 meters of a primary road, compared to 22 percent of
students in all U.S. schools.
We also reviewed existing scholarly literature examining the potential for disproportionate
exposure among people of color and people with low socioeconomic status (SES). Numerous
studies evaluating the demographics and socioeconomic status of populations or schools near
roadways have found that they include a greater percentage of residents of color, as well as lower
SES populations (as indicated by variables such as median household income). Locations in
these studies include Los Angeles, CA; Seattle, WA; Wayne County, MI; Orange County, FL;
v This variable primarily represents roadway proximity. According to the Central Intelligence Agency's World
Factbook, in 2010, the United States had 6,506,204 km of roadways, 224,792 km of railways, and 15,079 airports.
Highways thus represent the overwhelming majority of transportation facilities described by this factor in the AHS.
w Here, "major roads" refer to those TIGER classifies as either "Primary" or "Secondary." The Census Bureau
describes primary roads as "generally divided limited-access highways within the Federal interstate system or under
state management." Secondary roads are "main arteries, usually in the U.S. highway, state highway, or county
highway system."
x For this analysis we analyzed a 200-meter distance based on the understanding that roadways generally influence
air quality within a few hundred meters from the vicinity of heavily traveled roadways or along corridors with
significant trucking traffic. See U.S. EPA, 2014. Near Roadway Air Pollution and Health: Frequently Asked
Questions. EPA-420-F-14-044.
199
-------
and the State of California, and nationally.202'203'204'205'206'207'208 Such disparities may be due to
multiple factors . 209>21 °>21 !>212>213
People with low SES often live in neighborhoods with multiple stressors and health risk
factors, including reduced health insurance coverage rates, higher smoking and drug use rates,
limited access to fresh food, visible neighborhood violence, and elevated rates of obesity and
some diseases such as asthma, diabetes, and ischemic heart disease. Although questions remain,
several studies find stronger associations between air pollution and health in locations with such
chronic neighborhood stress, suggesting that populations in these areas may be more susceptible
to the effects of air pollution.214'215'216'217
Several publications report nationwide analyses that compare the demographic patterns of
people who do or do not live near major roadways.218'219'220'221'222'223 Three of these studies
found that people living near major roadways are more likely to be people of color or low in
SES.224'225'226 They also found that the outcomes of their analyses varied between regions within
the U.S. However, only one such study looked at whether such conclusions were confounded by
living in a location with higher population density and how demographics differ between
locations nationwide.227 In general, it found that higher density areas have higher proportions of
low-income residents and people of color. In other publications based on a city, county, or state,
the results are similar.228'229
Two recent studies provide strong evidence that reducing emissions from heavy-duty vehicles
is extremely likely to reduce the disparity in exposures to traffic-related air pollutants, both using
NO2 observations from the recently launched TROPospheric Ozone Monitoring Instrument
(TROPOMI) satellite sensor as a measure of air quality, which provides the highest-resolution
observations heretofore unavailable from any satellite.230
One study evaluated satellite-based NO2 concentrations during the COVID-19 lockdowns in
2020 and compared them to NO2 concentrations from the same dates in 2019.231 That study
found that average NO2 concentrations were highest in areas with the lowest percentage of white
populations, and that the areas with the greatest percentages of non-White or Hispanic
populations experienced the greatest declines in NO2 concentrations during the lockdown. These
NO2 reductions were associated with the density of highways in the local area.
In the second study, satellite-based NO2 measured from 2018-2020 was averaged by racial
groups and income levels in 52 large U.S. cities.232 Using census tract-level NO2, the study
reported average population-weighted NO2 levels to be 28% higher in low-income non-White
people compared with high-income White people. The study also used weekday-weekend
differences and bottom-up emission estimates to estimate that diesel traffic is the dominant
source of NO2 disparities in the studied cities. Overall, there is substantial evidence that people
who live or attend school near major roadways are more likely to be of a non-White race,
Hispanic, and/or have a low SES. Although proximity to an emissions source is an indicator of
potential exposure, it is important to note that the impacts of emissions from tailpipe sources are
not limited to communities in close proximity to these sources. For example, the effects of
potential decreases in emissions from sources affected by this final rule might also be felt many
miles away, including in communities with EJ concerns. The spatial extent of these impacts
depends on a range of interacting and complex factors including the amount of pollutant emitted,
atmospheric lifetime of the pollutant, terrain, atmospheric chemistry and meteorology. However,
200
-------
recent studies using satellite-based NO2 measurements provide evidence that reducing emission
from heavy-duty vehicles is likely to reduce disparities in exposure to traffic-related pollution.
In Chapter 6.4.9 of this RIA, we also present an analysis of how the air quality impacts
from this rule are distributed among different populations, specifically focusing on PM2.5 and
ozone concentrations in the contiguous U.S. Using air quality modeling results from the
proposal, we assessed whether areas with the worst projected baseline air quality in 2045 have
larger numbers of people of color living in them, and if those with the worst projected air quality
would benefit more from the final rule. We found that in the 2045 baseline, nearly double the
number of people of color live within areas with the worst air quality, compared to non-Hispanic
Whites (NH-Whites). We also found that the largest improvements in both ozone and PM2.5 are
estimated to occur in these areas with the worst baseline air quality. When we consider the
national implications of this rule on specific race and ethnic groups, non-Hispanic Blacks will
benefit the most from PM2.5 and ozone air quality improvements in both absolute and relative
(percent change from baseline) terms. Although the spatial resolution of the air quality modeling
is not sufficient to capture very local heterogeneity of human exposures, particularly the
pollution concentration gradients near roads, the analysis does allow estimates of demographic
trends at a national scale. See Chapter 6.4.9 of the RIA for additional information on the
demographic analysis.
In summary, there is substantial evidence that people who live or attend school near
major roadways are more likely to be people of color, Hispanic ethnicity, and/or low
socioeconomic status. This final rule will reduce emissions that contribute to NO2 and other
near-roadway pollution, improving air quality for the 72 million people who live near major
truck routes and are already overburdened by air pollution from diesel emissions. Heavy-duty
vehicles also contribute to regional concentrations of ozone and PM2.5 The emission reductions
from this final rule will result in widespread reductions in such pollution. The largest predicted
improvements in both ozone and PM2.5 are estimated to occur in areas with the worst baseline air
quality, and a larger number of people of color are projected to reside in these areas.
Chapter 4 References
1 EPA (2009) Metabolically-derived ventilation rates: A revised approach based upon oxygen
consumption rates. Washington, DC: Office of Research and Development. EPA/600/R-
06/129F. http://cfpub. epa.gov/ncea/cfm/recordisplay. cfm ?deid=202543.
2 U.S. EPA Integrated Science Assessment for Particulate Matter (Final Report, 2019). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019. Chapter 4
"Overall Conclusions" p. 4-1.
3 Foos, B.; Marty, M.; Schwartz, J.; Bennet, W.; Moya, J.; Jarabek, A.M.; Salmon, A.G. (2008)
Focusing on children's inhalation dosimetry and health effects for risk assessment: An
introduction. J Toxicol Environ Health 71 A: 149-165.
201
-------
4 EPA (2006) A Framework for Assessing Health Risks of Environmental Exposures to Children.
EPA, Washington, DC, EPA/600/R-05/093F, 2006.
5 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. https://www3.epa.gov/airtoxics/childrens supplementJinal.pdf.
6 U.S. EPA. America's Children and the Environment. . Available at:
https://www.epa.gov/americaschildrenenvironment.
7 U.S. EPA. Integrated Science Assessment (ISA) for Ozone and Related Photochemical
Oxidants (Final Report). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-
20/012, 2020.
8 U.S. EPA. Policy Assessment (PA) for the Review of the National Ambient Air Quality
Standards for Particulate Matter (Final Report, 2020). U.S. Environmental Protection Agency,
Washington, DC, EPA/452/R-20/002, 2020.
9 U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019. Table 2-1.
10 U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019. Table 2-1.
11 U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
12 U.S. EPA. Supplement to the 2019 Integrated Science Assessment for Particulate Matter (Final
Report, 2022). U.S. Environmental Protection Agency, Washington, DC, EPA/635/R-22/028,
2022.
13 U.S. EPA. Policy Assessment (PA) for the Reconsideration of the National Ambient Air
Quality Standards for Particulate Matter (Final Report, 2022). U.S. Environmental Protection
Agency, Washington, DC, EPA/452/R-22-004, 2022
14 U.S. EPA. (2009). Integrated Science Assessment for Particulate Matter (Final Report). U.S.
Environmental Protection Agency, Washington, DC, EPA/600/R-08/139F.
15 U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
16 U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
17 U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
18 U.S. EPA. Policy Assessment (PA) for the Reconsideration of the National Ambient Air
Quality Standards for Particulate Matter (Final Report, 2022). U.S. Environmental Protection
Agency, Washington, DC, EPA-452/R-22-004, 2022, p. 3-53.
202
-------
19 U.S. EPA. Integrated Science Assessment for Oxides of Nitrogen - Health Criteria (2016 Final
Report). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-15/068, 2016.
20 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/cfim/recordisplay. cfm?deid=218686.
21 U.S. EPA. (1999). Guidelines for Carcinogen Risk Assessment. Review Draft. NCEA-F-
0644, July. Washington, DC: U.S. EPA. Retrieved on March 19, 2009 from
http://cfpub. epa.gov/ncea/cfm/recordisplay. cfm?deid=54932.
22 U.S. EPA (2002). Health Assessment Document for Diesel Engine Exhaust. EPA/600/8-
90/057F Office of research and Development, Washington DC. Retrieved on March 17, 2009
from http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=29060. pp. 1-1 1-2.
23 Garshick, Eric, Francine Laden, Jaime E. Hart, Mary E. Davis, Ellen A. Eisen, and Thomas J.
Smith. 2012. Lung cancer and elemental carbon exposure in trucking industry workers.
Environmental Health Perspectives 120(9): 1301-1306.
24 Silverman, D. T., Samanic, C. M., Lubin, J. H., Blair, A. E., Stewart, P. A., Vermeulen, R., &
Attfield, M. D. (2012). The diesel exhaust in miners study: a nested case-control study of lung
cancer and diesel exhaust. Journal of the National Cancer Institute.
25 Olsson, Ann C., et al. "Exposure to diesel motor exhaust and lung cancer risk in a pooled
analysis from case-control studies in Europe and Canada." American journal of respiratory and
critical care medicine 183.7 (2011): 941-948.
26IARC [International Agency for Research on Cancer], (2013). Diesel and gasoline engine
exhausts and some nitroarenes. IARC Monographs Volume 105. [Online at
http ://monographs.iarc.fr/ENG/Monographs/vol 105/index.php]
27 U.S. EPA (2022) Technical Support Document EPA Air Toxics Screening Assessment.
2017AirToxScreen TSD. https://www.epa.gov/system/files/documents/2022-
03/airtoxscreen_2017tsd.pdf
28 U.S. EPA (2022) 2018 AirToxScreen Risk Drivers.
https://www.epa.gov/AirToxScreen/airtoxscreen-risk-drivers.
29 U.S. EPA. (2000). Integrated Risk Information System File for Benzene. This material is
available electronically at:
https://cfpub.epa. gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=276.
30 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.
31 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.
203
-------
32 U.S. EPA. (2000). Integrated Risk Information System File for Benzene. This material is
available electronically at:
https://cfpub. epa.gov/ncea/iris2/chemicalLanding. cfm?substance_nmbr=276.
33 International Agency for Research on Cancer (IARC, 2018. Monographs on the evaluation of
carcinogenic risks to humans, volume 120. World Health Organization - Lyon, France.
http://publications.iarc.fr/Book-And-Report-Series/Iarc-Monographs-On-The-Identification-Of-
Carcinogenic-Hazards-To-Humans/Benzene-2018.
34NTP (National Toxicology Program). 2016. Report on Carcinogens, Fourteenth Edition.;
Research Triangle Park, NC: U.S. Department of Health and Human Services, Public Health
Service, https://ntp.niehs.nih.gov/go/rocl4
35 Aksoy, M. (1989). Hematotoxicity and carcinogenicity of benzene. Environ. Health Perspect.
82: 193-197. EPA-HQ-OAR-2011-0135.
36 Goldstein, B.D. (1988). Benzene toxicity. Occupational medicine. State of the Art Reviews.
3: 541-554.
37Rothman, 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.
38 U.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
https://cfpub.epa.gov/ncea/iris/iris_documents/documents/toxreviews/0276tr.pdf.
39 Qu, 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.
40 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.
41 Lan, Qing, Zhang, L., Li, G., Vermeulen, R., et al. (2004). Hematotoxically in Workers
Exposed to Low Levels of Benzene. Science 306: 1774-1776.
42 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.
43 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.
204
-------
44 Corti, M; Snyder, CA. (1996) Influences of gender, development, pregnancy and ethanol
consumption on the hematotoxicity of inhaled 10 ppm benzene. Arch Toxicol 70:209-217.
45McKinney P.A.; Alexander, F.E.; Cartwright, R.A.; et al. (1991) Parental occupations of
children with leukemia in west Cumbria, north Humberside, and Gateshead, Br Med J 302:681-
686.
46 Keller, KA; Snyder, CA. (1986) Mice exposed in utero to low concentrations of benzene
exhibit enduring changes in their colony forming hematopoietic cells. Toxicology 42:171-181.
47Keller, KA; Snyder, CA. (1988) Mice exposed in utero to 20 ppm benzene exhibit altered
numbers of recognizable hematopoietic cells up to seven weeks after exposure. Fundam Appl
Toxicol 10:224-232.
48 Corti, M; Snyder, CA. (1996) Influences of gender, development, pregnancy and ethanol
consumption on the hematotoxicity of inhaled 10 ppm benzene. Arch Toxicol 70:209-217.
49 U. S. EPA. (2002). Toxicological Review of Benzene (Noncancer Effects). National Center for
Environmental Assessment, Washington, DC. Report No. EPA/635/R-02/001F.
https://cfpub.epa.gov/ncea/iris/iris documents/documents/toxreviews/0276tr.pdf.
50 Ford, AM; Pombo-de-Oliveira, MS; McCarthy, KP; MacLean, JM; Carrico, KC; Vincent, RF;
Greaves, M. (1997) Monoclonal origin of concordant T-cell malignancy in identical twins. Blood
89:281-285.
51 EPA. Integrated Risk Information System. Formaldehyde (CASRN 50-00-0)
https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=419.
52 NTP (National Toxicology Program). 2016. Report on Carcinogens, Fourteenth Edition.;
Research Triangle Park, NC: U.S. Department of Health and Human Services, Public Health
Service, https://ntp.niehs.nih.gov/go/rocl4.
53IARC Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 88 (2006):
Formaldehyde, 2-Butoxyethanol and l-tert-Butoxypropan-2-ol.
54 IARC Monographs on the Evaluation of Carcinogenic Risks to Humans Volume 100F (2012):
Formaldehyde.
55 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.
56 Hauptmann, 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.
57 Beane 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.
205
-------
58 Pinkerton, L. E. 2004. Mortality among a cohort of garment workers exposed to
formaldehyde: an update. Occup. Environ. Med. 61: 193-200.
59 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.
60 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.
61ATSDR. 1999. Toxicological Profile for Formaldehyde, U.S. Department of Health and
Human Services (HHS), July 1999.
62 ATSDR. 2010. Addendum to the Toxicological Profile for Formaldehyde. U.S. Department of
Health and Human Services (HHS), October 2010.
63 IPCS. 2002. Concise International Chemical Assessment Document 40. Formaldehyde. World
Health Organization.
64 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/63 5/R-10/002A. U.S.
Environmental Protection Agency, Washington DC [online]. Available:
http://cfpub. epa.gov/ncea/iris_drafts/recordisplay. cfm?deid=223614.
65 NRC (National Research Council). 2011. Review of the Environmental Protection Agency' s
Draft IRIS Assessment of Formaldehyde. Washington DC: National Academies Press.
http: //b ooks. nap. edu/op enb ook. php ? record_i d= 13142.
66U.S. EPA (2018). See
https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=419.
67 U.S. EPA. IRIS Toxicological Review of Formaldehyde-Inhalation (Interagency Science
Consultation Draft, 2021). U.S. Environmental Protection Agency, Washington, DC,
EPA/635/R-21/286, 2021.
68 https://www.nationalacademies.org/our-work/review-of-epas-2021-draft-formaldehyde-
assessment
69U.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
https://cfpub.epa. gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=290.
70U.S. EPA (1991). Integrated Risk Information System File of Acetaldehyde. This material is
available electronically at
https://cfpub. epa.gov/ncea/iris2/chemicalLanding. cfm?substance_nmbr=290.
206
-------
71 NTP (National Toxicology Program). 2016. Report on Carcinogens, Fourteenth Edition.;
Research Triangle Park, NC: U.S. Department of Health and Human Services, Public Health
Service, https://ntp.niehs.nih.gov/go/rocl4.
72 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.
73 U.S. EPA (1991). Integrated Risk Information System File of Acetaldehyde. This material is
available electronically at
https://cfpub.epa. gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=290.
74 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
https://cfpub.epa. gov/ncea/iris2/chemicalLanding.cfm?substance_nmbr=364.
75 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.
76Myou, 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.
77 California OEHHA, 2014. TSD for Noncancer RELs: Appendix D. Individual, Acute, 8-Hour,
and Chronic Reference Exposure Level Summaries. December 2008 (updated July 2014).
https://oehha.ca.gov/media/downloads/crnr/appendixdlfinal.pdf
78 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
https://cfpub.epa. gov/ncea/iris_drafts/recordisplay.cfm?deid=56434.
79 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
https://cfpub.epa. gov/ncea/iris_drafts/recordisplay.cfm?deid=56434.
80 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
https://cfpub.epa. gov/ncea/iris_drafts/recordisplay.cfm?deid=56434.
81U. 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
207
-------
material is available electronically at
https://cfpub.epa.gov/ncea/iris_drafts/recordisplay.cfm?deid=56434.
82 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/recordisplay. cfm?deid=84403.
83U.S. EPA. (2021) See: https://cfpub.epa.gov/ncea/iris drafts/recordisplay.cfm?deid=340791.
84 NTP (National Toxicology Program). 2016. Report on Carcinogens, Fourteenth Edition.;
Research Triangle Park, NC: U.S. Department of Health and Human Services, Public Health
Service, https://ntp.niehs.nih.gov/go/rocl4.
85 International Agency for Research on Cancer (IARC). (2002). Monographs on the Evaluation
of the Carcinogenic Risk of Chemicals for Humans. Vol. 82. Lyon, France.
86 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
https://cfpub.epa. gov/ncea/iris_drafts/recordisplay.cfm?deid=56434.
87 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
https://cfpub.epa. gov/ncea/iris_drafts/recordisplay.cfm?deid=56434.
88ATSDR. Toxicological Profile for Naphthalene, 1-Methylnaphthalene, and 2-
Methylnaphthalene (2005). https://www.atsdr.cdc.gov/ToxProfiles/tp67-p.pdf
89ATSDR. Letter Health Consultation, Radiac Abrasives, Inc., Chicago, Illinois (2014).
https://www.atsdr.cdc.gov/HAC/pha/RadiacAbrasives/Radiac%20Abrasives,%20Inc.%20_%20L
HC%20(Final)%20_%2003-24-2014%20(2)_508.pdf
90 U. S. EPA. Derivation of an acute reference concentration for inhalation exposure to
naphthalene. Report No. EPA/600/R-21/292.
https://cfpub.epa.gov/ncea/risk/recordisplay.cfm?deid=355035.
91 U.S. EPA Integrated Risk Information System (IRIS) database is available at:
www.epa.gov/iris.
92 Karner, A.A.; Eisinger, D.S.; Niemeier, D.A. (2010). Near-roadway air quality: synthesizing
the findings from real-world data. Environ Sci Technol 44: 5334-5344.
93 McDonald, B.C.; McBride, Z.C.; Martin, E.W.; Harley, R.A. (2014) High-resolution mapping
of motor vehicle carbon dioxide emissions. J. Geophys. Res.Atmos.,119, 5283-5298,
doi: 10.1002/2013JD021219.
94Kimbrough, S.; Baldauf, R.W.; Hagler, G.S.W.; Shores, R.C.; Mitchell, W.; Whitaker, D.A.;
Croghan, C.W.; Vallero, D.A. (2013) Long-term continuous measurement of near-road air
208
-------
pollution in Las Vegas: seasonal variability in traffic emissions impact on air quality. Air Qual
AtmosHealth 6: 295-305. DOI 10.1007/sll869-012-0171-x
95 Kimbrough, S.; Palma, T.; Baldauf, R.W. (2014) Analysis of mobile source air toxics
(MSATs)—Near-road VOC and carbonyl concentrations. Journal of the Air &Waste
Management Association, 64:3, 349-359, DOI: 10.1080/10962247.2013.863814
96 Kimbrough, S.; Owen, R.C.; Snyder, M.; Richmond-Bryant, J. (2017) NO to NO2 Conversion
Rate Analysis and Implications for Dispersion Model Chemistry Methods using Las Vegas,
Nevada Near-Road Field Measurements. Atmos Environ 165: 23-24.
97 Hilker, N.; Wang, J.W.; Jong, C-H.; Healy, R.M.; Sofowote, U.; Debosz, J.; Su, Y.; Noble, M.;
Munoz, A.; Doerkson, G.; White, L.; Audette, C.; Herod, D.; Brook, J.R.; Evans, G.J. (2019)
Traffic-related air pollution near roadways: discerning local impacts from background. Atmos.
Meas. Tech., 12, 5247-5261. https://doi.org/10.5194/amt-12-5247-2019.
98 Grivas, G.; Stavroulas, I.; Liakakou, E.; Kaskaoutis, D.G.; Bougiatioti, A.; Paraskevopoulou,
D.; Gerasopoulos, E.; Mihalopoulos, N. (2019) Measuring the spatial variability of black carbon
in Athens during wintertime. Air Quality, Atmosphere & Health (2019) 12:1405-1417.
https://doi. org/10.1007/si1869-019-00756-y.
99 Apte, J.S.; Messier, K.P.; Gani, S.; Brauer, M.; Kirchstetter, T.W.; Lunden, M.M.; Marshall,
J.D.; Portier, C.J.; Vermeulen, R.C.H.; Hamburg, S.P. (2017) High-Resolution Air Pollution
Mapping with Google Street View Cars: Exploiting Big Data. Environ Sci Technol 51: 6999-
7008. https://d0i.0rg/l0.1021/acs.est. 7b00891.
100 Dabek-Zlotorzynska, E.; Celo, V.; Ding, L.; Herod, D.; Jeong, C-H.; Evans, G.; Hilker, N.
(2019) Characteristics and sources of PM2.5 and reactive gases near roadways in two
metropolitan areas in Canada. Atmos Environ 218: 116980.
https://d0i.0rg/l 0.1016/j.atmosenv. 2019.116980.
101 Sarnat, J. A.; Russell, A.; Liang, D.; Moutinho, J.L; Golan, R.; Weber, R.; Gao, D.; Sarnat, S.;
Chang, H.H.; Greenwald, R.; Yu, T. (2018) Developing Multipollutant Exposure Indicators of
Traffic Pollution: The Dorm Room Inhalation to Vehicle Emissions (DRIVE) Study. Health
Effects Institute Research Report Number 196. [Online at:
https://www.healtheffects.org/publication/developing-multipollutant-exposure-indicators-traffic-
pollution-dorm-room-inhalatiori\
102 Gantt, B; Owen, R.C.; Watkins, N. (2021) Characterizing nitrogen oxides and fine particulate
matter near major highways in the United States using the National Near-road Monitoring
Network. Environ Sci Technol 55: 2831-2838. [Online at
https://doi.org/10.1021/acs.est.0c05851]
103 Apte, J.S.; Messier, K.R.; Gani, S.; et al. (2017) High-resolution air pollution mapping with
Google Street View cars: exploiting big data. Environ Sci Technol 51: 6999-7018, [Online at
https://d0i.0rg/l 0.102 l/acs.est.7b00891 ]
209
-------
104 Gu, P.; Li, H.Z.; Ye, Q.; et al. (2018) Intercity variability of particulate matter is driven by
carbonaceous sources and correlated with land-use variables. Environ Sci Technol 52: 52:
11545-11554. [Online at http://dx.doi.org/10.1021/acs.est.8b03833]
105 Liu, W.; Zhang, J.; Kwon, J.l; et 1. (2006). Concentrations and source characteristics of
airborne carbonyl compounds measured outside urban residences. J Air Waste Manage Assoc
56: 1196-1204.
106 Cahill, T.M.; Charles, M.J.; Seaman, V.Y. (2010). Development and application of a sensitive
method to determine concentrations of acrolein and other carbonyls in ambient air. Health
Effects Institute Research Report 149. Available at
https://www.healtheffects.org/system/files/Cahilll49.pdf.
107 Laden, F.; Hart, J.E.; Smith, T.J.; Davis, M.E.; Garshick, E. (2007) Cause-specific mortality
in the unionized U.S. trucking industry. Environmental Health Perspect 115:1192-1196.
108 Peters, A.; von Klot, S.; Heier, M.; Trentinaglia, I.; Hormann, A.; Wichmann, H.E.; Lowel, H.
(2004) Exposure to traffic and the onset of myocardial infarction. New England J Med 351:
1721-1730.
109 Zanobetti, A.; Stone, P.H.; Spelzer, F.E.; Schwartz, J.D.; Coull, B.A.; Suh, H.H.; Nearling,
B.D.; Mittleman, M.A.; Verrier, R.L.; Gold, D.R. (2009) T-wave alternans, air pollution and
traffic in high-risk subjects. Am J Cardiol 104: 665-670.
110 Adar, S.; Adamkiewicz, G.; Gold, D.R.; Schwartz, J.; Coull, B.A.; Suh, H. (2007) Ambient
and microenvironmental particles and exhaled nitric oxide before and after a group bus trip.
Environ Health Perspect 115: 507-512.
111 HEI Panel on the Health Effects of Long-Term Exposure to Traffic-Related Air Pollution
(2022) Systematic review and meta-analysis of selected health effects of long-term exposure to
traffic-related air pollution. Health Effects Institute Special Report 23. [Online at
https://www.healtheffects.org/system/files/hei-special-report-23 l.pdf~\ This more recent review
focused on health outcomes related to birth effects, respiratory effects, cardiometabolic effects,
and mortality.
112 Boogaard, H.; Patton. A.P.; Atkinson, R.W.; Brook, J.R.; Chang, H.H.; Crouse, D.L.; Fussell,
J.C.; Hoek, G.; Hoffman, B.; Kappeler, R.; Kutlar Joss, M.; Ondras, M.; Sagiv, S.K.; Somoli, E.;
Shaikh, R.; Szpiro, A. A.; Van Vliet E.D.S.; Vinneau, D.; Weuve, J.; Lurmann, F.W.; Forastiere,
F. (2022) Long-term exposure to traffic-related air pollution and selected health outcomes: a
systematic review and meta-analysis. Environ Intl 164: 107262. [Online at
https://doi.Org/10.1016/j.envint.2022.107262]
113 Boothe, V.L.; Shendell, D.G. (2008). Potential health effects associated with residential
proximity to freeways and primary roads: review of scientific literature, 1999-2006. J Environ
Health 70: 33-41.
114 Salam, M.T.; Islam, T.; Gilliland, F.D. (2008). Recent evidence for adverse effects of
residential proximity to traffic sources on asthma. Curr Opin Pulm Med 14: 3-8.
210
-------
115 Sun, X.; Zhang, S.; Ma, X. (2014) No association between traffic density and risk of
childhood leukemia: a meta-analysis. Asia Pac J Cancer Prev 15: 5229-5232.
116 Raaschou-Nielsen, O.; Reynolds, P. (2006). Air pollution and childhood cancer: a review of
the epidemiological literature. Int J Cancer 118: 2920-9.
117Boothe, VL.; Boehmer, T.K.; Wendel, A.M.; Yip, F.Y. (2014) Residential traffic exposure
and childhood leukemia: a systematic review and meta-analysis. Am J Prev Med 46: 413-422.
118 National Toxicology Program (2019) NTP Monograph n the Systematic Review of Traffic-
related Air Pollution and Hypertensive Disorders of Pregnancy. NTP Monograph 7.
https://ntp.niehs.nih.gov/ntp/ohat/trap/mgraph/trap final 50H.pdf.
119 Volk, H.E.; Hertz-Picciotto, I.; Delwiche, L.; et al. (2011). Residential proximity to freeways
and autism in the CHARGE study. Environ Health Perspect 119: 873-877.
120 Franco-Suglia, S.; Gryparis, A.; Wright, R.O.; et al. (2007). Association of black carbon with
cognition among children in a prospective birth cohort study. Am J Epidemiol, doi:
10.1093/aje/kwm308. [Online at http://dx.doi.org],
121 Power, M.C.; Weisskopf, M.G.; Alexeef, S.E.; et al. (2011). Traffic-related air pollution and
cognitive function in a cohort of older men. Environ Health Perspect 2011: 682-687.
122 Wu, J.; Wilhelm, M.; Chung, J.; et al. (2011). Comparing exposure assessment methods for
traffic-related air pollution in and adverse pregnancy outcome study. Environ Res 111: 685-
6692.
123 Stenson, C.; Wheeler, A.J.; Carver, A.; et al. (2021) The impact of traffic-related air pollution
on child and adolescent academic performance: a systematic review. Environ Intl 155: 106696
[Online at https://doi.Org/l 0.1016/j .envint.2021.106696]
124 Riediker, M. (2007). Cardiovascular effects of fine particulate matter components in highway
patrol officers. Inhal Toxicol 19: 99-105. doi: 10.1080/08958370701495238
125 Alexeef, S.E.; Coull, B.A.; Gryparis, A.; et al. (2011). Medium-term exposure to traffic-
related air pollution and markers of inflammation and endothelial function. Environ Health
Perspect 119: 481-486. doi:10.1289/ehp. 1002560
126 Eckel. S.P.; Berhane, K.; Salam, M.T.; et al. (2011). Residential Traffic-related pollution
exposure and exhaled nitric oxide in the Children's Health Study. Environ Health Perspect.
doi:10.1289/ehp,1103516.
127 Zhang, J.; McCreanor, J.E.; Cullinan, P.; et al. (2009). Health effects of real-world exposure
diesel exhaust in persons with asthma. Res Rep Health Effects Inst 138. [Online at
http://www .healtheffects. org\.
128 Adar, S.D.; Klein, R.; Klein, E.K.; et al. (2010). Air pollution and the microvasculature: a
cross-sectional assessment of in vivo retinal images in the population-based Multi-Ethnic Study
of Atherosclerosis. PLoS Med 7(11): E1000372. doi: 10.1371/journal.pmed. 1000372. Available
at http://dx.doi.org.
211
-------
129 Kan, H.; Heiss, G.; Rose, K.M.; et al. (2008). Prospective analysis of traffic exposure as a risk
factor for incident coronary heart disease: The Atherosclerosis Risk in Communities (ARIC)
study. Environ Health Perspect 116: 1463-1468. doi: 10.1289/ehp. 11290. Available at
http ://dx. doi. org.
130 McConnell, R.; Islam, T.; Shankardass, K.; et al. (2010). Childhood incident asthma and
traffic-related air pollution at home and school. Environ Health Perspect 1021-1026.
131 Islam, T.; Urban, R.; Gauderman, W.J.; et al. (2011). Parental stress increases the detrimental
effect of traffic exposure on children's lung function. Am J Respir Crit Care Med.
132 Clougherty, J.E.; Levy, J.I.; Kubzansky, L.D.; et al. (2007). Synergistic effects of traffic-
related air pollution and exposure to violence on urban asthma etiology. Environ Health
Perspect 115: 1140-1146.
133 Chen, E.; Schrier, H.M.; Strunk, R.C.; et al. (2008). Chronic traffic-related air pollution and
stress interact to predict biologic and clinical outcomes in asthma. Environ Health Perspect 116:
970-5.
134 Long, D.; Lewis, D.; Langpap, C. (2021) Negative traffic externalities and infant health: the
role of income heterogeneity and residential sorting. Environ and Resource Econ 80: 637-674.
[Online at https://doi.Org/l 0.1007/sl 0640-021 -00601 -w]
135 U.S. EPA (2021). Estimation of Population Size and Demographic Characteristics among
People Living Near Truck Routes in the Conterminous United States. Memorandum to the
Docket.
136 EPA. (2011) Exposure Factors Handbook: 2011 Edition. Chapter 16. Online at
https://www.epa.gov/sites/production/files/2015-09/documents/eJh-Chapter 16.pdf.
137 Pedde, M.; Bailey, C. (2011) Identification of Schools within 200 Meters of U.S. Primary and
Secondary Roads. Memorandum to the docket.
138 Stenson, C.; Wheeler, A.J.; Carver, A.; et al. (2021) The impact of traffic-related air pollution
on child and adolescent academic performance: a systematic review. Environ Intl 155: 106696.
[Online at https://d0i.0rg/l 0.1016/j .envint.2021.106696]
139 Riediker, M.; Cascio, W.E.; Griggs, T.R.; et al. (2004) Particulate matter exposure in cars is
associated with cardiovascular effects in healthy young men. Am J Respir Crit Care Med 169.
[Online at https://doi.Org/10.l 164/rccm.200310-1463OC]
140 Peters, A.; von Klot, S.; Heier, M.; et al. (2004) Exposure to traffic and the onset of
myocardial infarction. New Engl J Med 1721-1730. [Online at
https://d0i.0rg/l 0.1056/NEJMoa040203 ]
141 Adar, S.D.; Gold, D.R.; Coull, B.A.; (2007) Focused exposure to airborne traffic particles and
heart rate variability in the elderly. Epidemiology 18: 95-103 [Online at 351:
https://d0i.0rg/l 0.1097/01 .ede.0000249409.81050.46]
212
-------
142 Sabin, L.; Behrentz, E.; Winer, A.M.; et al. Characterizing the range of children's air pollutant
exposure during school bus commutes. J Expo Anal Environ Epidemiol 15: 377-387. [Online
at http s:// doi. org/10.103 8/sj .jea.7500414]
143 Li, C.; N, Q.; Ryan, P.H.; School bus pollution and changes in the air quality at schools: a
case study. J Environ Monit 11: 1037-1042. [https://doi.org/10.1039/b819458k]
144 Austin, W.; Heutel, G.; Kreisman, D. (2019) School bus emissions, student health and
academic performance. Econ Edu Rev 70: 108-12
145 Adar, S.D.; D.Souza, J.; Sheppard, L.; Adopting clean fuels and technologies on school buses.
Pollution and health impacts in children. Am J Respir Crit Care Med 191. [Online at
http://doi.Org/10.l 164/rccm.201410-19240C]
146 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 https://www.nap.edu/catalog/2097/protecting-visibility-in-national-
parks-and-wilderness-areas
147 Hand et al. "Spatial and Temporal Variability of Haze and its Constituents in the United
States" - IMPROVE Report 0737-5352 (HERO #3121721).
148 Sisler, J.F. 1996. Spatial and seasonal patterns and long-term variability of the composition of
the haze in the United States: an analysis of data from the IMPROVE network. CIRA Report,
ISSN 0737-5352-32, Colorado State University.
149 U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
150 Malm, William. Visibility - The Seeing of Near and Distant Landscape Features. Elsevier,
2016.
151 U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
152 Hand, JL; ; Prenni, AJ; Copeland, S; Schichtel, BA; Malm, WC. (2020). Thirty years of the
Clean Air Act Amendments: Impacts on haze in remote regions of the United States (1990-
2018). Atmos Environ 243: 117865.
153 See Section 169(a) of the Clean Air Act.
154U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
155 73 FR 16486 (March 27, 2008).
156 U.S. EPA. Integrated Science Assessment (ISA) for Ozone and Related Photochemical
Oxidants (Final Report). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-
20/012, 2020.
213
-------
157 73 FR 16492 (March 27, 2008).
158 73 FR 16493-16494 (March 27, 2008).
159 7 3 FR 16490/ 16497 (March 27, 2008).
160 U.S. EPA. Integrated Science Assessment (ISA) for Ozone and Related Photochemical
Oxidants (Final Report). U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-
20/012, 2020.
161 U.S. EPA. Integrated Science Assessment (ISA) for Oxides of Nitrogen, Oxides of Sulfur and
Particulate Matter Ecological Criteria (Final Report). U.S. Environmental Protection Agency,
Washington, DC, EPA/600/R-20/278, 2020.
162 U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
163 U.S. EPA. Integrated Science Assessment (ISA) for Oxides of Nitrogen, Oxides of Sulfur and
Particulate Matter Ecological Criteria (Final Report). U.S. Environmental Protection Agency,
Washington, DC, EPA/600/R-20/278, 2020. See Figure IS-2.
164 Hutchison, R., and C.E. Kraft. 1994. "Hmong Fishing Activity and Fish Consumption."
Journal of Great Lakes Research 20(2):471-487.
165 Peterson, D.E., M.S. Kanarek, M.A. Kuykendall, J.M. Diedrich, H.A. Anderson, P.L.
Remington, and T.B. Sheffy. 1994. "Fish Consumption Patterns and Blood Mercury Levels in
Wisconsin Chippewa Indians." Archives of Environmental Health 49(l):53-58.
166 Joslin, J.D., Kelly, J.M., and van Miegroet, H. 1992. "Soil chemistry and nutrition of North
American spruce-fir stands: evidence for recent change." Journal of Environmental Quality, 21,
12-30.
167DeHayes, D.H., P.G. Schaberg, G.J. Hawley, and G.R. Strimbeck. 1999. "Acid rain impacts
on calcium nutrition and forest health." Bioscience 49(10):789-800.
168 Bricker, S., B. Longstaff, W. Dennison, A. Jones, K. Boicourt, C. Wicks, and J. Woerner.
2007. Effects of Nutrient Enrichment In the Nation's Estuaries: A Decade of Change. NOAA
Coastal Ocean Program Decision Analysis Series No. 26. National Centers for Coastal Ocean
Science, Silver Spring, MD. 328 pp.
169 Valigura, R.A., R.B. Alexander, M.S. Castro, T.P. Meyers, H.W. Paerl, P.E. Stacy, and R.E.
Turner. 2001. Nitrogen Loading in Coastal Water Bodies: An Atmospheric Perspective.
Washington, DC: American Geophysical Union.
170 Hutchinson J; Maynard D; Geiser L. (1996). Air quality and lichens - a literature review
emphasizing the Pacific Northwest, USA. Pacific Northwest Region Air Resource Management
Program; U.S. Forest Service; U.S. Department of Agriculture (USDA).
171 Riddell et al, 2008. "The effect of HN03 gas on the lichen Ramalina menziesii." Flora, 203:
47-54.
214
-------
172 Grantz DA; Garner JHB; Johnson DW. 2003. "Ecological effects of particulate matter."
Environ Int, 29: 213-239.
173 Chameides. W.L., Yu, H., Liu, S.C., Bergin, M., Zhou, X., Mearns, L., Wang, G., Kiang,
C.S., Saylor, R.D., Luo, C. Huang, Y., Steiner, A., and Giorgi, F. 1999. "Case study of the
effects of atmospheric aerosols and regional haze on agriculture: an opportunity to enhance crop
yields in China through emission controls?" PNAS, 96: 13626-13633.
174 Simcik M.F., Eisenreich S.J., 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.
175 Simcik M.F., Eisenreich, S.J., Golden K.A., et al. 1996. "Atmospheric Loading of Polycyclic
Aromatic Hydrocarbons to Lake Michigan as Recorded in the Sediments." Environmental
Science and Technology, 30, 3039-3046.
176 Arzavus 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.
177 Cotrufo, M.F., De Santo A.V., Alfani A., Bartoli G., De Cristofaro A. 1995. "Effects of urban
heavy metal pollution on organic matter decomposition in Quercus ilex L. Woods."
Environmental Pollution, 89(1), 81-87.
178 Niklinska 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.
179 Landers DH; Simonich SL; Jaffe DA; Geiser LH; Campbell DH; Schwindt AR; Schreck CB;
Kent ML; Hafner WD; Taylor HE; Hageman KJ; Usenko S; Ackerman LK; Schrlau JE; Rose
NL; Blett TF; Erway MM. (2008). The Fate, Transport and Ecological Impacts of Airborne
Contaminants in Western National Parks (USA). EPA/600/R-07/138. U.S. Environmental
Protection Agency, Office of Research and Development, NHEERL, Western Ecology Division.
Corvallis, Oregon.
180 U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
181 Irving, P.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.
182 U.S. EPA. (1991). Effects of organic chemicals in the atmosphere on terrestrial plants.
EPA/600/3-91/001.
183 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.
215
-------
184 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.
185 Viskari E-L. (2000). Epicuticular wax of Norway spruce needles as indicator of traffic
pollutant deposition. Water, Air, and Soil Pollut. 121:327-337.
186Ugrekhelidze D, F Korte, GKvesitadze. (1997). Uptake and transformation of benzene and
toluene by plant leaves. Ecotox. Environ. Safety 37:24-29.
187Kammerbauer H, H Selinger, RRommelt, A Ziegler-Jons, D Knoppik, B Hock. (1987). Toxic
components of motor vehicle emissions for the spruce Picea abies. Environ. Pollut. 48:235-243.
188 "Technical Guidance for Assessing Environmental Justice in Regulatory Analysis." Epa.gov,
Environmental Protection Agency, https://www.epa.gov/sites/production/files/2016-
06 documents e/tg 5 6 16) \o. I.pdf. (June 2016).
189 Mohai, P.; Pellow, D.; Roberts Timmons, J. (2009) Environmental justice. Annual Reviews
34: 405-430. https://d0i.0rg/l0.1146/annurev-environ-082508-094348.
190 Rowangould, G.M. (2013) A census of the near-roadway population: public health and
environmental justice considerations. Trans Res D 25: 59-67.
http ://dx. doi. org/10.1016/j. trd. 2013.08.003.
191 Marshall, J.D., Swor, K.R.; Nguyen, N.P (2014) Prioritizing environmental justice and
equality: diesel emissions in Southern California. Environ Sci Technol 48: 4063-4068.
https://doi. org/10.1021/es405167f.
192 Marshall, J.D. (2008) Environmental inequality: air pollution exposures in California's South
Coast Air Basin. Atmos Environ 21: 5499-5503.
https://d0i.0rg/l 0.1016/j.atmosenv.2008.02.005.
193 C. W. Tessum, D. A. Paolella, S. E. Chambliss, J. S. Apte, J. D. Hill, J. D. Marshall, PM2.5
polluters disproportionately and systemically affect people of color in the United States. Sci.
Adv. 7, eabf4491 (2021).
194 http://www.cdc.gov/asthma/most recent data.htm.
195 Arias, E. Xu, J. (2022) United States Life Tables, 2019. National Vital Statistics Report,
Volume 70, Number 19. [Online at https://www.cdc.gov/nchs/data/nvsr/nvsr70/nvsr70-19.pdf]
196 U.S. EPA (2021). Estimation of Population Size and Demographic Characteristics among
People Living Near Truck Routes in the Conterminous United States. Memorandum to the
Docket.
197 U.S. Department of Housing and Urban Development, & U.S. Census Bureau, (n.d.). Age of
other residential buildings within 300 feet. In American Housing Survey for the United States:
2009 (pp. A-l). Retrieved from https://www.census.gov/programs-surveys/ahs/data/2009/ahs-
2009-summary-tables0/hl50-09.html.
198 http://nces.ed.gov/ccd/.
216
-------
199 Bailey, C. (2011) Demographic and Social Patterns in Housing Units Near Large Highways
and other Transportation Sources. Memorandum to docket.
200 http://nces.ed.gov/ccd/.
201 Pedde, M.; Bailey, C. (2011) Identification of Schools within 200 Meters of U.S. Primary and
Secondary Roads. Memorandum to the docket.
202 Marshall, J.D. (2008) Environmental inequality: air pollution exposures in California's South
Coast Air Basin.
203 Su, J.G.; Larson, T.; Gould, T.; Cohen, M.; Buzzelli, M. (2010) Transboundary air pollution
and environmental justice: Vancouver and Seattle compared. GeoJournal 57: 595-608.
doi:10.1007/sl 0708-009-9269-6
204 Chakraborty, J.; Zandbergen, P.A. (2007) Children at risk: measuring racial/ethnic disparities
in potential exposure to air pollution at school and home. J Epidemiol Community Health 61:
1074-1079. doi:10.1136/jech.2006.054130
205 Green, R.S.; Smorodinsky, S.; Kim, J.J.; McLaughlin, R.; Ostro, B. (20042004) Proximity of
California public schools to busy roads. Environ Health Perspect 112: 61-66.
doi:10.1289/ehp.6566
206 Wu, Y; Batterman, S.A. (2006) Proximity of schools in Detroit, Michigan to automobile and
truck traffic. J Exposure Sci & Environ Epidemiol, doi: 10.1038/sj.jes.7500484
207 Su, J.G.; Jerrett, M.; de Nazelle, A.; Wolch, J. (2011) Does exposure to air pollution in urban
parks have socioeconomic, racial, or ethnic gradients? Environ Res 111: 319-328.
208 Jones, M.R.; Diez-Roux, A.; Hajat, A.; et al. (2014) Race/ethnicity, residential segregation,
and exposure to ambient air pollution: The Multi-Ethnic Study of Atherosclerosis (MESA). Am
J Public Health 104: 2130-2137. [Online at: https://doi.org/10.2105/AJPH.2014.302135.]
209 Depro, B.; Timmins, C. (2008) Mobility and environmental equity: do housing choices
determine exposure to air pollution? Duke University Working Paper.
210 Rothstein, R. The Color of Law: A Forgotten History of How Our Government Segregated
America. New York: Liveright, 2018.
211 Lane, H.J.; Morello-Frosch, R.; Marshall, J.D.; Apte, J.S. (2022) Historical redlining is
associated with present-day air pollution disparities in US Cities. Environ Sci & Technol Letters
9: 345-350. DOI: [Online at: https://doi.org/10.1021/acs.estlett.lc01012]
212 Ware, L. (2021) Plessy's legacy: the government's role in the development and perpetuation
of segregated neighborhoods. RSF: The Russel Sage Foundation Journal of the Social Sciences,
7:92-109. DOI: DOI: 10.7758/RSF.2021.7.1.06.
213 Archer, D.N. (2020) "White Men's Roads through Black Men's Homes": advancing racial
equity through highway reconstruction. Vanderbilt Law Rev 73: 1259.
217
-------
214 Clougherty, J.E.; Kubzansky, L.D. (2009) A framework for examining social stress and
susceptibility to air pollution in respiratory health. Environ Health Perspect 117: 1351-1358.
Doi:10.1289/ehp.0900612
215 Clougherty, J.E.; Levy, J.I.; Kubzansky, L.D.; Ryan, P.B.; Franco Suglia, S.; Jacobson
Canner, M.; Wright, R.J. (2007) Synergistic effects of traffic-related air pollution and exposure
to violence on urban asthma etiology. Environ Health Perspect 115: 1140-1146.
doi:10.1289/ehp.9863
216 Finkelstein, M.M.; Jerrett, M.; DeLuca, P.; Finkelstein, N.; Verma, D.K.; Chapman, K.;
Sears, M.R. (2003) Relation between income, air pollution and mortality: a cohort study.
Canadian Med Assn J 169: 397-402.
217 Shankardass, K.; McConnell, R.; Jerrett, M.; Milam, J.; Richardson, J.; Berhane, K. (2009)
Parental stress increases the effect of traffic-related air pollution on childhood asthma incidence.
Proc Natl Acad Sci 106: 12406-12411. doi:10.1073/pnas.0812910106
218 Rowangould, G.M. (2013) A census of the U.S. near-roadway population: public health and
environmental justice considerations. Transportation Research Part D; 59-67.
219 Tian, N.; Xue, J.; Barzyk. T.M. (2013) Evaluating socioeconomic and racial differences in
traffic-related metrics in the United States using a GIS approach. J Exposure Sci Environ
Epidemiol 23: 215-222.
220 CDC (2013) Residential proximity to major highways - United States, 2010. Morbidity and
Mortality Weekly Report 62(3): 46-50.
221 Clark, L.P.; Millet, D.B., Marshall, J.D. (2017) Changes in transportation-related air pollution
exposures by race-ethnicity and socioeconomic status: outdoor nitrogen dioxide in the United
States in 2000 and 2010. Environ Health Perspect https://doi.org/10.1289/EHP959.
222 Mikati, I.; Benson, A.F.; Luben, T.J.; Sacks, J.D.; Richmond-Bryant, J. (2018) Disparities in
distribution of particulate matter emission sources by race and poverty status. Am J Pub Health
https://ajph.aphapublications.org/doi/abs/10.2105/AJPH.2017.304297?journalCode=ajph.
223 Alotaibi, R.; Bechle, M.; Marshall, J.D.; Ramani, T.; Zietsman, J.; Nieuwenhuijsen, M.J.;
Khreis, H. (2019) Traffic related air pollution and the burden of childhood asthma in the
continuous United States in 2000 and 2010. Environ International 127: 858-867.
https://www.sciencedirect.com/science/article/pii/S0160412018325388.
224 Tian, N.; Xue, J.; Barzyk. T.M. (2013) Evaluating socioeconomic and racial differences in
traffic-related metrics in the United States using a GIS approach. J Exposure Sci Environ
Epidemiol 23: 215-222.
225 Rowangould, G.M. (2013) A census of the U.S. near-roadway population: public health and
environmental justice considerations. Transportation Research Part D; 59-67.
226 CDC (2013) Residential proximity to major highways - United States, 2010. Morbidity and
Mortality Weekly Report 62(3): 46-50.
218
-------
227 Rowangould, G.M. (2013) A census of the U.S. near-roadway population: public health and
environmental justice considerations. Transportation Research Part D; 59-67.
228 Pratt, G.C.; Vadali, M.L.; Kvale, D.L.; Ellickson, K.M. (2015) Traffic, air pollution, minority,
and socio-economic status: addressing inequities in exposure and risk. Int J Environ Res Public
Health 12: 5355-5372. http://dx.doi.org/10.3390/ijerphl20505355.
229 Sohrabi, S.; Zietsman, J.; Khreis, H. (2020) Burden of disease assessment of ambient air
pollution and premature mortality in urban areas: the role of socioeconomic status and
transportation. Int J Env Res Public Health doi:10.3390/ijerphl7041166.
230 TROPospheric Ozone Monitoring Instrument (TROPOMI) is part of the Copernicus Sentinel-
5 Precursor satellite.
231 Kerr, G.H.; Goldberg, D.L.; Anenberg, S.C. (2021) COVID-19 pandemic reveals persistent
disparities in nitrogen dioxide pollution. PNAS118. [Online at
https://doi.Org/l 0.1073/pnas.2022409118]
232 Demetillo, M.A.; Harkins, C.; McDonald, B.C.; et al. (2021) Space-based observational
constraints on N02 air pollution inequality from diesel traffic in major US cities. Geophys Res
Lett 48, e2021GL094333. [Online at https://doi.org/10.1029/2021GL094333]
219
-------
Chapter 5 Emissions Inventory
5.1 Introduction
This chapter presents our analysis of the national emissions impacts of the final standards for
calendar years 2027 through 2045. In addition, this chapter describes the methods used to
estimate the spatially and temporally-resolved emission inventories in 2016 and 2045 that
supported the air quality modeling analysis documented in Chapter Chapter 6.
As described in detail in Chapter 5.2, the onroad national inventories were estimated using the
public version of EPA's Motor Vehicle Emission Simulator (MOVES) model, MOVES3. The
onroad national emission inventories were developed using a single national modeling domain,
referred to as "national-scale" in MOVES. Inputs developed to model the national emission
inventories for the final standards are discussed in Chapter 5.2.2. The national emissions
inventory impacts for calendar years 2030, 2040, and 2045 for the final standards are presented
in Chapter 5.3. In addition, the national emissions results for calendar years 2027 through 2045
are presented in Chapter 5.5.4.
As described in Chapter 5.4, MOVES was also used to estimate the emission inventories for
air quality modeling. However, as described in Section VII of the preamble to this rule and also
in Chapter Chapter 6, we did not perform new air quality modeling for the final rule; as
described in Chapter 5.4.2., the emission reductions modeled in the air quality analysis compare
well with those estimated for the final standards.
5.2 Model and Data Updates
To quantify the emissions impacts of the final standards, EPA used the public version of
MOVES available at the time of the FRM analysis, MOVES3.1 MOVES3 includes all the model
updates previously made for "MOVES CTINPRM" (the MOVES model used for the NPRM
analysis), as well as other more recent updates. The additional updates included in MOVES3 are
described in Table 5-1. Detailed descriptions of the underlying data and analyses that informed
the model updates are documented in peer-reviewed technical reports referenced in Table 5-1. In
addition, the peer review materials are available on the EPA's science inventory webpage.2
Finally, the MOVES3 version used to quantify the emissions impacts of the final standards can
be found in the docket.3
220
-------
Table 5-1: Updates to MOVES3 from MOVES CTINPRM
MOVES updates
Description
Heavy-duty running
exhaust emission rates 4
Further update heavy-duty diesel running exhaust emission rates for 2010 and later
model year vehicles using the latest data from the Heavy-Duty In-Use Testing
(HDIUT) Program. In this update, the MY2010+ vehicles are further divided into two
groups - "model year 2010-2013 vehicles" and "model year 2014 and later vehicles"
-to account for differences in emission performance of more recent engines and
aftertreatment systems
Heavy-duty crank case
emission rates Error! liook,m"'k
not defined.
Update heavy-duty diesel crank case emission rates for model year 2007-2009
vehicles (based on certification test data) and for model year 2010 and later vehicles
(using US EPA's National Vehicle and Fuel Emissions Laboratory (NFVEL) testing
data)
Heavy-duty population and
activity information6
Update information on heavy-duty vehicle populations based on vehicle registration
data, FHWA vehicle miles traveled data, and updated projections from the Annual
Energy Outlook (AEO) 2019 5
Glider trucks 6
Update projected glider vehicle sales estimates for model year 2018 and later
Light-duty vehicle and
other changes 7
Updated gasoline fuel properties, light-duty vehicle emissions rates, light-duty
activity, and other changes
5.2.1 Methodology Overview
We used MOVES3 to estimate the emissions impacts of the HD2027 final standards. First, we
estimated emissions for a baseline scenario in which there are no new heavy-duty engine
emission standards. We then estimated emissions for the control scenario, described in Chapter
1. The emissions impacts of the final standards were estimated by calculating the difference
between the emissions estimated in the baseline and the control scenario. All of the model inputs,
MOVES runspec files, and the scripts used for the analysis, as well as the version of MOVES
used to generate the emissions inventories, can be found in the docket.3
In modeling the baseline scenario, we used the MOVES3 default heavy-duty emission rates
updated based on the latest data, as described in Table 5-1. In particular, the updates to the
heavy-duty exhaust and crankcase emission rates resulted in a lower national-scale baseline
inventory for NOx and PM2.5, compared to the baseline national-scale emissions inventory
estimated for the proposal. A Note that because the national emissions inventories used a single
national modeling domain, the baseline scenario did not account for different emission standards
in California or other states that have adopted the California emission standards.6
A Note that as described in preamble Section III, we also made a programmatic adjustment to the proposed
requirement to close crankcases; the combination of this programmatic change and the update to crankcase emission
rates influence the national-scale PM2 5 emissions inventory results for the final rule (see RIA Chapter 5.2.2.4 for
additional detail).
B EPA is reviewing a waiver request under CAA section 209(b) from California for the Omnibus rule; until EPA
grants the waiver, the HD Omnibus program is not enforceable. For more information on the California Air
Resources Board Omnibus rule see, "Heavy-Duty Engine and Vehicle Omnibus Regulation and Associated
Amendments," December 22, 2021. https://ww2.arb.ca.gov/rulemaking/2020/hdomnibuslownox. Last accessed
September 21, 2022. See also "California State Motor Vehicle Pollution Control Standards and Nonroad Engine
Pollution Control Standards; The ' 'Omnibus'' Low NOx Regulation; Request for Waivers of Preemption;
Opportunity for Public Hearing and Public Comment" at 87 FR 35765 (June 13, 2022).
221
-------
The vehicle activity (e.g., fleet age distributions, vehicle miles traveled by vehicle type and
road type, vehicle speeds, off-network idling, hotelling hours, and start activity) and fuel inputs
were kept the same for both baseline and control scenarios, using the default values in
MOVES3.6'7 For example, as shown in Table 5-1 above, future year projections of vehicle
populations and vehicle miles travelled were updated0 to reflect the estimates from the
Department of Energy's Annual Energy Outlook 2019.5 Note that the vehicle activity data for the
emissions inventory used for the air quality analyses in the proposal included local activity data
(as documented in Chapter 5.4) which captures more detailed information than the national
default values.
We used the default data in MOVES for fuel usage and fuel property for the national
inventory runs for calendar years from 2027 through 2045.7 The national emissions inventory
runs aggregate the gasoline fuel properties into a single region representing the United States,
whereas the county-level MOVES runs for the air quality modeling analysis account for county
and regional differences in gasoline fuel properties. Diesel and CNG fuel properties are also
aggregated into one set of nationally-representative fuel properties for each calendar year and
fuel type.
The emission rate inputs developed for modeling the final standards and the proposed Option
2 are discussed in detail in the following section.
5.2,2 MOVES Emission Rates for Control Scenarios
We developed separate MOVES emission rates for the final standards and the proposed
Option 2 (collectively referred to as the control scenarios).0 The methodologies used to develop
the emission rates reflect the effects of this rulemaking's program elements (duty-cycle
standards, off-cycle standards, closed crankcase requirements, refueling standards, regulatory
useful life, and emissions warranty) on vehicles subject to the final rule as described below. We
did not estimate the emission impacts of certain compliance provisions that target long-term
compliance assurance. As we describe in Section IV of the preamble to this rule, we expect the
improved serviceability and updated approach to inducements that we are finalizing will
discourage owners from tampering their engines or emission control systems; however, we had
insufficient data to estimate the impact of these program elements in the final rule.
MOVES has separate emission rates for heavy-duty vehicles according to three different fuel
types—diesel, gasoline, and natural gas (NG) —and six different regulatory vehicle classes,
shown in Table 5-2. The final rule includes new duty-cycle standards for all heavy-duty vehicles
in the LHD45, MHD, HHD, and Urban Bus regulatory classes for all fuel types. The Urban Bus
c The version of MOVES used for the NPRM analysis, MOVES CTI NPRM, had projections based on Annual
Energy Outlook 2018. Most of the changes in the baseline inventory come from the differences between the two
AEO projections.
D The control scenario analyzed for the final standards differs slightly from the final program; we note these minor
differences in relevant tables throughout this chapter.
222
-------
regulatory class was modeled using the same zero-mile emission rates as HHD for the control
scenarios.E
Light heavy-duty Class 2b and 3 trucks (LHD2b3) are composed of vehicles that are both
chassis- and engine-certified. The final standards will apply to all engine-certified LHD2b3
vehicles, which are estimated to be a small fraction of the diesel LHD2b3 vehicles. All Class 2b
and 3 gasoline-fueled vehicles are chassis-certified and will not be affected by the final rule.
A glider vehicle is a new motor vehicle produced with a used or remanufactured engine, and
typically does not include the aftertreatment systems needed to meet the 2007 or 2010 heavy-
duty emission standards.17 In MOVES, glider vehicles are modeled as heavy heavy-duty diesel
vehicles (Class 8 Trucks) that are assumed to emit at a level equivalent to the model year 2000
HHD vehicles.6 Annual glider sales are fixed at 4,000 units per year for 2018 and later based on
the 2018 and later sales volume cap for glider vehicles set forth in the Heavy-duty Greenhouse
Gas Phase 2 Rulemaking8, as well as the number of glider manufacturers and their historic
production levels, as documented in the MOVES population and activity report.6 For the final
rule analysis, we assumed there is no change to the sales of glider vehicles and no change to
glider vehicle emission rates from the baseline scenario.0
Table 5-2: MOVES Heavy-duty Regulatory Classes and Relevant MOVES Fuel Types
MOVES Regulatory Class Description
regClassName
regClassID
Gross Vehicle
Weight Rating
(GVWR) [lb.]
MOVES
Fuel Types
Light Heavy-Duty Class 2b and 3 trucks
LHD2b3
41
8,501 - 14,000
Gasoline,
Diesel
Light Heavy-Duty Class 4 and 5 Trucks
LHD45
42
14,001 - 19,500
Gasoline,
Diesel
Medium Heavy-Duty (Class 6 and 7 Trucks)
MHD
46
19,501 -33,000
Gasoline,
Diesel
Heavy Heavy-Duty (Class 8 Trucks)
HHD
47
>33,000
Diesel, NG
Urban Bus
Urban Bus
48
> 33,000
Diesel, NG
Gliders (Class 8 Trucks)
Glider Vehicles
49
>33,000
Diesel
For heavy-duty diesel vehicles, we developed nitrogen oxide (NOx) emission rates for
running, start, and extended idle processes, as discussed in Chapters 5.2.2.1, 5.2.2.2, and 5.2.2.3,
respectively, in response to the final and proposed Option 2 duty-cycle and off-cycle standards.
E Urban Bus vehicles are also considered HHD vehicles. They are modeled as a separate regulatory class in MOVES
because they had stricter PM emission standards for 1994 through 2006 model years. For the control scenarios, the
Urban Bus emission rates were assumed to be equivalent to HHD except for the age effects discussed in Section
5.2.2.1.2.
F See the definition of "glider vehicle" in 40 CFR 1037.801 and our discussion in the preamble of the heavy-duty
GHG Phase 2 rulemaking (81 FR 73941, October 25, 2016).
G It may be possible that future sales of glider vehicles are lower than assumed in MOVES3 or glider vehicles emit
at a lower level due to the final standards, resulting in lower emissions from gliders. However, due to the uncertainty
associated with these assumptions, the emissions from gliders were kept the same between the baseline and control
scenarios in the current analysis.
223
-------
The lower NOx emissions are anticipated to be achieved using the technologies discussed in
Chapter 3, including cylinder deactivation (CDA) and dual selective catalytic reduction exhaust
aftertreatment system. We also revised heavy-duty diesel emission running rates for NOx, total
hydrocarbons (THC), carbon monoxide (CO), and particulate matter below 2.5 microns (PM2.5),
due to the changes to the regulatory useful life and warranty, by modifying MOVES' tampering
and mal-maintenance (T&M) calculations, as discussed in Chapter 5.2.2.1.2. We also evaluated
the impacts on the PM2.5 crankcase emissions from heavy-duty diesel vehicles due to the closed
crankcase design option for heavy-duty diesel engines11 in the final and proposed Option 2
standards as discussed in Section 5.2.2.4.
For heavy-duty gasoline vehicles, we developed revised NOx, THC, CO, and PM2.5 emission
rates for running exhaust in response to the final and proposed Option 2 duty-cycle standards, as
discussed in Chapter 5.2.2.6. We also revised THC refueling emission rates in response to the
final and proposed Option 2 refueling standard discussed in Chapter 5.2.2.7. We did not revise
the start emission rates for heavy-duty gasoline vehicles to account for the final and proposed
Option 2 standards due to lack of sufficient data to model the impact.
For heavy-duty NG vehicles, we did not estimate reductions in NOx or other pollutants due to
the final and proposed Option 2 heavy-duty spark-ignition duty-cycle standards (discussed in
Section III of the Preamble). As shown in the MOVES heavy-duty exhaust report1, the average
FTP emissions for MY 2010-2017 CNG engine families is close to 0.1 g/hp-hr. We expect small
reductions in NOx emissions from heavy-duty NG engines with the final and proposed Option 2
scenarios. However, in part due to the small contribution of NG emissions to total emissions, we
did not estimate the reductions in NG engines in this analysis.
Similarly, we did not estimate emission reductions from the final and proposed Option 2
useful life and warranty requirements for gasoline and NG vehicles. This is because, in MOVES,
we do not tie emission estimates to heavy-duty gasoline or NG vehicle warranty and useful life
periods, but rather estimate age effects from emissions data, or adapt light-duty gasoline or
heavy-duty diesel age effects. Gasoline and NG-fueled vehicles' contributions to the heavy-duty
vehicle NOx emissions inventory are small (see Figure 5-15), and thus, we expect minimal
impacts on our emission inventory estimates from our conservative approach of estimating
reductions solely from running emissions for these vehicles/
While we estimated different emission rates for the control scenarios, we did not estimate
differences in the speciation of the total organic gases and particulate matter emissions between
the baseline and control scenarios. Emissions speciation refers to the calculation of individual
compounds or classes of compounds within a broader pollutant. For example, MOVES conducts
speciation to estimate benzene from volatile organic compound (VOC) emissions and particle-
H As described in Preamble Section III.B, EPA is finalizing a requirement for manufacturers to use one of two
options for controlling crankcase emissions, either: 1) as proposed, closing the crankcase, or 2) an updated version
of the current requirements for an open crankcase that includes additional requirements for measuring and
accounting for crankcase emissions.
1 See Table 4-2 in Exhaust Emission Rates for Heavy-Duty Onroad Vehicles in MOVES3
1 According to MOVES, the contributions from NG vehicles to HD NOx inventory in calendar year 2045 are: 0.6%
(baseline), 1.1% (final standards), and 1.0% (proposed Option 2). These contributions are dependent on the current
projections of NG in the future heavy-duty fleet and the emission rates of NG vehicles, main of which are already
meeting the 0.02 g/hp-hr NOx standard.
224
-------
phase naphthalene from particulate matter (PM) emissions. Speciation is important to estimate
individual toxics and necessary for air quality modeling. However, we do not have sufficient data
to support differences in the emissions speciation between the control case and baseline
scenarios. Heavy-duty diesel vehicles in the control case scenarios apply the same speciation
values as the baseline scenario for heavy-duty diesel vehicles model year 2010 and later.
Similarly, emissions from heavy-duty gasoline vehicles in the control scenarios use the same
speciation as heavy-duty gasoline vehicles in the baseline scenario. Changes in emissions for
individual compounds presented in Section 5.3 (acetaldehyde, benzene, formaldehyde, and
naphthalene) are due to changes in the hydrocarbon and particulate matter emission rates for the
control scenarios presented in this section.
5.2.2.1 Heavy-Duty Diesel Running Emission Rates
This section documents the approach used to develop the heavy-duty diesel emission running
rates for the control scenarios. We first estimated new zero-mile NOx emission rates in response
to the final and proposed Option 2 duty-cycle and off-cycle standards as discussed in Chapter
5.2.2.1.1. The zero-mile emission rate in MOVES is defined as the emission rate of a new
vehicle without any tampering and mal-maintenance effects. We then estimated the effects of
lengthened regulatory useful life and warranty periods and applied them to the aged NOx, THC,
CO and PM2.5 emission rates for the final and proposed Option 2 standards in Chapter 5.2.2.1.2
Emission rates in MOVES are defined by the operating mode, regulatory class, model year,
and fuel type. The running operating modes are defined in Table 5-3.
225
-------
Table 5-3: MOVES Running Operating Mode Definitions
OpModelD
Operating Mode
Vehicle Speed
(v, mph)
Scaled Tractive
Power (STP, skW)
0
Deceleration/B raking
All speeds
1
Idle
v < 1.0
11
Coast
1
-------
Table 5-4: Heavy-duty Compression Ignition Duty-Cycle NOx Standards for the Final and Proposed Option 2
Scenarios
Scenario
Applicable Model Years
Regulatory
Classes
LLC (g/hp-hr)
FTP
(g/hp-hr)
SET-RMC
(g/hp-hr)
Baseline
Model Year 2010+
LHD, MHD,
HHD
-
0.2
0.2
Final
Standards
Model Year 2027+
LHD, MHD,
HHD
0.05 [0.065]a
0.035 [0.05]A
0.035 [0.05]A
Proposed
Option 2
Model Year 2027+
LHD, MHD,
HHD
0.1
0.05
0.05
A Values in brackets denote standards that were only applied to HHD engines in the modeling of the final control
scenario; in the final program, these values will apply to bothMHD and HHD, see preamble Section III.B for details.
The final and proposed Option 2 scenarios also include revised standards for PM emissions as
discussed in Section III of the preamble. We did not model the revised PM standards for heavy-
duty diesel vehicles in MOVES, because the revised PM standards are intended to prevent
backsliding of PM reductions already achieved with current heavy-duty diesel emission control
systems. We estimate reductions in heavy-duty diesel PM emissions will occur due to the
lengthened warranty and useful life periods, discussed in Chapter 5.2.2.1.2, and due to the
crankcase emissions control, as discussed in Chapter 5.2.2.4.
We used the final and proposed Option 2 standards for the FTP and (SET-RMC) to estimate
the effect of the duty-cycle standards on MOVES NOx emission rates. Because we do not have
sufficient (LLC) test data on existing heavy-duty diesel vehicles to develop the modeling inputs
specific for the LLC standard in MOVES, we used the FTP standard to model the impact of the
standards on low-power operation.
Equation 5-1 through Equation 5-6 were used to incorporate the effects of more stringent FTP
and SET-RMC engine duty-cycle emission standards on MOVES running exhaust NOx emission
rates for model years subject to the final and proposed Option 2 standards. The term Rduty is the
ratio between the final or proposed Option 2 emission standards and the current FTP and SET-
RMC duty-cycle standards (0.2 g/hp-hr).
Equation 5-1
Final or proposed Option 2 FTP or SET RMC standard
duty Current standard
Rduty ranges between 17 and 25 percent for the control scenarios considered, as shown in
Table 5-5.
227
-------
Table 5-5: Rduty Ratios Calculated for Each Scenario
Scenario
Applicable Model
Years
Regulatory Classes
FTP and SET
standard
(g/hp-hr)
Rduty
Final
Standards
2027+
LHD, MHD
0.035
17.5%
HHD
0.05A
25%
Proposed
Option 2
2027+
LHD, MHD, HHD
0.05
25%
A Values in this row denote standards that were only applied to HHD engines in the modeling
scenario for the final rule analysis; in the final program, these values will apply to both MHD and
HHD, see preamble Section III.B for details
To estimate the effect of final and proposed Option 2 engine dynamometer duty-cycle
standards on in-use emissions, we used the relationship between reductions in the most recent
duty-cycle standard compared to reductions in in-use emissions. The 2010 0.2 g/hp-hr NOx
emission standard9 is the most recent heavy-duty NOx emission standard. To evaluate the in-use
effectiveness of the 2010 standard, we compared the in-use NOx emission rates from vehicles
that were certified to the previous heavy-duty NOx standard and the 2010 standard. Equation 5-2
defines Rin_use as the ratio between the percent change observed in-use from vehicles compliant
with the 2010 NOx standard relative to vehicles compliant with the previous standard, and the
percent change in the 2010 standard FTP standards relative to the previous standard.K In other
words, Rin Use is the ratio between the relative effectiveness of reducing in-use emissions
compared to the relative reduction in the FTP duty-cycle emission standard.
Equation 5-2
% Change in the inuse emission rates from 2010 compliant vehicles
^in-use % Change in the 2010 FTP standard
The percent change in in-use emission rates from vehicles certified to the 2010 standard (the
numerator in Equation 5-2) was estimated using Equation 5-3. The MOVES emission rates for
HHD vehicles certified to the 2010 0.2 g/hp-hr standard (the numerator) are calculated from 93
MY 2010-2015 HHD vehicles with engine family emission limit (FEL) certified below the 0.2
g/hp-hr NOx emissions level and tested as part of the Heavy-Duty In-Use Testing program.Error!
Bookmark not defined, MOVES emission rates for HHD vehicles certified to the 2004-2006
standard (the denominator) are based on 91 MY 2003-2006 trucks from two in-use datasets:
ROVER data collected by the US EPA and the Heavy-Duty Diesel Consent Decree data
collected by West Virginia University. The in-use datasets and analysis used to derive the
emission rates in MOVES3 are documented in the MOVES3 heavy-duty exhaust technical
ranAt4 Error! Bookmark not defined.
K In 2004-2006, the NMHC+NOx emission standard was 2.4 g/hp-hr; the 0.2 g/hp-hr NOx standard began to be
phased-in starting in 2007, with a full-phase in 2010.
228
-------
Equation 5-3
% Change in the in_use emission rates from 2010 compliant vehicles
Emission rate from HHD 2010 compliant vehicles
HHD MY 2006 MOVES emission rate
The percent change from Equation 5-3 was applied to each MOVES operating mode to
evaluate the effectiveness of the 2010 standard across different ranges of in-use operating
conditions, as shown in Figure 5-1 and Table 5-6. The emission reduction is larger for the higher
speed and higher load MOVES operating modes, with the largest decrease observed for speeds
above 50 mph (operating modes 33 through 35). The lowest effectiveness of the standards is
observed for low speed and several low power operating modes (operating mode 1,11, and 21),
with an exception of the deceleration bin (operating mode 0).
Equation 5-4 was used to estimate the percent reduction between the 2010 standard and the
2004-2006 emission standard. This term is also the denominator of Equation 5-2. The percent
reduction in the NOx emission standard was estimated assuming thatNOx emissions consist of
70 percent of the combined NMHC (Non-Methane Hydrocarbons) + NOx 2004-2006 emission
standard (2.4 g/hp-hr), consistent with the assumption used in MOVES3.Error! Bookmark not defined.
This value is also plotted as a line in Figure 5-1 to compare to the in-use emission rate
reductions.
Equation 5-4
0.2
% Change in the 2010 FTP standard = — 1 = —88.1%
5 (70% x 2.4)
229
-------
dl
¦*->
ro
tao
x
O
50
230
-------
Table 5-6: Calculation of Rin_use by MOVES Operation Mode
MOVES
OpMode
HHD MOVES
MY 2006 NOx
emission rates
(g/hr)
NOx Emission rate
from 2010 compliant
HHD vehicles (g/hr)a
Percent change in in-use
NOx emission rates from
2010 compliant vehicles
(%)
Rin use
0
0.038
0.0031
-91.8
1.04
1
0.015
0.0063
-57.9
0.66
11
0.015
0.0106
-28.8
0.33
12
0.058
0.0210
-63.5
0.72
13
0.093
0.0339
-63.7
0.72
14
0.127
0.0453
-64.4
0.73
15
0.145
0.0565
-60.9
0.69
16
0.188
0.0734
-60.9
0.69
21
0.010
0.0066
-32.1
0.36
22
0.064
0.0163
-74.4
0.84
23
0.093
0.0243
-73.9
0.84
24
0.132
0.0359
-72.7
0.83
25
0.165
0.0485
-70.6
0.80
27
0.225
0.0633
-71.8
0.82
28
0.244
0.0558
-77.2
0.88
29
0.314
0.88b
30
0.384
0.88
33
0.051
0.0062
-87.8
1.00
35
0.148
0.0130
-91.2
1.04
37
0.226
0.0323
-85.7
0.97
38
0.268
0.0306
-88.6
1.01
39
0.345
1.01
40
0.422
1.01
Notes:
a The HHD rates in this table are based on fscaie of 17.1 metric tons to be consistent with the
fScaie of the MY 2006 HHD emission rates in MOVES3. Note that the fscaie for model year
2010 and later in MOVES3 is 10 for HHD, 7 for MHD, and 5 for LHD45 and LHD2b3.Error!
Bookmark not defined.
b For operating modes lacking data, we used the same Rm use for the closest operating mode.
Equation 5-5 is used to estimate the percentage reduction to NOx running emissions from the
change in the duty-cycle standard for each operating mode.L
L We assumed that the RIM lse values calculated by MOVES operating mode can be applied to the MOVES rates that
are derived using a different fscaie. The change in fscaie does not change the definition of operating modes that are not
defined by Scaled Tractive Power, STP (deceleration operating mode 0, and idle operating mode 1), or operating
modes with negative STP values (operating mode 11 and 21), defined in Table 5-3. Changing the fscaie values does
change the definition of vehicle operation in the other operating modes. However, the R,„ ,.s,. values are relatively
constant for the positive power operating modes within each speed range as observed in Table 5-6. In Figure 5-1, we
deemed it was not necessary to attempt to account for the fscaie differences when applying the Rm use values.
231
-------
Equation 5-5
^duty_in_use — (l — ^duty) ^ ^in_use
Where:
Rduty_inuse= the percent emission reductions in the in-use running NOx emissions estimated from changing the FTP
duty-cycle standard.
Equation 5-6 is used to estimate the heavy heavy-duty diesel NOx running emission rates
from the changes in the duty-cycle standard. The same calculations were applied to estimate the
other heavy-duty diesel regulatory classes.M
Equation 5-6
ERduty_in_use — (l — ^duty_in_use) * ER]y[ovES_baseline
Where:
ER duty m use= the MOVES running NOx emission rates for the control scenarios based on reduction in the duty-
cycle standard
Rduty_inuse= the percent emission reductions in running NOx emissions estimated from changing to FTP duty-cycle
standard
ER moves _baseiine= the MOVES baseline running NOx emission rates for each regulatory class
The estimated HHD MOVES running NOx emission rates for the control scenarios, estimated
based on the duty-cycle standards in Table 5-4, are shown in Figure 5-2.
M We applied the Rduty m use developed on HHD data to both the LHD45 and MHD regulatory classes, which have
different fscaie values in MOVE3 for the 2010 and later model years. We believe this approximation is defensible for
the same reason provided in footnote L.
232
-------
300
250
-C
CD
-t—i
03
a:
¦2 200
U)
E
LU
6 150
z
CO
LU
>100
50
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
MOVES Operating Mode
Figure 5-2: Duty-cycle-based running NOx emissions, ERduty_in_use, for HHD diesel for the control scenarios
5.2.2.1.1.2 Emission Rates Based on Off-cycle Standards
In this section, we document the methods used to estimate MOVES NOx running emission
rates based on the final and proposed Option 2 off-cycle standards for heavy-duty diesel
vehicles. Table 5-7 presents the calculated off-cycle standards used to develop MOVES inputs
for the control scenarios. The final and proposed Option 2 off-cycle standards all include
requirements for operating conditions in three bins: idle, low-load and medium-to-high load.N
The off-cycle operation in the control scenarios is defined as follows: idle is less than 6 percent
of maximum power; low-load is 6 to 20 percent of maximum power; and medium-to-high load is
above 20 percent of maximum power.0 In developing inputs to MOVES, we did not apply a
scaling factor to the off-cycle idling operation and assumed manufacturers will comply with the
voluntary idle standard in all off-cycle idle operation. We then developed the off-cycle standards
for the control scenarios using the procedures as described below.
N At the time of analyzing the final standards, we used the 3-bin approach as proposed in the NPRM (while setting
the same numeric standards for the low-load and the medium-high load bins). In the final rule, the low-load and the
medium-high load bins are consolidated into a single "non-idle operation" bin, see preamble Section III.C.
u See preamble Section III.C for more discussion on defining off-cycle operations in the final program.
233
-------
Table 5-7. Calculated Off-Cycle NOx Standards used for the Control Scenarios
Reference
Off-cycle
NOx
Standard
Off-Cycle NOx
Standards
Scenario
Model
Year
Regulatory
Class
Engine
Cycle
Off-cycle Bin
(g/hr for idling,
g/hp-hr for
low-load and
medium to
high-load)
Idle (g/hr)A
5
Idle, < 6%
power
5
LHD,
MHD
LLC (g/hp-
hr)
0.05
Low-load, 6-
20% power
0.058
Final
2027+
FTP &
SET (g/hp-
hr)
0.035
Medium to
High Load,
>20% power
0.058
Standards
Idle (g/hr)A
5
Idle, < 6%
power
5
HHDB
LLC (g/hp-
hr)
0.075
Low-load, 6-
20% power
0.088
FTP &
SET (g/hp-
hr)
0.05
Medium to
High Load,
>20% power
0.088
Proposed
Option 2
LHD,
Idle (g/hr)A
10
Idle, < 6%
power
10
2027+
MHD,
HHD
LLC (g/hp-
hr)
0.1
Low-load, 6-
20% power
0.15
FTP &
SET (g/hp-
hr)
0.05
Medium to
High Load,
>20% power
0.075
A Note that the voluntary idle standard in the final control scenario that we modeled is different than the voluntary
idle standard in the final program, see preamble Sections III.B and III.C for details on the voluntary idle standard in
the final program and the off-cycle standard for idle emissions, respectively.
B Note that, in both control case scenarios, we modeled all engine categories complying with off-cycle standards
during in-use operations. For the final control scenario, the off-cycle standards for HHD include an in-use
compliance margin of 30 mg/hp-hr. As discussed in preamble Section III.C, the compliance margin for MHDE and
HHDE in the final rule is 15 mg/hp-hr for both off-cycle and duty-cycle emissions.
We calculated the voluntary idle NOx g/hr standard in units of NOx g/C02 kg using Equation
5-7, and the resulting values are displayed in Table 5-8.
234
-------
Equation 5-7
Voluntary Idle standard = [voluntary Idle NOx standard
Idle CO
(if)]
Where Idle C02 (j^j= the MOVES average CO2 (kg/hr) emission rate for HHD diesel vehicles for MOVES idle
(operating mode 1 defined in Table 5-3).
Table 5-8: Calculation of Voluntary Idle NOx/CCh Standard (g/kg)
Scenario
Applicable
Model
Years
Voluntary Idle NOx standard (g/hr)
Average HHD Idle
(Operating Mode=l) CO2
emission rate (kg/hr)
Voluntary Idle
NOx/COi
standard (g/kg)
Final
Standards
2027+
5
7.68
0.65
Proposed
Option 2
2027+
10
7.68
1.30
A Note that the voluntary idle standard in the final control scenario that we modeled is different than the voluntary
idle standard in the final program, see preamble Sections III.B and III.C for details on the voluntary idle standard in
the final program and the off-cycle standard for idle emissions, respectively.
Next, we converted the reference off-cycle NOx standards into units of gram per hour (g/hr)
for each MOVES operating mode. We refer to g/hr rates as the off-cycle standard compliant
emission rates, which are shown in Table 5-9 in Columns (F) for HHD vehicles for the final
standards.
In Table 5-9, Column (B) contains the MOVES CO2 emission rate for Model Year 2027 HHD
diesel vehicles for each MOVES operating mode. Column (C) includes the mean power for each
operating mode bin calculated from the Heavy-Duty In-Use Testing data, which is the same data
set that was used to derive the MOVES CO2 emission rates.10 The percent load, Column (D), is
calculated for each operating mode bin using Equation 5-8
Equation 5-8
Mean Power 0pMode=i
Percent Load0pMode=i = 71 „
Mean PoweropMode=40
Where i= one of the 23 MOVES running exhaust operating modes from 0 to 40
Mean Power 0pMode= t= the mean power for each of the MOVES operating modes shown in Column C.
Mean Power0pMode=40= assumed maximum power bin = 470 hp for HHD diesel vehicles.
We then assigned each MOVES operating mode into a power classification (Column (E))
based on the percent load (Column (D)), where percent load less than 6 percent of maximum
power is idle, 6 to 20 percent of maximum power is low-load, and above 20 percent of maximum
power is medium-to-high load.
235
-------
Table 5-9: Calculation of the Off-cycle NOx Standard Compliant Emission Rate for HHD Diesel Vehicles for
the Final Control Scenario
A
B
C
D
E
F
MOVES
operating
mode
MOVES
MY 2027
HHD C02
emission
rate
(kg/hr)
Mean
power
(hp)
Percent load
Power classification
MY 2027+
off-cycle
compliant
emission
rate (g/hr)
0
4.92
6.04
1.3%
Idle
3.20
1
7.68
8.10
1.7%
Idle
5.00
11
13.42
1.04
0.2%
Idle
8.73
12
21.69
28.90
6.2%
Low Load
2.54
13
37.31
75.16
16.1%
Low Load
6.61
14
52.20
121.18
26.0%
Medium to High Load
10.66
15
68.68
166.98
35.8%
Medium to High Load
14.69
16
110.42
282.24
60.5%
Medium to High Load
24.84
21
13.92
-1.61
-0.3%
Idle
9.06
22
32.99
34.43
7.4%
Low Load
3.03
23
44.71
77.71
16.7%
Low Load
6.84
24
59.82
121.62
26.1%
Medium to High Load
10.70
25
77.03
167.82
36.0%
Medium to High Load
14.77
27
102.53
230.56
49.4%
Medium to High Load
20.29
28
142.09
327.41
70.2%
Medium to High Load
28.81
29
181.90
403.76
86.5%
Medium to High Load
35.53
30
212.63
470.01
100.7%
Medium to High Load
41.36
33
28.36
34.76
7.4%
Low Load
3.06
35
71.87
145.03
31.1%
Medium to High Load
12.76
37
106.93
227.23
48.7%
Medium to High Load
20.00
38
148.35
323.23
69.3%
Medium to High Load
28.44
39
183.17
396.00
84.9%
Medium to High Load
34.85
40
196.42
466.62
100.0%
Medium to High Load
41.06
The off-cycle NOx standard compliant emission rate in Column (F) is then calculated based
on the power classification and the stringency of the off-cycle standard. For operating modes
classified as idle, we multiplied the MOVES CO2 emission rate in Column (B) with the
N0x/C02 off-cycle idle standard calculated in Table 5-8 for the corresponding control scenario
using Equation 5-9.
236
-------
Equation 5-9
Idle Emission Rate
= MOVES MY 2027 HHD C02 Emission Rate
NO
x
x Voluntary Idle in_use standard
LUo
(D
(-)
For the operating modes classified as low-load and medium- to high-load, we multiplied the
off-cycle (g/hp-hr) standard in Table 5-7 for the corresponding control scenario and power
classification by the mean power (Column C), as shown in Equation 5-10.
Equation 5-10
Low Load and Medium to High Load Emission Rate
= Mean Power (hp) x In_use standard (-—)
\hp ¦ hr/
The estimated off-cycle NOx standard compliant emission rates for heavy heavy-duty diesel
vehicles are shown in Figure 5-3. Similarly, we applied Equation 5-8 through Equation 5-10 to
estimate the off-cycle standard compliant emission rates for the other MOVES regulatory classes
using corresponding CO2 rates and the mean power for those vehicles.
Base Rates
Final Standards Off-Cycle Standard Compliant Rate
Proposed Option 2 Off-Cycle Standard Compliant Rate
^ 300
CD
-I—'
05
C£
| 250
1 200
X
O
z 150
CD
LU
>
2 100
50
nil.
1 11.
i.i.ni
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
MOVES Operating Mode
Figure 5-3: Base NOx rates and off-cycle NOx standard compliant emission rates for HHD diesel
237
-------
5.2.2.1.1.3 Emission Rates Based on Combination of Duty-Cycle and Off-Cycle
Standards
In this section, we document the methods used to develop MOVES NOx emission rates for
heavy-duty diesel vehicles that reflect the effects of both duty-cycle standards and off-cycle
standards. As an example, Figure 5-4 shows that the final HHD duty-cycle and off-cycle
standards for MYs 2027 and later affect running emission rates differently across MOVES
operating modes. The duty-cycle standard is estimated to have a larger impact than the off-cycle
standard in five operating modes (operating modes 0, 33, 35, 37, 38, and 40), while the off-cycle
standard is estimated to have a larger impact in the remaining operating modes.
MOVES Operating Mode
Figure 5-4: Comparison of Running NOx emission rates for diesel-fueled HHD compliant with the MY2027+
final duty-cycle and off-cycle standards
Because manufacturers will need to comply with both the duty-cycle and off-cycle standards,
we estimated the final MOVES NOx emission rate for each operating mode as the lower of the
two rates generated from the duty-cycle and the off-cycle standards (e.g., the emission rate based
on the final off-cycle standards is selected for operating mode 12, but the emission rate based on
the final duty-cycle standards is selected for operating mode 35). Figure 5-5 presents the
estimated emission rates for HHD diesel vehicles that meet both the final duty-cycle and off-
cycle standards. The same approach was used to estimate the emission rates for proposed Option
2 scenario. The final standards emission rates for MHD, LHD45 regulatory classes are shown in
Appendix 5.5.1.
238
-------
300
250
¦2 200
CO
in
£
LU
a150
-Z.
tt)
LU
>100
50
i Base Rate
Final Standards
Proposed Option 2
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
MOVES Operating Mode
Figure 5-5: Estimated zero-mile NOx emission rates for HHD diesel vehicles due to the final and proposed
Option 2 duty-cycle and off-cycle standards
5.2.2.1.2 Emission Rates Based on Final Changes in Warranty and Useful Life
The MOVES NOx, THC, CO, and PM2.5 emission rates for heavy-duty diesel engines in the
control scenarios are adjusted to reflect the useful life and warranty periods for the final
standards and proposed Option 2 scenario shown in Table 5-10. The emission rate adjustments
due to updated useful life and warranty periods are collectively considered adjustments to
account for "age effects."
239
-------
Table 5-10: Useful Life and Warranty Periods for Heavy-duty DieselA Engines and Aftertreatment Systems in
the Control Scenarios
Scenario
Applicable Model
Years
Warranty
Useful Life
LHD
MHD
HHD
LHD
MHD
HHD
Baseline
Model Year 2010+
5yr/50kB
5yr/ 100k
5yr/ 100k
lOyr/
110k
lOyr/
185k
lOyr/
435k
Final
Standards
Model Year 2027+
lOyr/
210k
lOyr/
280k
7yrc /
450k
15yr/
270k
12yr/
350k
llyr/
650k
Proposed
Option 2
Model Year 2027+
5yr/ 110k
5yr/ 150k
5yr/ 350k
lOyr/
250k
lOyr/
325k
lOyr/
650k
A The age effects for heavy-duty gasoline vehicles in MOVES are estimated directly from emissions data or
adapted from light-duty gasoline or heavy-duty diesel vehicles and are not tied to warranty and useful life
periods; thus, the heavy-duty gasoline or NG engine emission rates were not adjusted to account for the final
and proposed Option 2 warranty and useful life periods.
B A warranty mileage of 100k instead of 50k was assumed in the MOVES baseline emission rates for LHD
diesel, and thus, we underestimated the emissions impact of the longer warranty periods for LHD diesel vehicles
in the final standards and proposed Option 2 scenarios.
c We analyzed a warranty years value of 7 years instead of 10 years in the final standards scenario for HHD
diesel, and thus, underestimated the emissions impact of the longer warranty periods for Urban Buses in the
final standards scenario.
We used the existing methodologyp in MOVES to estimate the impact of lengthened useful
life and warranty periods on heavy-duty diesel engine emissions for each of the two control
scenarios (final and proposed Option 2 standards). In that approach, new vehicles/engines have
zero-mile emission rates for each operating mode and maintain that rate until the age of the
vehicle/engine matches the warranty period (Figure 5-6). Once the warranty period ends, the
emission rate increases linearly until the vehicle/engine reaches its useful life age. At the end of
the useful life, the emissions rates remain constant at a level calculated from the tampering and
mal-maintenance (T&M) adjustment factor. In MOVES, we assume that tampering and mal-
maintenance effects are the dominant source of emissions deterioration of fleet-wide heavy-duty
diesel emissions. Although MOVES does not explicitly account for normal deterioration of
heavy-duty diesel emissions, such as due to catalyst aging, tampering and mal-maintenance
effects assume emission increases due to aging and deterioration.
p The existing methodology is documented in Appendix B "Tampering and Mal-maintenance" of the reference
"Exhaust Emission Rates for Heavy-Duty Onroad Vehicles in MOVES3 "Error! liook,m"'k not dtli,Kd-
240
-------
Emission rate
Final emission rate
due to T&M
Zero-rnile
emission rate
End of warranty End of useful life A"e
period
Figure 5-6: Methodology to model the effects of tampering and mal-maintenance (T&M) on emission rates
according to warranty and useful life
For the baseline and control scenarios, we estimated the vehicle age at which heavy-duty
diesel vehicles would reach the end of their warranty period and the end of their useful life
period (Table 5-12). Table 5-11 shows example calculations for the final standards. Row (A)
shows the age limit of the standards for warranty and useful life periods. Row (B) shows the
mileage limit of the standards. Row (C) shows the typical miles driven per yearQ, which is used
to calculate Row (D), the calculated age rounded to the nearest whole number when the mileage
limit is reached. Row (E) is the smaller of the age at which the vehicle meets the end of its age
limit, Row (A), or mileage limit, Row (D).
Q The typical miles per year used in Table 5-11 are the same values used to derive the vehicle age at the end of the
warranty period and useful life in MOVES3.
241
-------
Table 5-11: Estimated Vehicle Age at the End of the Warranty Period and the Useful Life for Each Heavy-
duty Diesel Regulatory Class for the Final Control Scenario
Row
Warranty
Useful Life
LHD
MHD
HHD
Bus
LHD
MHD
HHD
Bus
(A)
Age limit
10
10
7A
7A
15
12
11
11
(B)
Mileage
limit
210,000
280,000
450,000
450,000
270,000
350,000
650,000
650,000
(C)
Typical
miles/year
driven
26,000
41,000
105,000
44,000
26,000
41,000
105,000
44,000
(D)
Calculated
age when the
mileage
limit is
reached
8
7
4
10
10
9
6
15
(E)
Estimated
age
8
7
4
7
10
9
6
11
A We analyzed a warranty years value of 7 years instead of 10 years in the final standards scenario for HHD diesel and
Urban Bus, and thus, underestimated the emissions impact of the longer warranty periods for Urban Buses in the final
standards scenario.
Similar calculations were performed for other regulatory classes for the baseline and control
scenarios using the same estimated mileage per year; the resulting estimates of vehicle age at the
end of the warranty and useful life periods are shown in Table 5-12.
Table 5-12: Estimated Vehicle Age at the End of the Warranty Period and the Useful Life for Each Heavy-
duty Diesel Regulatory Class in the Baseline and Control Scenarios
Warranty
Useful Life
Vehicle age
LHD
MHD
HHD
Bus
LHD
MHD
HHD
Bus
Baseline
4
2
1
2
4
5
4
10
Final Standards
Model Year 2027+
8
7
4
7
10
9
6
11
Proposed Option 2
Model Year 2027+
4
4
3
5
10
8
6
10
The T&M adjustment factor is calculated as the sum of the product of the T&M frequency for
each failure z, and the corresponding T&M emission effect, as shown in Table 5-13.
Equation 5-11
^T&M,p = I, (T&M frequency; x T&M emission effectp j)
Where:
fx&M= the tampering and mal-maintenance adjustment factor for pollutant p
T&M frequency; = estimated fleet average frequency of a tampering & mal-maintenance failure i.
T&M emission effect^ estimated emission effect for pollutant p associated with tampering & mal-maintenance
failure i.
242
-------
The emission rate at the end of useful life is then calculated using Equation 5-12.
Equation 5-12
E^End of useful life,p,r,o — ERzero mile,p,r,o ^ (l
Where:
ERUsefuiiife,P,r,o = the heavy-duty diesel emission rate at the end of warranty for each pollutant p, regulatory class, r,
and operating mode, o
ERZero miie~ the zero-mile heavy-duty diesel emission rate for each pollutant p, regulatory class, r, and operating
mode, o
fx&M= the tampering and mal-maintenance adjustment factor for each pollutant p (Equation 5-11)
We used both the T&M frequency values and T&M emission effects for THC, CO, and PM2.5
in MOVES3 for the baseline and control scenarios.Error! Bookmarknot defined.
For the NOx T&M emissions effects in the baseline scenario, we used the existing MOVES3
emission effects shown in Table 5-13, but for the control scenarios, we adjusted the emission
effects to reflect the final numeric standards. As NOx emissions become more tightly controlled
with the application of advanced technologies to meet the final standards, we anticipate the NOx
T&M emission effects will increase (i.e., there will be a relatively larger impact of T&M because
the emission control system is reducing a greater percentage of the NOx produced by the
engine). To estimate the NOx T&M emission effects for the control scenarios, we first calculated
the average zero-mile NOx emission rate ERzeromiieNOx based on the weighted average of the
different operating modes o, and regulatory class r, using Equation 5-13.
Equation 5-13
-prpf _ 2r,o(.ERzero mile,NOx,r,o ^ ^r,o )
k^zero mile,NOX — y 7
2_,r,o ^r,o
Where:
ERzero mile,nox = the average heavy-duty diesel NOx emission rate
ERZero miie,Nox,r,o = the zero-mile heavy-duty diesel NOx emission rate for regulatory class, r, and operating mode, 0
tr,0= operation time by regulatory class and operating mode estimated by MOVES3 for calendar year 2045
Next, we estimated the NOx emission rate of vehicles with a tampering and mal-maintenance
failure i, using Equation 5-14, which was derived from Equation 5-12 using the fleet average
emission rate from Equation 5-13 assuming the T&M frequency is 100 percent.
Equation 5-14
ERT&Mi,«ox = ERzero mile,nox x (l + T&M emission effectj WOx)
We then derived Equation 5-15, assuming that a NOx aftertreatment equipment failure i, in
the control scenario, would cause the average of the failed emission rates, ERt8lM £jNOx , to be the
same as a NOx aftertreatment failure in the baseline case, Baseline ERX&M [ NOx .
243
-------
Equation 5-15
Baseline ERT&M lNOx = Control ERT&M lNOx
Baseline ERzero miie,wox x (l + Baseline T&M emission effect; WOx)
= Control ERzero mile,wox x (l + Control T&M emission effect; NOx)
By rearranging Equation 5-15, we derived Equation 5-16 to estimate the control scenario NOx
T&M emissions effects.
Equation 5-16
Control T&M emission effect; NOx
"Baseline ERzero mile,wox x (l + Baseline T&M emission effect; WOx)l
Control ERzero mjie WOx
Table 5-13 presents the T&M NOx emission effects for the NOx aftertreatment failures for
the control scenarios calculated from Equation 5-16. The T&M NOx emission effects for the
NOx aftertreatment failures are much larger than the baseline scenario, because the zero-mile
NOx emission rate in the control scenarios are lower than the baseline zero-mile NOx emission
rates. As shown in Table 5-13, the NOx T&M emission effects for the other T&M failures (e.g.,
Timing Advanced and EGR Disabled/Low-Flow) in the control scenarios use the same NOx
T&M emissions effects as the baseline.
244
-------
Table 5-13: NOx Tampering & Mal-maintenance (T&M) Emission Effects for HHD
Baseline
Final
Standards
MY 2027+
Proposed
Option 2
MY 2027+
Timing Advanced
6%
6%
6%
Timing Retarded
-20%
-20%
-20%
Injector Problem (all)
-1%
-1%
-1%
Puff Limiter Mis-set
0%
0%
0%
Puff Limited Disabled
0%
0%
0%
Max Fuel High
0%
0%
0%
Clogged Air Filter
0%
0%
0%
Wrong/W orn Turbo
0%
0%
0%
Intercooler Clogged
3%
3%
3%
Other Air Problem
0%
0%
0%
Engine Mechanical Failure
-10%
-10%
-10%
Excessive Oil Consumption
0%
0%
0%
Electronics Failed
0%
0%
0%
Electronics Tampered
8%
8%
8%
EGR Stuck Open
-20%
-20%
-20%
EGR Disabled/Low-Flow
5%
5%
5%
NOx Aftertreatment Sensor1
200%
1505%
1407%
Replacement NOx Aftertreatment Sensor A
200%
1505%
1407%
NOx Aftertreatment Malfunction A
500%
3111%
2914%
PM Filter Leak
0%
0%
0%
PM Filter Disabled
0%
0%
0%
Oxidation Catalyst Malfunction/Remove
0%
0%
0%
A For detailed descriptions of these failure modes, refer to Appendix B in Exhaust Emission Rates for Heavy-Duty
On-road Vehicles in MOVES3.Error! Bookmark not defined.
Using the NOx T&M emission effects in Table 5-13, we then calculated T&M adjustment
factors fx&M.NOX f°r each scenario using Equation 5-11 and the baseline T&M frequency values.
For THC, CO, and PM2.5, we used the existing T&M adjustment factors fx&M,p in MOVES3.
Then, we calculated the heavy-duty diesel emission rate for each pollutant p, age a, regulatory
class r, and operating mode o, using Equation 5-17.
Equation 5-17
E^p,r,a,o — ERzero mjie p r o X (1 + Sa X fx&ivi)
Where:
ERPir,0,a = the heavy-duty diesel emission rate for each pollutant p, regulatory class r, age a, operating mode, 0,
ERZero miie~ the zero-mile heavy-duty diesel emission rate for each pollutant p, regulatory class r, operating mode, 0
sa= scaled age effect at age a
fx&M= the tampering and mal-maintenance adjustment factor (Equation 5-11)
245
-------
The scaled age effect, sa, is calculated using the age of the vehicle in comparison to the
warranty and useful life requirements, as shown in Table 5-14. When the vehicle age is between
the end of the warranty and the useful life, sa is interpolated between 0 and 1.
Table 5-14: Calculation of sa
Where:
0
age < end of warranty age
(age — end of warranty age)
(Useful life age — end of warranty age)
end of warranty age < age < useful life
1
age > useful life
As the final step, the age-adjusted emission rates calculated in Equation 5-17 were averaged
according to the age ranges shown in Table 5-15 that are used to define emission rates in
MOVES for LHD45, MHD, HHD, and Urban Bus regulatory classes. The resulting age-adjusted
running emissions have a relationship with vehicle age as shown in Figure 5-7 and Figure 5-8 for
HHD NOx emissions.11
Table 5-15: MOVES ageGroupID Which Are Used to Define Running and Start Emission Rates
ageGroupID
Lower bound
(years)
Upper bound
(years)
3
0
3
405
4
5
607
6
7
809
8
9
1014
10
14
1519
15
19
2099
20
30
5.2.2.1,3 Summary of Diesel NOx Running Emission Rates
Figure 5-7 shows average running NOx emission rates (g/mile) in MOVES3 for the model
year 2027 fleet across vehicle age for the baseline and control scenarios. The MOVES running
emission rates for the control scenarios reflect the adjustments to the duty-cycle and off-cycle
standards, and the extended warranty and useful life as discussed above in Chapter 5.2.2.1.1 and
Chapter 5.2.2.1.2, respectively. The gram per mile average running emissions are also a function
of the default activity assumptions in MOVES3 6
Figure 5-7 shows that the average zero-mile NOx emission rates for the control scenarios are
significantly lower than the baseline scenario. The figure also demonstrates the larger T&M NOx
R The average emission rate accounts for the frequency of different operating modes according to MOVES estimate
of in-use vehicle activity. The trend in individual operating modes will be slightly different than the average trend
shown in Figure 5 7. For example, the zero-mile idle operating mode is not reduced as much as the average emission
rates in the control scenarios. Because the control case T&M emission effects were calculated using the average
emission rates in Equation 5 16, individual emission rates in the control case, such as for idle, can be higher than the
baseline scenario when fully aged. This is a feature of the method used to derive the aging effects, but the effect is
averaged out when conducting emission inventory analysis.
246
-------
emission effect for the control scenarios than for the baseline scenario as explained in Chapter
5.2.2.1.2. Although not shown, the emission rate is constant from age 15 through age 30.
Baseline
n c
Final Standards
Proposed Option 2
T? 1
/
_OJ z
E
a] 15
/
Q_
to
E
2 l
-5S
X
5 OS
J
A
7
^ U.j
0
c
) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Age
Figure 5-7: NOx emission rates (g/mile) in MOVES for HHD diesel long-haul combination trucks for the
model year 2027 fleet across vehicle age for the baseline and control scenarios
5.2.2.2 Heavy-Duty Diesel Start Emission Rates
In this section, we describe the methods used in MOVES3 to estimate lower start NOx
emission rates for the control scenarios due to the final and proposed Option 2 duty-cycle
standards. We did not estimate the impact of the off-cycle standard on start emissions, in part
because the baseline MY 2010 and later start emission rates in MOVES3 are not based on in-use
data but are based on emissions data from the FTP duty-cycle.Error! Bookmarknotdeflned-
Additionally, because the baseline heavy-duty diesel start emission rates in MOVES3 do not
vary with age due to insufficient dataError! Bookmarknotdeflned-5 we did not estimate changes due to
the changes in warranty and useful life in the control scenarios.
Start emission rates in MOVES are defined by regulatory class, fuel type, vehicle age and
operating mode. Start operating modes in MOVES are also defined by different lengths of engine
soak time (the time between the preceding engine off event and the engine start). The length of
soak time for each MOVES operating mode bin is defined in Figure 5-9. Operating mode 108
represents a start with a soak longer than 720 minutes or 12 hours and is referred to as a 12-hour
cold-start.
To estimate the start emissions under the control scenarios, we estimated the NOx cold start
emission rate (g/start) from the CARB Stage 1 HDD engine tested on the FTP duty-cycle cycle.
Table 5-16 contains the NOx Cold and Hot FTP measurements in Columns (B) and (C) for
247
-------
different aging lengths. Cold - Hot, Column (E), is calculated as the difference between the
Columns (B) and (C). The cold start, Column (F), is then calculated by multiplying the
difference in Column (E) by the work performed on the FTP cycle, Column (D), as shown in
Equation 5-18.
Equation 5-18
NOx Cold Start (——) = [cold (—-—) — Hot (—-—) x FTP work (hp ¦ hr)
x Vstart/ L Vhp ¦ hr/ Vhp ¦ hr/J v p 7
Table 5-16: Calculation of NOx 12-hour Cold Starts from the CARB Stage 1 HHD Engine from the Cold and
Hot FTP Cycle
(A)
(B)
(C)
(E)
(D)
(F)
Aged hours
FTP composite
(g/hp-hr)
Cold
(g/hp-hr)
Hot
(g/hp-hr)
Cold - Hot
(g/hp-hr)
FTP Work
(hp-hr)
Cold Start
(g/start)
0
0.008
0.025
0.005
0.02
31.4
0.63
333
0.012
0.042
0.006
0.036
31.4
1.13
656
0.018
0.061
0.009
0.052
31.4
1.64
1000
0.024
0.092
0.01
0.082
31.4
2.58
1000 hrPost Ash
Clean
0.026
0.109
0.009
0.1
31.4
3.14
The Stage 1 HHD engine is deemed representative of an engine-certified to a 0.02 g/hp-hr
NOx standard based on the FTP composite measurements in Column (A). Table 5-16
demonstrates that there was larger cold start measured with increase in aged hours, and after the
ash clean out at 1000 hours. We used the 1000 hr, Post Ash Clean cold start emission rate (3.14
g/start shown in Table 5-16) to represent the 12-hour cold-start (operating mode 108) emission
rate.
To estimate the 12-hour cold-start NOx emission rate for HHD diesel vehicles in the control
scenarios, we interpolated the HHD 12-hour cold-start between the Stage 1 cold start (3.14
g/start) and the MOVES baseline 12-hour cold-start (8.4 g/start), and their respective FTP duty-
cycle standards using Equation 5-19 as shown in Figure 5-8 and Table 5-17. For example, the
interpolation yielded an estimated 12-hour cold start of 4.02 g/start for the 0.05 g/hp-hr FTP
standard.
248
-------
Equation 5-19
Start ERFXPx HHD12 hour
= /MOVES start Hhd,i2 hour ~ Stagel start\ _
V Baseline FTP - Stagel FTP / x
+ MOVES start HHd,i2 hour
Where:
Start ERDuty_cycle standard x,hhd,i2 hour = the estimated NOx start emissions for an FTP duty-cycle standard, x, for
heavy heavy-duty diesel emissions for a 12-hour cold-start (operating mode 108).
Stagel start = 1000 Post Ash Clean start emission rate from the CARB Stage 1 HHD diesel engine =3.14 g/start
(Table 5-16)
Stagel FTP = Composite FTP level of the CARB Stage 1 engine = 0.02 g/hp-hr
MOVES start HHD12 hour= MOVES3 baseline start emission rate (= 8.4 g/start) for MY 2027 heavy heavy-duty
diesel engine for a 12-hour soak (operating mode 108)
Baseline FTP = baseline FTP composite NOx standard = 0.2 g/hp-hr
FTPX = composite FTP standard in the control scenarios, either 0.035, 0.029, 0.05, or 0.02 g/hp-hr (Table 5-4)
9
_ 8
-M
!_
«j 7
IS)
*9 6
4->
5 5
l/l
T3
o 4
(_>
^ 3
o
T 2
rsi
T—1
Q 1
X
xo
c
) 0.05 0.1 0.15 0.2 0.25
FTP Composite Standard (g/hp-hr)
Figure 5-8: Estimated relationship between the HHD NOx 12-hour cold-start and the composite FTP NOx
standards
249
-------
Table 5-17: HHD Cold Start Emissions for Baseline and Control Scenarios
Scenario
Applicable Model
Years
Weighted
Average FTP
standard (g/hp-
hr)
Cold Start
emissions (g/start)
Baseline
Model Year 2010+
0.2
8.40
Final Standards
Model Year 2027+
0.05
4.02
Proposed Option 2
Model Year 2027+
0.05
4.02
We then used Equation 5-20 to estimate the MOVES NOx emission rates for each MOVES
heavy-duty regulatory class (LHD45, MHD, and HHD) and for each MOVES start operating
mode classified by different soak times (Figure 5-9). We assumed that the relative difference in
emission rates by regulatory class and by operating mode is the same in the baseline and control
scenarios.
Equation 5-20
Start ERpTp=x,reg class=y,soak=z
_ r T7r) /'MOVES start reg ciass=y,soak=zN\
- start hKDutycyclestandardx,hhd,12 X ^ MOVES start hhd,12-hour )
Where:
Start ERfxp= the start NOx emission rates for the control scenarios with FTP x (0.035, 0.029,0.05, or 0.02) for
regulatory class y (LHD45, MHD, and HHD), and soak length z
Start ERDuty cycle standard x,hhd,i2—hour = the estimated start emissions for an FTP duty-cycle standard, x, for heavy
heavy-duty diesel emissions for a 12-hour soak (operating mode 108)
MOVES start reg ciass=y,soak=z = MOVES3 baseline start emission rate for MY 2027 for regulatory class y (LHD45,
MHD, and HHD), and soak length z
MOVES start HHD,i2-hour= MOVES3 baseline start emission rate for MY 2027 HHD diesel engine for a 12-hour
soak (operating mode 108)
Figure 5-9 shows the estimated MOVES NOx start emission rates for HHD diesel vehicles for
the baseline scenario, as well as the final standards and proposed Option 2 scenarios.
250
-------
¦ Base Rate ¦ Final Standards » Proposed Option 2
9
< 6 min 6-30 min 30-60 min 60-90 min 90-120 min 120-360 360-720 720+min
min min
MOVES Start Operating Mode
Figure 5-9: Duty-cycle-based NOx start emissions for HHD Diesel for the baseline, final standards and
proposed Option 2 scenarios
5.2.2.3 Heavy-Duty Extended Idle Emission Rates
In MOVES, extended idling is defined as idling for more than an hour, which occurs during
hotelling activity when long-haul combination trucks idle during rest periods. MOVES has
extended idle emission rates for long-haul combination trucks that include HHD, MHDS and
glider vehicles1. All idling activity by other regulatory classes is modeled using the running idle
emission rates (Table 5-3), which are different than extended idle emission rates. We anticipate
that reductions in the HHD and MHD NOx extended idle emissions rates will be driven by the
idle standard, rather than the duty-cycle standards in the final rule. The duty-cycle standards do
not contain high duration extended idling (> 1 hour) that is representative of truck hotelling
activity. In addition, we did not estimate lower extended idle emission rates due to the
lengthened warranty or useful life periods.u
First, we estimated extended idle emission rates that would comply with the off-cycle
N0x/C02 g/kg standard calculated in Table 5-8. We then used Equation 5-9 to calculate the
extended idle off-cycle NOx g/hr emission rate based on the MOVES extended idle CO2 g/hr
emission rate, as shown in Table 5-18.
s HHD and MHD have the same extended idle emission rates in MOVES.
T We assumed there are no changes to glider emission rates due to the final rule.
u Extended idle emission rates in MOVES are not differentiated by vehicle age.
251
-------
Table 5-18: Calculation of HHD and MHD Extended Idle NOx g/hr Emission Rates
Control Scenario
Model Years
2027 MY
Baseline
Rates
NOx
(g/hr)
2027 MY
Baseline
Rates
CO2
(g/hr)
Idle
Standard
(g/hr)A
Idle
Standard
NOx/CCh
(g/kg)
Idle-
standard
compliant
NOx
emission
rate
(g/hr)
%
Change
in NOx
emission
rate
Final Standards
Model Year 2027+
42.6
7191
5
0.65
4.68
-89%
Proposed
Option 2
Model Year 2027+
42.6
7191
10
1.30
9.36
-78%
A Note that the voluntary idle standard in the final control scenario we modeled is different than the voluntary idle
standard in the final program, see preamble Section III.B for details on the voluntary idle standard in the final
program.
5.2.2.4 Heavy-duty Diesel Crankcase Emissions
Since the proposal, we made improvements in MOVES to better model the emissions from
crankcase for both open and closed systems, by including the fraction of the fleet with closed
crankcase systems in the baseline as well as incorporating more recent data from MY2027 and
later HD vehicles Error! Bookmark not deflned- With these updates, the inventory analysis done for the
final rule estimates more accurately the emissions benefits of the closed crankcase for THC, CO,
NOx, and PM2.5
As described in Section III.B of the preamble, EPA is finalizing a requirement for
manufacturers to use one of two options for controlling crankcase emissions, either: 1) as
proposed, closing the crankcase, or 2) an updated version of the current requirements for an open
crankcase that includes additional requirements for measuring and accounting for crankcase
emissions. For the emissions impact analysis of the final standards and proposed Option 2
presented in Chapter 5.3 below, the emission reductions were estimated assuming that closing
the crankcase would be the preferred option to meet the final standards.
In modeling the control scenarios, the PM2.5 crankcase emissions from HHD, MHD, and
LHD45 diesel vehicles were set to zero. For LHD2b3 diesel vehicles, we reduced the crankcase
emissions by 94.9%, assuming that 5.1% of the LHD2b3 diesel vehicles are engine-certified (see
Chapter 5.2.2.5).
252
-------
5.2.2.5 Light Heavy-Duty Class 2b and 3 Diesel Emission Rates
We assumed that in 2027 and later model years, 5.1 percent of the diesel-fueled LHD2b3
vehicles will be engine-certified and therefore, will be impacted by the final rule.v'w To develop
the emission rates for LHD2b3 vehicles, for the control scenarios, we assumed that 5.1 percent of
the emissions from LHD2b3 are equivalent to the controlled emissions of LHD45 regulatory
class vehicles. This is consistent with the analysis for model year 2010 and later diesel vehicles,
where we used the same data from the HDIUT program to estimate emission rates for both
LHD2b3 and LHD45 vehicles.Error! Bookmarknot defined. jn ^d^ion, We assumed that the final duty-
cycle, off-cycle, and warranty and useful life requirements are the same for all engine-certified
LHD vehicles.
We did not estimate the contribution of engine-certified vehicles on the emission rates for
diesel-fueled LHD2b3 in the baseline scenario.x Because the LHD2b3 diesel emission rates
certified to engine standards are higher than the emission rates certified to chassis standards, the
control scenarios generally increase NOx emissions compared to the baseline scenario for diesel
LHD2b3 vehicles for most calendar years, even though we anticipate the final rule will reduce
NOx emissions for engine-certified LHD2b3 vehicles. We included the emission contribution
from engine-certified diesel-fueled LHD2b3 vehicles in the control scenarios to better account
for future NOx emissions from these vehicles. We acknowledge that we are underestimating the
benefits of controlling these vehicles due to their absence from the baseline scenario.
5.2.2.6 Heavy-Duty Gasoline Running Emission Rates
In this section, we describe the methods used to develop the running exhaust emission rates
for NOx, THC, CO, and PM2.5 from heavy-duty gasoline vehicles in MOVES for the control
scenarios. The final rule does not include off-cycle standards for vehicles fueled by gasoline or
NG. Furthermore, even though we anticipate emission benefits from the lengthened warranty and
useful life periods from gasoline and NG-fueled vehicles, they were not included in the analysis
done for the final rule.
v Class 2b and 3 vehicles with GVWR between 8,500 and 14,000 pounds are primarily commercial pickup trucks
and vans and are sometimes referred to as "medium-duty vehicles." The majority of Class 2b and 3 vehicles are
certified as complete vehicles and which EPA intends to include in a future combined light-duty and medium-duty
rulemaking action, consistent with Executive Order 14037, Section 2a.
w In Appendix 5.5.1 of the draft RIA, we presented an analysis suggesting that 4.2% of MY 2027 diesel-fueled
LHD2b3 vehicles would be engine-certified. However, we used 5.1% in developing the MOVES rates for LHD2b3
vehicles and subsequent inventory analysis (including this final rulemaking analysis). Given the small contribution
of engine-certified LHD2b3 to the total emissions inventory, we expect this would have only a negligible impact on
the emission reductions estimated in the final rule. In addition, we deem that both estimates (4.2% and 5.1%) are
within the range of feasible values for the fraction of engine-certified LHD2b3 vehicles in future years.
x In the baseline case created for MOVES3, we assumed, for simplicity, that 100% of diesel-fueled LHD2b3
vehicles are chassis-certified and are subject to the light-duty Tier 3 emission standard. Note that the estimated NOx
emission rates for engine-certified diesel LHD2b3 vehicles (subject to the final rule) are higher than chassis-certified
diesel LHD2b3 vehicles (subject to the light-duty Tier 3 standard).
253
-------
The final FTP duty-cycle standards shown in Table 5-4 apply to both heavy-duty
compression-ignition engines and heavy-duty spark-ignition engines.Y For the control scenarios,
we updated the NOx exhaust emission rates for gasoline, assuming that emissions are reduced for
all operating modes based on the reduction in the NOx FTP standards from the current 0.2 g/hp-
hr standard. Table 5-19 shows the estimated reduction in NOx emission rates, which is consistent
with the ratio of the current FTP emission standards and the final and proposed Option 2 FTP
standards shown in Table 5-5.
In addition to modeling the final and proposed Option 2 standards for NOx, we estimated
emission rate reductions due to the final and proposed Option 2 standards for HC, CO and PM2.5.
As discussed in the preamble Section III.D, the final emissions standards for HC, CO, and PM2.5
heavy-duty spark ignition are lower than the current MY 2010 standards. We estimated reduced
THC and CO emission rates assuming that those emissions would be reduced due to
improvements in the three-way catalyst emission controls. We used initial data from our
production HD SI engines and from the heavy-duty gasoline technology demonstration program
presented in Chapter 3.2 to estimate our modeled emissions levels. We assumed a 65 percent
reduction in THC emissions would occur at a NOx standard of 0.1 g/hp-hr.z We assumed
additional decreases in THC emissions to reflect tighter final NOx standards in the control
scenarios in MY 2027. We derived Equation 5-21 assuming a linear decrease in THC emissions
between the estimated THC emissions emitted at the 0.1 g/hp-hrNOxFTP level, and zero THC
emissions at a hypothetical 0 g/hp-hr NOx FTP level. We then used Equation 5-21 to estimate
the reductions in THC emissions using the NOx levels for the control scenarios (Table 5-19).
Equation 5-21
I NOx FTP Standard \ ^ ^
^gasoline,THC,NOx FTP = 1 — [ ~ g ] X (1 — ^gasoline,THC,0.1 NOx FTP )
\ bhp ¦ hr J
. I NOx FTP Standard\ , .
= 1 - —oT^~— x d -65%)
\ " bhp hr /
Where:
Rgasoline, THC, NOx ftp - percent emission reductions in heavy-duty gasoline THC emissions for NOx FTP standards
more stringent than the 0.1 NOx FTP standard, calculated values shown in Table 5-19
NOx FTP Standard = NOx FTP standards in the control scenarios
We assumed a single CO standard for MY 2027 and later HD SI engines and we maintained a
60 percent reduction in CO for all scenarios (see Table 5-19). To meet the final PM standards,
manufacturers are expected to improve fuel control and limit the need for catalyst protection.
Therefore, we assumed a 50 percent reduction in PM2.5, consistent with the 50 percent lower PM
standard, for all scenarios. Table 5-19 contains the emission rate reductions, Rgasoiine-, applied in
MOVES for the emission inventory analysis.
Y Our inventory analysis for HD SI engines only evaluated the impact of the final FTP duty-cycle standards. We did
not analyze the impact of our final SET duty-cycle standards or idle provisions for HD SI engines.
z The scenario analyzed for the air quality modeling assumed an FTP standard at 0.1 g/hp-hr.
254
-------
Table 5-19: Running Emission Rate Reductions From Heavy-duty Gasoline Vehicles Due to Final and
Proposed Option 2 Standards, Rgasoune, Across All Heavy-duty Gasoline Regulatory Classes and Operating
Modes
Control
Scenario
Model
Years
Regulatory
ClassA
FTP/SET NOx standard
(g/hp-hr)
NOx
THC
CO
PM25
Final
Standards
2027+
LHD,
MHD,
HHD
0.035
82.5%
87.8%
60%
50%
Proposed
Option 2
2027+
LHD,
MHD,
HHD
0.035
82.5%
87.8%
60%
50%
A We applied the same standards for the final and proposed Option 2 scenarios to represent the SI engines modeled
by the LHD, MHD, and HHD regulatory classes, unlike the final standards for compression-ignition engines
(Preamble Section III.D)
We used Equation 5-22 to estimate the MOVES NOx emission rates for the control scenarios
using the Rgasoime values for heavy-duty gasoline vehicles. Since the final rule does not require an
in-use testing program for spark-ignition engines, we did not estimate operating mode-specific
effectiveness of reductions of the in-use emissions compared to duty-cycle standard emissions,
as was done for diesel running emissions. Instead, we assumed these reductions apply uniformly
across all running exhaust operating modes. As such, we used Equation 5-22 to estimate the
MOVES emission rates proportionally for all operating modes.
Equation 5-22
ERcontrol (l Rgasoline) ^ ^^MOVES_baseline
Where:
ERControi = MOVES running exhaust emission rates for the control scenarios based on the reduction in the FTP duty-
cycle standard
Rgasoiine = percent emission reductions in heavy-duty gasoline emissions from Table 5-19
ERmoves baseline = MOVES running exhaust emission rates for the baseline
The estimated heavy-duty gasoline MOVES running emission rates for the baseline, final
standards, and proposed Option 2 scenarios are shown for NOx and THC emissions in Figure
5-10 and Figure 5-11, respectively. CO and PM2.5 were similarly estimated from the reductions
shown in Table 5-19, but they have the same emission rates within each regulatory class for all
the control scenarios and, therefore, are not plotted.
255
-------
Base Rate
Final Standards
Proposed Option 2
., lI
¦ l.i I.I
1 i i I 1 1
I
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
MOVES Operating Mode
Figure 5-10: Duty-cycle-based running NOx emission rates for LHD gasoline for the control scenarios
256
-------
MOVES Operating Mode
Figure 5-11: LHD gasoline Duty-cycle-based running THC emission rates for LHD gasoline for the control
scenarios
5.2.2.7 Heavy-Duty Gasoline Refueling Emission Rates
In this section, we describe the methods used to estimate lower refueling emission rates in
MOVES for the control scenarios due to the final Onboard Refueling Vapor Recovery (ORVR)
requirements. Refueling emissions result when the pumped gasoline displaces the vapor in the
vehicle tank. The THC emissions from refueling are a function of temperature and the gasoline
Reid Vapor Pressure (RVP). ^ The emissions control technology which collects the vapor from
the refueling events is the ORVR system. ORVR requirements for light-duty vehicles started
phasing in as part of EPA's Refueling Emission Regulations for Light-Duty Vehicles and Light-
Duty Trucks Final Rule11 in 1998. Under the EPA's Tier 2 vehicle program, all complete
vehicles with a gross vehicle weight rating (GVWR) of 8,500 lbs. up to 14,000 lbs. (MOVES
regulatory class LHD2b3) were required to meet the ORVR requirements between 2004 and
2006 model years.12 With the Tier 3 rulemaking, all heavy-duty trucks up to 14,000 lbs. and all
AA See additional discussion of refueling updates in the Evaporative Emissions from Onroad Vehicles in MOVES3
14
257
-------
complete vehicles greater than 14,000 lbs. are required to meet a refueling standard of 0.2 grams
of HC per gallon of gasoline dispensed by 2022.13
Table 5-20 shows the ORVR adoption rates applied to all heavy-duty gasoline trucks in
MOVES. For the baseline scenario, we estimated that all heavy-duty gasoline trucks with
GVWRup to 14,000 lbs. will have ORVR control by 2018 (as shown in Table 5-20). No heavy-
duty gasoline vehicles over 14,000 lb GVWR are being certified todayBB as complete vehicles,
and our baseline scenario reflects a population of 100 percent incomplete vehicles that have not
adopted ORVR technologies and are not expected to adopt ORVR without a regulatory action,
due to the costs and added complexities.
As part of this final rulemaking, all heavy-duty gasoline vehicles, including those sold as
incomplete vehicles, will be required to have an ORVR system and be certified to the same
standard as light-duty by model year 2027. Therefore, for the control scenarios, we assumed
manufacturers will fully implement ORVR technologies for HD vehicles over 14,000 GVWR
starting in MY2027. For a 100 percent phase-in of ORVR, we estimated a 98 percent reduction
in refueling emissions, because we assume that some ORVR systems would fail or may not be
fully effective, similar to our assumptions made for current ORVR systems in light-duty
vehicles.14 The emissions inventory impact of the final ORVR control is summarized in Chapter
5.3.3.
Table 5-20: Phase-In of Onboard Refueling Vapor Recovery (ORVR) for Heavy-duty Trucks
Model Year
Light Heavy-Duty Trucks
8,500-10,000 lbs GVWR
(Class 2b)
Heavy-Duty Trucks
10,000-14,000 lbs GVWR
(Class 3)
Heavy-Duty Trucks
> 14,000 lbs GVWR
(LHD45 and MHD)
including incompletes
Baseline
Final
Standard
2003 and earlier
0%
0%
0%
0%
2004
40%
0%
0%
0%
2005
80%
0%
0%
0%
2006-2017
100%
0%
0%
0%
2018-2026
100%
100%
0%
0%
2027 and later
100%
100%
0%
100%
5.3 National Emissions Inventory Results
In the following sections, we present the estimated emissions impacts of the final control
scenario and the proposed Option 2 in three select calendar years.cc The national (50 states and
Washington DC, excluding Puerto Rico and the Virgin Islands) highway heavy-duty vehicle
emission inventory was generated using the national-scale option in MOVES3 with the
BB We expect that only one complete vehicle model will exist in 2022 and it is not yet certified.
cc The final rule is expected to have minimal impacts on C02 emissions. We estimated a small fuel savings for
heavy-duty gasoline vehicles in Section 7.2.2 due to ORVR control (Chapter 5.2.2.7). However, for MOVES
emissions inventories, we estimated no differences in the CO2 emission rates for the baseline and control scenarios.
See RIA Chapter 1 for more discussion of the technologies evaluated to control NOx emissions without impacting
CO2 emissions.
258
-------
methodology and the model inputs as described in Chapter 5.2.DD For comparison, we also
present the emission impacts of the Proposed Option 2 using the same methodology.
5.3.1 Final Standards
Table 5-21 summarizes the emission impacts of the final control scenario for three select
calendar years. Chapter 5.5.4 shows NOx, VOC, PM2.5, and CO inventories for all calendar years
between 2027 and 2045.
Table 5-21: National Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,2040, and 2045
— Final Control Scenario Emissions Relative to Heavy-Duty Vehicle Emissions Baseline
Pollutant
2030
2040
2045
US Short
Tons
%
Reduction
US Short
Tons
%
Reduction
US Short Tons
%
Reduction
NOx
139,677
14.0%
398,864
43.5%
453,239
47.9%
VOC
5,018
4.9%
17,139
19.6%
20,758
22.6%
Primary Exhaust PM2 5
- Total
115
0.9%
491
6.6%
566
7.7%
Carbon Monoxide (CO)
43,978
3.0%
208,935
15.5%
260,750
18.3%
Acetaldehyde
36
1.5%
124
6.0%
145
6.7%
Benzene
40
3.7%
177
23.3%
221
27.6%
Formaldehyde
29
1.1%
112
6.6%
134
7.5%
Naphthalene
2
1.0%
7
13.2%
9
15.7%
More details about the impacts of the final standards can be found in Chapter 5.5.2, where the
emission reductions are categorized by vehicle fuel type with further splits by emission process
and by heavy-duty regulatory class. Contributions to NOx emissions from different engine
operational processes in calendar year 2045 are also provided in Chapter 5.5.3 for both the final
and proposed Option 2 control scenarios.
5.3.2 Proposed Option 2
Table 5-22 summarizes the emissions impacts in three selected calendar years for proposed
Option 2.
DD Because of the differences in the control scenarios and the differences in the emission inventory methodology
between the proposal and the final rule, no direct comparison should be made between the emission impacts of the
final rule presented in Table 5-21 and the emission impacts estimated in the proposal.
259
-------
Table 5-22: National Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,2040, and 2045
— Proposed Option 2 Program Emissions Relative to Heavy-Duty Vehicle Emissions Baseline
2030
2040
2045
Pollutant
US Short
%
US Short
%
US Short
%
Tons
Reduction
Tons
Reduction
Tons
Reduction
NOx
133,699
13.4%
352,468
38.5%
400,024
42.3%
VOC
5,018
4.9%
17,067
19.5%
20,681
22.5%
Primary Exhaust
115
0.9%
443
6.0%
515
7.0%
PM2 5 - Total
Carbon Monoxide
43,978
3.0%
204,178
15.1%
255,653
17.9%
(CO)
Acetaldehyde
36
1.5%
121
5.9%
142
6.6%
Benzene
40
3.7%
177
23.3%
221
27.6%
Formaldehyde
29
1.1%
110
6.5%
132
7.4%
Naphthalene
2
1.0%
7
13.1%
9
15.6%
5.3.3 Impacts of Heavy-Duty Gasoline Refueling Controls
Table 5-23 shows the estimated impact on refueling emissions from heavy-duty vehicles due
to the final refueling emission standard. For heavy-duty vehicles, MOVES3 only estimates
refueling emissions from gasoline-fueled vehicles. Thus, the reductions reflect the control of the
refueling emissions from only the heavy-duty gasoline vehicles above 14,000 lbs. Because
benzene is calculated as a fraction of VOC emissions, the percent reductions are the same for
both pollutants as shown in Table 5-23.
Table 5-23: Emission Reductions Due to Adoption of ORVR for Heavy-Duty Vehicles Relative to Heavy-Duty
Vehicle Emissions Baseline
Pollutant
Calendar Year
Reductions in US Short Tons
% Reduction
Benzene
2027
3
6.6%
2030
13
27.8%
2040
43
80.2%
2045
52
88.7%
VOC
2027
890
6.6%
2030
3,718
27.8%
2040
11,867
80.2%
2045
14,381
88.7%
5.4 Emissions Inventories for Air Quality Modeling
When feasible, we conduct full-scale photochemical air quality modeling to accurately project
levels of criteria and air toxic pollutants, because the atmospheric chemistry related to ambient
concentrations of PM2.5, ozone, and air toxics is very complex. Air quality modeling was
performed for the proposed rule and demonstrated improvements in concentrations of air
pollutants. We did not perform new air quality modeling for this final rule. This section of the
RIA describes the emissions inventories that were used in the air quality modeling and presents
the differences between the air quality modeling emissions inventories and those developed for
the final rule. As further described in this Chapter 5.4, despite the differences in the version of
260
-------
the MOVES model used and the control scenario modeled, the emission reductions used in the
air quality modeling analysis for the proposed rule compare well with the emission reductions
estimated for the final standards.
The air quality modeling analysis required emission inventories with greater geographical and
temporal resolution than the national-scale inventories described in Chapter 5.3. The approach
for estimating emission inventories for an air quality modeling analysis is extremely complex
and time and resource intensive since it involves modeling of each 12 km grid cell, for each hour
of the day, for the entire year.
The methodology for developing the onroad emission inventories for the air quality modeling,
also referred to as SMOKE-MOVES emission inventories, is summarized here. Additional
details, including information for sectors other than onroad vehicles, are available in the Air
Quality Modeling Technical Support Document (AQM TSD).15 Figure 5-12 illustrates the
process involved in generating the onroad emissions inventories for use in air quality modeling.
First, MOVES county-level onroad emission factors (EF) by temperature and speed bins are
generated based on the output from the meteorological preprocessor, Met4Moves, which is used
to develop the temperature ranges, relative humidity, and temperature profiles. As discussed
below, the MOVES emission rates for the air quality modeling inventory differed from the
emission rates discussed in Chapter 5.2.2. Additionally, other MOVES data inputs for the
county-level runs used for the air quality modeling sometimes differ from the inputs used for the
national inventory. For example, the county-level MOVES runs incorporate county-specific
inputs including: fuel programs, inspection and maintenance programs, adoption of LEV
standards, and the age distribution of the local vehicle fleet. The emission factors for each
representative county were generated with multiple runs of the MOVES CTINPRM version of
the model.
The MOVES-generated onroad emission factors were then combined with activity data to
produce emissions within the Sparse Matrix Operator Kernel Emissions (SMOKE) modeling
system. The collection of tools that compute the onroad mobile source emissions (as one of the
sectors included in the air quality modeling) are known as SMOKE-MOVES. SMOKE-MOVES
uses a combination of vehicle activity data, emission factors from MOVES, meteorology data,
and temporal allocation information to estimate hourly onroad emissions. Additional types of
ancillary data are used for the emissions processing, such as spatial surrogates which spatially
allocate emissions to the 12 km grid cells used for air quality modeling.
261
-------
Use representative
EFs and county/
grid-specific
activity data and
meteorology to
create emissions
for all counties
i Liupviuiui i cii iyv.j,
rel. humidity
Run MOVES to get
emission factors
(EF) for
representative
counties for each
temperature and
speed needed
SMOKE
-0-
Activity
Data
AQ model-ready files
Figure 5-12: Modeling process of onroad emissions as part of the input for air quality modeling
5.4.1 Control Scenario Evaluated for the Air Quality Modeling Analysis
The control scenario evaluated for the air quality modeling analysis is different than the final
standards that are represented in the national emissions inventories discussed in Chapter 5.3.
Table 5-24 through Table 5-27 present the differences in duty-cycle NOx standards, warranty,
and useful life between the control scenario modeled for air quality modeling and the final
standards.
Table 5-24 compares the differences in the duty-cycle standards used in developing the
running, start, and extended idle emission rates in the final standards and the scenario used for
the air quality modeling analysis.
262
-------
Table 5-24: Duty-Cycle NOx Standards for the Final Standards and the Control Scenario Analyzed for Air
Quality Modeling
Model Year
Engine
Duty
Cycle
Duty-Cycle NOx Standards (mg/hp-hr)
Scenario Analyzed
for Air Quality
Modeling
Final Standards
2027
HHD, MHD,
LHD
FTP
100
35
[50]B
SET
50
35
[50]B
LLC
200
50
T651B
Idle0
18 g/hr
5 g/hr
HD SI
FTP
100
35
SET
50
35
2030A
HHD, MHD,
LHD
FTP
50
35
[50]B
SET
20
35
[50]B
LLC
100
50
T651B
Idle0
10 g/hr
5 g/hr
HD SI
FTP
50
35
SET
20
35
A The different duty-cycle NOx standards for MY2030 apply only to the air quality modeling control
scenario. The final standards have the same duty-cycle NOx standards for MY2027 and later.
B Values in brackets denote the 15 mg/hp-hr compliance margin for MHDE and HHDE that applies after the engines
are in-use in the final rule (see preamble Section III.B for details).
c In both scenarios, we assumed compliance with the Voluntary Idle standard which is more stringent than
the off-cycle standard as discussed in Chapter 5.2.2.1.1.2. Note that the Voluntary Idle standard in the final
control scenario that we modeled is different than the Voluntary Idle standard in the final program, see
preamble Sections III.B for details on the Voluntary Idle standard in the final program.
In the control scenario analyzed for air quality modeling, the FTP and SET standards are
different from one another. The Rauty values calculated from the FTP are applied to MOVES
running operating modes for vehicle speeds below 50 miles per hour, which aligns with the
transient behavior of the FTP cycle. The i?duty from the SET standard are applied to MOVES
operating modes above 50 mph (operating mode 33 and above), which aligns with the high-
speed activity that is targeted with the SET standard. However, the final FTP and SET standards
are equivalent, so we used the same i?duty values to calculate the emission rates for all running
operating modes, as discussed in Section 5.2.2.1.1.1.
263
-------
Table 5-25: Rduty Ratios Calculated for the Control Scenario Analyzed for Air Quality Modeling
Scenario
Applicable Model
Years
Emission standard
Rduty
(g/hp-hr)
FTP
SET-RMC
FTP
SET-RMC
Air Quality Modeling
Model Year 2027-2029
0.1
0.05
50%
25%
Model Year 2030+
0.05
0.02
25%
10%
Final Standards
Model Year 2027+
0.035
[0.05]A
0.035
[0.05]A
17.5%
[25%]A
17.5%
[25%]A
A Values in brackets denote the 15 mg/hp-hr compliance margin for MHDE and HHDE that applies after
the engines are in-use in the final rule (see preamble Section III.B for details).
The differences in the warranty and useful life periods analyzed for air quality modeling are
shown in Table 5-26 and Table 5-27.
Table 5-26: Warranty Mileages and Years in the Final Control Scenario and the Control Scenario Analyzed
for Air Quality Modeling
Model
Year
Engine
Warranty Mileage
Warranty Years
Air
Quality
Modeling
Control
Scenario
Final Standards
Air
Quality
Modeling
Control
Scenario
Final Standards
2027
HHD
350k
450k
5 years
7 years0
MHD
150k
280k
10 years
LHD
110k
210k
10 years
HD SI
110k
160k
5 years
10 years
2030A
HHD
600k
Same as 2027
7 years
Same as 2027
MHD
260k
Same as 2027
Same as 2027
LHD
200k
Same as 2027
Same as 2027
HD SI
160k
Same as 2027
7 years
Same as 2027
A The different warranty for MY2030 applies only to the air quality modeling
control scenario. The warranty for final standards is the same for MY2027 and
later.
B The FRM scenario we analyzed included a warranty years value of 7 years
instead of 10 years in the final standards scenario for HHD diesel.
264
-------
Table 5-27:Useful Life Mileages and Years in the Final Control Scenario and the Control Scenario Analyzed
for Air Quality Modeling
Model
Year
Engine
Useful Life Mileage
Useful Life Years
Air Quality
Modeling
Control Scenario
Final Standards
Air Quality
Modeling
Control Scenario
Final Standards
2027
HHD
650k
650k
10 years
11 years
MHD
325k
350k
12 years
LHD
250k
270k
15 years
HD SI
200k
200k
10 years
15 years
2030A
HHD
850k
Same as 2027
10 years
Same as 2027
MHD
450k
Same as 2027
Same as 2027
LHD
350k
Same as 2027
Same as 2027
HD SI
200k
Same as 2027
10 years
Same as 2027
A The different useful life for MY2030 applies only to the air quality modeling control scenario.
The useful life for final standards is the same for MY2027 and later.
5.4.2 Estimated Differences in the Emission Reductions between the Final Control Scenario
and the Control Scenario Analyzed for Air Quality Modeling
In addition to the differences between the final control scenario and the scenario modeled in
the air quality analysis, we have used an updated version of MOVES to develop the emission
inventories for the final rule, as described in Chapter 5.2. The combined net impact of the
differences in the control scenarios and the differences in the emission inventory methodology
are presented in this Chapter 5.4.2.
Overall, the estimated reductions from the final rule compare well with the reductions from
the SMOKE-MOVES inventory used in the air quality modeling, despite the differences in
modeling approaches (Chapter 5.2) and the control scenarios between the proposal and the final
rule (Chapter 5.4.1). Table 5-28 shows that both scenarios estimate large reductions inNOx,
similar reductions in PM2.5, and meaningful reductions in VOC, CO, and toxics. Based on this
comparison and the findings from the air quality modeling done for the proposal, we conclude
the final rule will lead to improvements in air quality.
265
-------
Table 5-28: Comparison of the Onroad Vehicle Emission Reductions from the Air Quality Modeling Control
Scenario vs. the Final Control Scenario
CY2045 Reduction in SMOKE-
CY2045 Reduction in MOVES National
Pollutant
MOVES Inventory (50 states) Used in
the Air Quality Modeling
Inventory (50 states) from the Final
Control Scenario
US Short Tons
% Reduction
US Short Tons
% Reduction
NOx
449,408
48.0%
453,239
43.3%
VOC
7,854
1.7%
20,758
3.7%
PM2.5 - Primary
548
1.4%
566
1.3%
Carbon Monoxide (CO)
167,241
3.6%
260,750
3.8%
Acetaldehyde
35
0.9%
145
3.3%
Benzene
113
1.6%
221
2.7%
Formaldehyde
46
1.6%
134
4.2%
Naphthalene
4
1.5%
9
2.7%
5.5 Chapter 5 Appendix
5.5.1 Zero-Mile Emission Rates for the Control Scenarios
The zero-mile NOx emission rates for HHD diesel vehicles in the final standards and
proposed Option 2 scenarios due to the duty-cycle and off-cycle standards are displayed in
Figure 5-5.
Figure 5-13 and Figure 5-14 display the zero-mile NOx emission rates for LHD45 and MHD
diesel vehicles in the final standards and proposed Option 2 scenarios.
MOVES Operating Mode
Figure 5-13: Estimated zero-mile emission rates for LHD45 diesel vehicles due to the final and proposed
Option 2 duty-cycle and off-cycle standards
266
-------
120
100
¦§ 80
* 60
z
CO
LU
> 40
20
0 1 11 12 13 14 15 16 21 22 23 24 25 27 28 29 30 33 35 37 38 39 40
MOVES Operating Mode
Figure 5-14: Estimated zero-mile emission rates for MHD diesel vehicles due to the final and proposed Option
2 duty-cycle and off-cycle standards
5.5.2 Details of the Emission Impacts of the Final Standards
In this section, we provide details of the national emission reductions from the heavy-duty
vehicles due to the final standards (previously summarized in Chapter 5.3.1).
267
-------
Running ¦ Start ¦ Extended Idle & APU
500,000
450,000
400,000
c 350,000
|2
^ 300,000
£Z
~ 250,000
o
tJ 200,000
~a
£ 150,000
100,000
50,000
0
I
• <& >£> . ,e> v^> . ^ ,e> s£>
0? e?> a?
CY2030 CY2040 CY2045
Figure 5-15: National NOx Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,2040,
and 2045 — for Each Fuel Type Category by Emission Process
~~M LHD2b3 ¦LHD45 ¦ MHD ¦HHP ¦ Urban Bus |
500,000
450,000
400,000
c 350,000
£
^ 300,000
250,000
o
o 200,000
ZJ
~o
£ 150,000
100,000
50,000
0
1
•
/• ^ ^
& c?
Figure 5-16: National NOx Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,2040,
and 2045 — for Each Fuel Type Category by HD Regulatory Class
268
-------
Running ¦ Start ¦ Extended Idle & APU ¦ Evap ¦ Refueling |
20,000
18,000
16,000
14,000
12,000
10,000
8,000
6,000
4,000
2,000
0
I
CY2030
// *
CY2040
• <& c$>
J** *
CY2045
Figure 5-17: National VOC Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,2040,
and 2045 — for Each Fuel Type Category by Emission Process
LHD2b3
LHD45
MHD
HHD
Urban Bus
20,000
18,000
16,000
£ 14,000
£
^ 12,000
~ 10,000
O
o 8,000
-a
g 6,000
4,000
2,000
0
I
&
CY2030
0>"
CY2040
6>
to"
CY2045
Figure 5-18: National VOC Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,2040,
and 2045 — for Each Fuel Type Category by HD Regulatory Class
269
-------
¦ Running ¦ Start ¦ Extended Idle & APU |
500
450
CY2030 CY2040 CY2045
Figure 5-19: National Exhaust PM2.5 Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,
2040, and 2045 — for Each Fuel Type Category by Emission Process
¦ LHD2b3 ¦ LHD45 ¦ MHD ¦HHP ¦ Urban Bus |
500
450
CY2030 | CY2040 CY2045
Figure 5-20: National Exhaust PM2.5 Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,
2040, and 2045 — for Each Fuel Type Category by HD Regulatory Class
270
-------
¦ Running ¦ Start ¦ Extended Idle & APU |
250,000
CY2030 CY2040 CY2045
Figure 5-21: National CO Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,2040, and
2045 — for Each Fuel Type Category by Emission Process
¦ LHD2b3 ¦ LHD45 ¦ MHD ¦HHP ¦ Urban Bus |
250,000
CY2030 | CY2040 CY2045
Figure 5-22: National CO Emission Reductions from Heavy-duty Vehicles in Calendar Years 2030,2040, and
2045 — for Each Fuel Type Category by HD Regulatory Class
5.5.3 Onroad Heavy-Duty NOx Emissions by Engine Operational Process for the Baseline,
Final and Proposed Option 2 Standards
Figure 5-23 displays the estimated national onroad heavy-duty NOx emissions in 2045 from
the baseline, final, and proposed Option 2 standards by engine operation process for the MY
2027 and later fleet impacted by the rule. See Section VI of the preamble for more discussion on
these comparisons.
271
-------
1,000,000
900,000
800,000
700,000
600,000
CO
=5
C 500,000
x
O
CO
13
C
C
<
400,000
300,000
200,000
100,000
0
Baseline
Extended Idle + APU
Starts
Running, Medium/High-Load
Running, Low-Load
Running, Age Effects
MY 2010-2026 Fleet
Final Standards Proposed Option 2
Figure 5-23: Comparison of Calendar Year 2045 Onroad Heavy-Duty NOx Emissions from Different Engine
Operational Process™ for the Baseline, Final Standards, and Proposed Option 2 Scenarios
5.5.4 Year-Over-Year Criteria Pollutant Emissions for Calendar Years Between 2027 and
2045
In this section, we present MOVES national inventory emissions (for selected criteria
pollutants) across multiple calendar years (2027-2045) for baseline and control scenarios.
The national heavy-duty vehicle emissions inventories are summarized in Table 5-29 through
Table 5-32 below for NOx, VOC, PM2.5 (exhaust), and CO, respectively, for the baseline, final
standards and proposed Option 2 scenarios. The same results are also displayed graphically in
Figure 5-24 through Figure 5-27.
EE In this graph, the "low-load running" emissions refer to the running exhaust emissions (for model year 2027 and
later without age effects) from MOVES running operating mode 0, 1, 11-14, 21-24, 33, plus 50% of operating mode
35 (see Table 5-3 for MOVES operating mode definitions). The remainder are considered "medium/high-load
running" emissions. The contribution of the "Running, Age Effects" was estimated by conducting a series of
MOVES runs with and without running aging effects for the baseline, final standards, and proposed Option 2
scenarios. The MOVES inputs without the aging effects are available in the rulemaking docket3.
272
-------
Table 5-29: National Heavy-duty Vehicle NOx Emissions (Annual US Tons) For Calendar Years Between
2027 and 2045
Calendar Year
Baseline
Final Standards
Proposed Option 2
2027
1,102,102
1,068,381
1,069,823
2028
1,059,062
990,842
993,761
2029
1,026,615
923,239
927,661
2030
996,202
856,525
862,503
2031
973,284
788,542
802,285
2032
954,582
726,649
747,755
2033
935,034
678,404
704,463
2034
925,914
641,732
672,513
2035
919,717
611,741
646,165
2036
917,205
587,210
625,054
2037
912,217
562,616
602,916
2038
914,585
546,719
589,342
2039
916,613
532,510
577,147
2040
916,684
517,820
564,216
2041
920,924
509,044
556,976
2042
925,534
502,047
551,409
2043
930,833
497,358
548,009
2044
937,395
494,049
545,975
2045
945,323
492,084
545,299
^—Baseline Final Standards Proposed Option 2
1,200,000
w 1,000,000
c
£ 800,000
00
=> 600,000
1 400,000
c
< 200,000
0
V^oPVVV^oPVVV'V>'V3'V5'V5'VJ'Vy'Vy'V5
Calendar Year
Figure 5-24: National Heavy-duty Vehicle NOx Emissions (Annual US Tons) For Calendar Years Between
2027 and 2045
273
-------
Table 5-30: National Heavy-Duty Vehicle VOC Emissions (Annual US Tons) For Calendar Years Between
2027 and 2045
Calendar Year
Baseline
Final Standards
Proposed Option 2
2027
119,297
118,094
118,094
2028
111,657
109,219
109,219
2029
108,273
104,569
104,569
2030
102,788
97,770
97,770
2031
99,012
92,632
92,652
2032
95,537
87,815
87,855
2033
92,868
83,728
83,777
2034
90,841
80,325
80,382
2035
88,878
77,076
77,138
2036
88,593
75,572
75,639
2037
86,610
72,446
72,515
2038
87,251
72,013
72,083
2039
87,826
71,607
71,678
2040
87,657
70,519
70,590
2041
88,448
70,481
70,554
2042
89,209
70,489
70,563
2043
89,964
70,559
70,633
2044
90,777
70,694
70,770
2045
91,810
71,053
71,130
^—Baseline Final Standards Proposed Option 2
l/)
c
£
LO
=>
ro
D
C
C
<
Calendar Year
Figure 5-25: National Heavy-Duty Vehicle VOC Emissions (Annual US Tons) For Calendar Years Between
2027 and 2045
130,000
120,000
110,000
100,000
90,000
80,000
70,000
60,000
274
-------
Table 5-31: National Heavy-duty Vehicle PM2.5 (Exhaust Only) Emissions (Annual US Tons) For Calendar
Years Between 2027 and 2045
Calendar Year
Baseline
Final Standards
Proposed Option 2
2027
17,790
17,762
17,762
2028
15,385
15,328
15,328
2029
14,101
14,015
14,015
2030
12,665
12,550
12,550
2031
11,627
11,446
11,454
2032
10,683
10,438
10,454
2033
9,889
9,597
9,624
2034
9,302
8,965
9,001
2035
8,795
8,426
8,466
2036
8,442
8,041
8,085
2037
7,544
7,118
7,163
2038
7,522
7,072
7,118
2039
7,495
7,024
7,070
2040
7,410
6,919
6,967
2041
7,390
6,882
6,930
2042
7,369
6,845
6,894
2043
7,356
6,818
6,868
2044
7,348
6,796
6,846
2045
7,357
6,791
6,842
Baseline Final Standards Proposed Option 2
20,000
14,000
CO
=) 12,000
E 10,000
CZ
^ '"ir *Xr '\r 'Va 'vi "ir "i? "i? 'T>> 'Vy 'Vy /vy '\r "ir 'vy
Calendar Year
18,000
Figure 5-26: National Heavy-duty Vehicle PM2.5 (Exhaust Only) Emissions (Annual US Tons) For Calendar
Years Between 2027 and 2045
275
-------
Table 5-32: National Heavy-Duty Vehicle CO Emissions (Annual US Tons) For Calendar Years Between 2027
and 2045
Calendar Year
Baseline
Final Standards
Proposed Option 2
2027
1,626,057
1,615,472
1,615,472
2028
1,538,135
1,516,707
1,516,707
2029
1,510,201
1,477,697
1,477,697
2030
1,469,592
1,425,614
1,425,614
2031
1,436,598
1,377,778
1,379,276
2032
1,403,726
1,330,388
1,333,299
2033
1,388,211
1,294,176
1,297,553
2034
1,370,037
1,255,992
1,259,824
2035
1,355,645
1,223,161
1,227,301
2036
1,356,898
1,206,858
1,211,295
2037
1,344,840
1,178,441
1,182,966
2038
1,346,192
1,164,406
1,169,030
2039
1,351,367
1,155,598
1,160,299
2040
1,347,716
1,138,782
1,143,539
2041
1,364,985
1,144,328
1,149,136
2042
1,378,756
1,147,320
1,152,196
2043
1,392,564
1,151,108
1,156,056
2044
1,407,837
1,156,631
1,161,649
2045
1,426,370
1,165,620
1,170,717
¦Baseline
-Final Standards
Proposed Option 2
1,700,000
1,600,000
^ 1,500,000
d
^ 1,400,000
1,300,000
c.
c=
^ 1,200,000
1,100,000
1,000,000
<5? <<£* <3? ^ ^ ^
•Vs it3 it3 if5 'V"> if? 'V3 "£? if3 'V3 'V"> If3 it? it^ /Vy 'V? it3 1>^ 'V?
Calendar Year
Figure 5-27: National Heavy-Duty Vehicle CO Emissions (Annual US Tons) For Calendar Years Between
2027 and 2045
276
-------
5.5.5 Sensitivity Analysis of Emissions Impacts of 2026 Service Class Pull Ahead Credits
Pathway
This section presents a sensitivity analysis of the estimated emission inventory impacts from
one of the transitional credit pathways under the ABT program included in the final rule. As
described in preamble Section IV.G.7, we are finalizing a transitional credit program that
includes several pathways for manufacturers to generate transitional credits in MYs 2022
through 2026 that they can then use in MYs 2027 and later. We conducted the sensitivity
analysis presented in this Appendix to evaluate the potential for additional emissions reductions,
particularly in the early years of the program, from allowing manufacturers to generate
transitional credits. We focused on the transitional credit pathway that provides the most
flexibility to use credits in MYs 2027 and later in order to assess whether these additional
flexibilities might impact the expected, additional early emissions reductions from the
transitional credits.
The results of the sensitivity analysis presented in this Appendix show that, compared to the
emissions reductions expected from the final rule, allowing manufacturers to generate
transitional credits through the pathway selected for analysis (i.e., the "2026 Service Class Pull
Ahead Credits Pathway") would result in additional emissions reductions in the early years of the
program. In the later years of the program, the emissions reductions would be essentially the
same with or without this transitional credits pathway. Below, we describe the methods used to
analyze the three scenarios included in the sensitivity analysis, and then present detailed results
of the comparison.
5.5.5.1 Modeling Scenario and MOVES Inputs
We estimated the emission impacts of the 2026 Service Class Pull Ahead Credits Pathway
using the same version of MOVES (MOVES3) as the final rule. The MOVES inputs for the final
rule are described in Chapter 5.2.2. The MOVES inputs for the 2026 Service Class Pull Ahead
Credits Pathway are based on the duty-cycle test standards, warranty and useful life requirements
described in preamble Section IV.G.7 and outlined immediately below.
For the purposes of this sensitivity analysis, we assumed that all heavy heavy-duty engines
produced in MY 2026 participated in the 2026 Service Class Pull Ahead Credits Pathway. As
such, we assumed that manufacturers certified all MY 2026 heavy-duty engines to FEL of 50
mg/hp-hr or less and met all other EPA requirements for MYs 2027 and later. Table 5-33
through Table 5-36 present our specific modeling inputs. We then assumed that manufacturers
used the credits generated by MY2026 engines to produce both heavy-heavy and medium-heavy-
duty engines at a FEL of 50 mg/hp-hr in MYs 2027 and later until the credits ran out; based on
our analysis, credits generated by heavy heavy-duty engines in MY 2026 are estimated to run out
approximately eight years later (i.e., in MY 2034).
277
-------
Table 5-33: Duty-Cycle NOx Standards for the 2026 Service Class Pull Ahead Credits Pathway and Final
ProgramA
Model
Year
Engine
Duty
Cycle
2026 Service
Class Pull Ahead
Credits
Pathway®
Final Program
2026
HHD, MHD
FTP
50 [65]D
(200)
SET
50 [65]D
(200)
LLC
71 [86]D
-
Idle0
7 g/hr
-
HD SI
FTP
-
(200)
SET
-
-
2027 and
later8
HHD, MHD, LHD
FTP
50 [65]D
35
SET
50 [65]D
35 [50]D
LLC
71 [86]D
50 [65]D
Idle0
7 g/hr
5 g/hr
HD SI
FTP
-
35
SET
-
35
A (#) = final standards with no change in the noted model year
B The 2026 Service Class Pull Ahead Credits program allows credits generated in MY 2026 to be used through MY
2034; thus, standards in these rows apply through MY 2034 for the 2026 Service Class Pull Ahead Credits Pathway
c We assumed compliance with the voluntary idle standard; note that the voluntary idle standard that we modeled is
different than the voluntary idle standard in the final program, see preamble Section III.B for details on the
voluntary idle standard in the final program.
D [HHDE in-use compliance margin]. Note that in the final program, the in-use compliance margin applies to both
HHDE and MHDE (see preamble Section III.B for details); for this sensitivity analysis, we modeled the compliance
margin applying only to HHDE.
E For this sensitivity analysis, we modeled only HHDE certifying to the requirements in the 2026 Service Class Pull
Ahead Credits Pathway in MY 2026; as discussed in preamble Section IV.G.7, only HHDE and MHDE can certify
to requirements of the 2026 Service Class Pull Ahead Credits Pathway and generate credits through this pathway.
Our analysis included both HHDE and MHDE using credits in MYs 2027 and later.
278
-------
Table 5-34: Off-Cycle Standards in 2026 Service Class Pull Ahead Credits Pathway and Final ProgramAC
Scenario
Model
Year
Regulatory
Class
Engine
Cycle
Off-
cycle
Bin
Off-Cycle NOx Standards (g/hr for
idling, g/hp-hr for low-load and
medium to high-load)
Idle
(g/hr)B
Idle, <
6%
power
5
LLC
Low-
load, 6-
20%
power
LHD,
MHD
(g/hp-
hr)
0.058
FTP &
SET
(g/hp-
hr)
Medium
Final
2027+
to High
Load,
>20%
power
0.058
Standards
Idle
Idle, <
6%
power
(g/hr)B
LLC
Low-
load, 6-
20%
power
HHD
(g/hp-
hr)
0.088
FTP &
SET
(g/hp-
hr)
Medium
to High
Load,
>20%
0.088
power
Idle
Idle, <
6%
power
2026
(g/hr)B
Service
Class Pull
Ahead
Credits
2026-
2034
MHD,
HHD
LLC
(g/hp-
hr)
Low-
load, 6-
20%
power
0.083 (0.113 HHDE)
PathwayA
FTP &
SET
(g/hp-
hr)
Medium
to High
Load,
>20%
0.083 (0.113 HHDE)
power
A For this sensitivity analysis, we modeled only HHDE certifying to the requirements in the 2026 Service Class Pull
Ahead Credits Pathway inMY2026; as discussed in preamble Section IV.G.7, only HHDE and MHDE can certify to
requirements of the 2026 Service Class Pull Ahead Credits Pathway and generate credits through this pathway. Our
analysis included both HHDE and MHDE using credits in MYs 2027 and later.
279
-------
B Note that the voluntary idle standard that we modeled is different than the voluntary idle standard in the final
program, see preamble Sections III.B and III.C for details on the voluntary idle standard in the final program and the
off-cycle standard for idle emissions, respectively.
c Note that we modeled all engine categories complying with off-cycle standards during in-use operations, which for
HHD includes an in-use compliance margin. In the final program, the in-use compliance margin applies to both
HHDE and MHDE (see preamble Section III.B for details); for this sensitivity analysis, we modeled the compliance
margin applying only to HHDE. The modeling for this sensitivity analysis also used a 30 mg/hp-hr compliance
margin for HHD off-cycle emissions and a 15 mg/hp-hr compliance margin for duty-cycle emissions; as discussed
in preamble Section III.B, the compliance margin for MHDE and HHDE in the final rule is 15 mg/hp-hr for both
off-cycle and duty-cycle emissions.
Table 5-35: Warranty Mileages and Years in 2026 Service Class Pull Ahead Credits Pathway and Final
ProgramA
Warranty Mileage
Warranty Years
Model
Year
Engine
2026
Service
Class Pull
Ahead
Credits
Pathway0
Final
Program
2026
Service
Class Pull
Ahead
Credits
Pathway
Final
Program
HHD
450k
(100k)
7 y
2026
MHD
-
(100k)
-
(5y)
LHD
-
(50k)
-
HD SI
-
(50k)
-
(5y)
HHD
450k
450k
2027+
MHD
280k
280k
7 y
7y
LHD
210k
210k
HD SI
160k
160k
7y
7y
A (#) = final standards with no change in the noted model year
B For this sensitivity analysis, we modeled only HHDE certifying to the requirements in the 2026 Service Class Pull
Ahead Credits Pathway inMY2026; as discussed in preamble Section IV.G.7, only HHDE and MHDE can certify to
requirements of the 2026 Service Class Pull Ahead Credits Pathway and generate credits through this pathway. Our
analysis included both HHDE and MHDE using credits in MYs 2027 and later.
280
-------
Table 5-36: Useful Life Mileages and Years in 2026 Service Class Pull Ahead Credits Pathway and Final
Program A
Model
Year
Engine
Useful Life Mileage
Useful Life Years
2026
Service
Class Pull
Ahead
Credits
Pathway0
Final
Program
2026
Service
Class Pull
Ahead
Credits
Pathway
Final
Program
2026
HHD
650k
(435k)
11
(10 y)
MHD
-
(185k)
-
LHD
-
(110k)
-
HD SI
-
(110k)
-
(10 y)
2027
HHD
650k
650k
ii y
ii y
MHD
350k
350k
12 y
12 y
LHD
270k
270k
15 y
15 y
HD SI
200k
200k
15 y
15 y
A (#) = final standards with no change in the noted model year
B For this sensitivity analysis, we modeled only HHDE certifying to the requirements in the 2026 Service Class Pull
Ahead Credits Pathway inMY2026; as discussed in preamble Section IV.G.7, only HHDE and MHDE can certify to
requirements of the 2026 Service Class Pull Ahead Credits Pathway and generate credits through this pathway. Our
analysis included both HHDE and MHDE using credits in MYs 2027 and later.
5.5.5.2 NOx Emissions Inventory Impacts of 2026 Service Class Pull Ahead Credits
The results of our sensitivity analysis, with the assumptions described in 5.5.5.1, are shown in
Figure 5-28. Our data show that including the 2026 Service Class Pull Ahead Credits pathway in
the final rule provides approximately 2 percent greater emissions reductions in CYs 2026
through 2031 compared to the final rule without the 2026 Service Class Pull Ahead Credits
pathway. In CYs 2032 through 2045, the emissions reductions from the final rule with the 2026
Service Class Pull Ahead Credits pathway in place are comparable to the final rule without the
transitional credits.
As described in 5.5.5.1, we assumed all heavy heavy-duty engines certified in MY 2026
would participate in the 2026 Service Class Pull Ahead Credits pathway, which likely
overestimates the volume of credits that would be generated and hence the magnitude of
additional emissions reductions in the early years of the program. However, our modeling did not
include the 10 percent discount that is part of the final pathway's requirements to move credits
between heavy heavy-duty engines and medium heavy-duty engine averaging sets; the 10
percent discount ensures that there will be a reduction of the overall emission level from
generating and using credits that are transferred between averaging sets. In addition, we assumed
that all credits generated in MY 2026 would be used; however, as the heavy-duty fleet continues
to transition to ZEVs, it is possible that manufacturers may not use all the credits generated,
which would result in greater emissions reductions than shown in our analysis.
As noted in 5.5.5.1, our sensitivity analysis represents one of several transitional credit
pathways in the final rule. However, since the transitional credit pathway we analyzed includes
281
-------
the most flexibilities for using credits in MYs 2027 and later, we believe the results are
indicative of the types of additional, early emissions reductions that the other transitional credit
pathways could provide.
Figure 5-28: Additional NOx Emissions Inventory Reductions in Early Program Years from 2026 Service
Class Pull Ahead Credits Program Compared to Final Program
-Baseline Final Program ^—2026 Service Class Pull Ahead Credits
1,200,000
1,100,000
1,000,000
l/l
l° 900,000
i/i
=> 800,000
i 700,000
c
<
600,000
500,000
400,000
^ ^ ^ ^ ^ ^ ^ ^
-------
5 US Energy Information Administration (EIA), Annual Energy Outlook 2019, Washington, DC:
January 2020, https://www.eia.gov/outlooks/archive/aeol9/
6 USEPA (2021). Population and Activity of Onroad Vehicles in MOVES3. EPA-420-R-21-
012. Assessment and Standards Division. Office of Transportation and Air Quality. US
Environmental Protection Agency. Ann Arbor, MI. April 2021.
https://www.epa.gov/moves/moves-onroad-technical-reports.
7 USEPA (2020). Fuel Effects on Exhaust Emissions from Onroad Vehicles in MOVES3. EPA-
420-R-20-016. Assessment and Standards Division. Office of Transportation and Air Quality.
US Environmental Protection Agency. Ann Arbor, MI. November 2020.
https://www.epa.gov/moves/moves-onroad-technical-reports.
8 Greenhouse Gas Emissions and Fuel Efficiency Standards for Medium- and Heavy-Duty
Engines and Vehicles—Phase 2. 81 FR 73941 (October 25, 2016)
9 2007/2010 Heavy-duty rulemaking. 66 FR 5002, January 18, 2001
10 USEPA (2020). Greenhouse Gas and Energy Consumption Rates for Onroad Vehicles
MOVES3. EPA-420-R-20-015. Assessment and Standards Division. Office of Transportation
and Air Quality. US Environmental Protection Agency. Ann Arbor, MI. November 2020.
https://www.epa.gov/moves/moves-onroad-technical-reports.
11 59 FR 16262, April 6, 1994
12 65 FR 6698, February 10, 2000.
13 79 FR 23414, April 28, 2014 and 80 FR 0978, February 19, 2015.
14 USEPA (2020). Evaporative Emissions from Onroad Vehicles in MOVES3. EPA-420-R-20-
012. Assessment and Standards Division. Office of Transportation and Air Quality. US
Environmental Protection Agency. Ann Arbor, MI. November 2020.
https://www.epa.gov/moves/moves-onroad-technical-reports.
15 U.S. EPA (2021) Technical Support Document: Air Quality Modeling for the HD 2027
Proposal.
283
-------
Chapter 6 Air Quality Impacts
This chapter presents information on air quality, including a discussion of current air quality
in Chapter 6.1, a discussion of air quality impacts from the final standards in Chapter 6.2, details
related to the methodology used for the proposal air quality modeling analysis in Chapter 6.3,
and results from the proposal air quality modeling analysis which are summarized in Chapter 6.4.
6.1 Current Air Quality
In this section we present information related to current levels of air pollutants, visibility
levels, and deposition amounts. This provides context for the need for this rule and a comparison
for the modeled projections from the rule.
6.1.1 Ozone
As described in Chapter 4 of this RIA, ozone causes adverse health effects, and EPA has set
national ambient air quality standards (NAAQS) to protect against those health effects. The
primary NAAQS for ozone, established in 2015 and retained in 2020, is an 8-hour standard with
a level of 0.07 ppm.A EPA recently announced that it will reconsider the decision to retain the
ozone NAAQS.B EPA is also implementing the previous 8-hour ozone primary standard, set in
2008 at a level of 0.075 ppm. As of August 31, 2022, there were 34 ozone nonattainment areas
for the 2008 primary ozone NAAQS, composed of 141 full or partial counties, with a population
of more than 90 million (see Figure 6-1); there were 49 ozone nonattainment areas for the 2015
primary ozone NAAQS, composed of 212 full or partial counties, with a population of more than
125 million (see Figure 6-2). In total, there were, as of August 31, 2022, 57 ozone nonattainment
areas with a population of more than 130 million people.c
A https://www. epa.gov/ground-level-ozone-pollution/ozone-national-ambient-air-quality-standards-naaqs.
B https://www.epa.gov/ground-level-ozone-pollution/epa-reconsider-previous-administrations-decision-retain-2015-
ozone.
c The total population is calculated by summing, without double counting, the 2008 and 2015 ozone nonattainment
populations contained in the Criteria Pollutant Nonattainment Summary report (https://www.epa.gov/green-
book/green-book-data-download).
284
-------
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 Classficatio
| Extreme
~ Severe 15
~ se,i„US
I Moderate
I Marginal
Figure 6-1: 8-Hour Ozone Nonattainment Areas (2008 Standard)
8-Hour Ozorie Nonattainment Areas (2015 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 Classification
t ¦ Extreme
I I Severe-17
~~Severe-15
I I Serious
~ Moderate
I I Marginal
| Marginal (R ural Transport)
For the Ozone-8Hr (2015) Cincinnati, OH-KY nonattainment area, the Ohio portion was redesignated on June 9,2022. The Kentucky portion has not been redesignated.
For the Ozone-8Hr (2015) Louisville, KY-IN nonattainment area, the Ohio portion was redesignated on July 5,2022. The Kentucky portion has not been redesignated.
The entire area is not considered in maintenance until all states in a multi-state area are redesignated.
Figure 6-2: 8-Hour Ozone Nonattainment Areas (2015 Standard)
285
-------
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. 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. Attainment dates for areas designated nonattainment for the 2015 ozone NAAQS are
in the 2021 to 2038 timeframe, again depending on the severity of the problem in each area.1 The
final standards will begin to take effect in 2027 and will assist areas with attaining the NAAQS
and may relieve areas with already stringent local regulations from some of the burden
associated with adopting additional local controls.0 The rule will also provide assistance to
counties with ambient concentrations near the level of the NAAQS who are working to ensure
long-term attainment or maintenance of the NAAQS.
6.1.2 PM2.5
As described in Chapter 4 of this RIA, PM causes adverse health effects, and EPA has set
NAAQS to protect against those health effects. There are two primary NAAQS for PM2.5: an
annual standard (12.0 micrograms per cubic meter ([j,g/m3)) and a 24-hour standard (35 (j,g/m3),
and there are two secondary NAAQS for PM2.5: an annual standard (15.0 jag/ m3) and a 24-hour
standard (35 (J,g/m3). The initial PM2.5 standards were set in 1997 and revisions to the standards
were finalized in 2006 and in December 2012, and then retained in 2020. On June 10, 2021, EPA
announced that it will reconsider the decision to retain the PM NAAQS.2
There are many areas of the country that are currently in nonattainment for the annual and 24-
hour primary PM2.5 NAAQS. As of August 31, 2022, more than 19 million people lived in the 4
areas that are designated as nonattainment for the 1997 annual PM2.5 NAAQS. Also, as of
August 31, 2022, more than 31 million people lived in the 14 areas that are designated as
nonattainment for the 2006 24-hour PM2.5 NAAQS, and more than 20 million people lived in the
5 areas designated as nonattainment for the 2012 annual PM2.5 NAAQS. In total, there are
currently 15 PM2.5 nonattainment areas with a population of more than 32 million people.E
Nonattainment areas for the PM2.5 NAAQS are pictured in Figure 6-3.
D While not quantified in the air quality modeling analysis for this rule, elements of the Averaging, Banking, and
Trading (ABT) program could encourage manufacturers to introduce new emission control technologies prior to the
2027 model year, which may help to accelerate some emission reductions of the final rule (See Preamble Section
IV.G for more details on the ABT program in the final rule).
E The population total is calculated by summing, without double counting, the 1997, 2006 and 2012 PM2 5
nonattainment populations contained in the Criteria Pollutant Nonattainment Summary report
(https://www.epa.gov/green-book/green-book-data-download).
286
-------
Counties Designated Nonattainment
for PM-2.5 (1997, 2006, and/or 2012 Standards)
Designated Nonattainment
three PM-2.5 Standards
I I Both 2006 and 2012 PM-2.5
I I Both 1997 and 2006 PM-2.5
M 2012 PM-2.5 only
I I 2006 PM-2.5 only
I I 1997 PM-2.5 only
Nonattainment areas are indicated by color.
VWien 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.
Figure 6-3: Counties Designated Nonattainment for PM2.5 (1997,2006, and/or 2012 standards)
The final standards will take effect in 2027 and will assist areas with attaining the NAAQS
and may relieve areas with already stringent local regulations from some of the burden
associated with adopting additional local controls.17 The rule will also provide assistance to
counties with ambient concentrations near the level of the NAAQS who are working to ensure
long-term attainment or maintenance of the PM2.5 NAAQS.
6.1.3 NO2
There are two primary NAAQS for NO2: an annual standard (53 ppb) and a 1-hour standard
(100 ppb).G In 2010, EPA established requirements for monitoring NO2 near roadways expected
to have the highest concentrations of NO2 within large cities. Monitoring within this near-
roadway network began in 2014, with additional sites deployed in the following years. At
present, there are no nonattainment areas for NO2.
F While not quantified in the air quality modeling analysis for this rule, elements of the Averaging, Banking, and
Trading (ABT) program could encourage manufacturers to introduce new emission control technologies prior to the
2027 model year, which may help to accelerate some emission reductions of the final rule (See Preamble Section
IV. G for more details on the ABT program in the final rule).
G The statistical form of the 1-hour NAAQS for N02 is the 3-year average of the yearly distribution of 1-hour daily
maximum concentrations.
287
-------
6.1.4 CO
There are two primary NAAQS for CO: an 8-hour standard (9 ppm) and a 1-hour standard (35
ppm). There are currently no CO nonattainment areas; as of September 27, 2010, all CO
nonattainment areas had been redesignated to attainment.
6.1.5 Air Toxics
The most recent available data indicate that millions of Americans live in areas where air
toxics pose potential health concerns.3 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 EPA's 2007 Mobile Source Air Toxics Rule.4 According to EPA's Air
Toxics Screening Assessment (AirToxScreen) for 2018, mobile sources were responsible for 40
percent of outdoor anthropogenic toxic emissions and were the largest contributor to national
average cancer and noncancer risk from directly emitted pollutants.5'11 Mobile sources are also
significant contributors to precursor emissions which react to form air toxics.6 Formaldehyde is
the largest contributor to cancer risk of all 71 pollutants quantitatively assessed in the 2018
AirToxScreen. Mobile sources were responsible for 26 percent of primary anthropogenic
emissions of this pollutant in 2018 and are significant contributors to formaldehyde precursor
emissions. Benzene is also a large contributor to cancer risk, and mobile sources account for
about 60 percent of average exposure to ambient concentrations.
6.1.6 Visibility
As of August 31, 2022, over 32 million people live in areas that are designated nonattainment
for the PM2.5 NAAQS. Overall, the evidence is sufficient to conclude that a causal relationship
exists between PM and visibility impairment.7 Thus, the populations who live in nonattainment
areas and travel to these areas will likely be experiencing visibility impairment. Additionally,
while visibility trends have improved in Mandatory Class I Federal areas, these areas continue to
suffer from visibility impairment.8'9'1 In summary, visibility impairment is experienced
throughout the U.S., in multi-state regions, urban areas, and remote Mandatory Class I Federal
areas.
6.1.7 Deposition
Over the past two decades, the EPA has undertaken numerous efforts to reduce nitrogen
deposition across the U.S. Analyses of monitoring data for the U.S. show that deposition of
nitrogen compounds has decreased over the last 25 years. At 34 long-term monitoring sites in the
eastern U.S., where data are most abundant, average total nitrogen deposition decreased by 43
percent between 1989-1991 and 2014-2016.10'n Although total nitrogen deposition has
decreased over time, many areas continue to be negatively impacted by deposition.
H AirToxScreen 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.
1 Mandatory Class I Federal areas are the 156 national parks and wilderness areas where state and federal agencies
work to improve visibility, https://www.epa.gov/visibility/regional-haze-program.
288
-------
6.2 Air Quality Impacts of the Final Rule
We expect the standards in the final rule to result in meaningful reductions in emissions of NOx,
VOC, CO and PM2.5 When feasible, we conduct full-scale photochemical air quality modeling
to accurately project levels of criteria and air toxic pollutants, because the atmospheric chemistry
related to ambient concentrations of PM2.5, ozone, and air toxics is very complex. Air quality
modeling was performed for the proposed rule and demonstrated improvements in
concentrations of air pollutants. We did not perform new air quality modeling for this final rule.
Chapter 5.4 of the RIA provides additional detail on the emissions inventory used for the
proposal's air quality modeling, including a comparison of the emission reductions modeled in
the air quality analysis and those from this final rule. Generally, despite the differences in the
version of the MOVES model used and the control scenario modeled, the emission reductions
used in the air quality modeling analysis compare well to the emission reductions estimated for
the final standards. Both scenarios result in reductions in emissions of VOC and PM2.5 and large
reductions in emissions of NOx, and we conclude that given the similar structure of the proposed
and final programs, the geographic distribution of emissions reductions and modeled
improvements in air quality are consistent and demonstrate that the final rule will lead to
substantial improvements in air quality.
We expect this rule will decrease ambient concentrations of air pollutants, including
significant improvements in ozone concentrations in 2045 as demonstrated in the air quality
modeling analysis. We also expect reductions in ambient PM2.5, NO2 and CO due to this rule.
Although the spatial resolution of the air quality modeling is not sufficient to quantify it, this
rule's emission reductions will also reduce air pollution in close proximity to major roadways,
where concentrations of many air pollutants are elevated and where people of color and people
with low income are disproportionately exposed. The emission reductions provided by the final
standards will be important in helping areas attain the NAAQS and prevent future nonattainment.
In addition, the final standards are expected to result in improvements in nitrogen deposition and
visibility. Additional information and maps showing modeled changes in ambient concentrations
of air pollutants in 2045 from the proposed standards are included in Chapter 6.4 of this RIA and
in the Air Quality Modeling Technical Support Document from the proposed rule.12
6.3 Air Quality Modeling Methodology for Proposal Analysis
6.3,1 Air Quality Model
CMAQ is a non-proprietary computer model that simulates the formation and fate of
photochemical oxidants, primary and secondary PM concentrations, acid deposition, and air
toxics, over regional and urban spatial scales for given inputs of meteorological conditions and
emissions. CMAQ includes numerous science modules that simulate the emission, production,
decay, deposition and transport of organic and inorganic gas-phase and particle pollutants in the
atmosphere. The CMAQ model is a well-known and well-respected tool and has been used in
numerous national and international applications/
1 More information available at: https://www.epa.gov/cmaq
289
-------
The air quality modeling analysis used the 2016vl platform with the most recent multi-
pollutant CMAQ code available at the time of air quality modeling (CMAQ version 5.3.1). The
2016 CMAQ runs utilized the CB6r3 chemical mechanism (Carbon Bond with linearized
halogen chemistry) for gas-phase chemistry, and AER07 (aerosol model with non-volatile
primary organic aerosol) for aerosols. The CMAQ model is regularly peer reviewed, with the
most recent review completed in 2019 on version 5.2 and 5.3beta.13
6.3.2 Model Domain and Configuration
The CMAQ modeling analyses used a domain covering the continental United States, as
shown in Figure 6-4. This single domain covers the entire continental U.S. (CONUS) and large
portions of Canada and Mexico using 12 km x 12 km horizontal grid spacing.K The 2016
simulation used a Lambert Conformal map projection centered at (-97, 40) with true latitudes at
33 and 45 degrees north. The model extends vertically from the surface to 50 millibars
(approximately 17,600 meters) using a sigma-pressure coordinate system with 35 vertical layers.
K The 12 km grid resolution of the air quality modeling domain does not allow us to analyze the concentration
gradients of NO2 and other pollutants which are likely to occur within a few hundred meters near roads.
290
-------
r% i i T"\
¦ AT\L }
12US2 domain "V k i \ | ,
x,y origin: -241200(mi, Vl62Ol0t)O(ri p v- 1j[ TV* \
col: 396 row:246 ^
Figure 6-4: Map of the CMAQ 12 km modeling domain (noted by the purple box)
6.3.3 Model Inputs
The key inputs to the CMAQ model include emissions from anthropogenic and biogenic
sectors, meteorological data, and initial and boundary conditions.
The onroad emissions inputs used for the 2045 reference and control scenarios are
summarized in Chapter 5 of the DRIA, and emissions inputs for other sectors are described in the
documentation for the 2016vl modeling platform.14 The reference scenario represents projected
2045 emissions without the proposed rule, and the control scenario represents projected 2045
emissions with the proposed rule. The AQM TSD also contains a detailed discussion of the
emissions inventory inputs used in our air quality modeling.22
The CMAQ meteorological input files were derived from simulations of the Weather
Research and Forecasting Model (WRF) version 3.8 for the entire 2016 year.1516 The WRF
Model is a state-of-the-science mesoscale numerical weather prediction system developed for
both operational forecasting and atmospheric research applications.17 The meteorological outputs
from WRF were processed to create 12 km model-ready inputs for CMAQ using the
Meteorology-Chemistry Interface Processor (MCIP) version 4.3. These inputs included hourly
varying horizontal wind components (i.e., speed and direction), temperature, moisture, vertical
diffusion rates, and rainfall rates for each grid cell in each vertical layer.18
291
-------
The boundary and initial species concentrations were provided by a northern hemispheric
CMAQ modeling platform for the year 2016.19'20 The hemispheric-scale platform uses a polar
stereographic projection at 108 km resolution to completely and continuously cover the northern
hemisphere for 2016. Meteorology is provided by WRF v3.8. Details on the emissions used for
hemispheric CMAQ can be found in the 2016 hemispheric emissions modeling platform TSD.21
The atmospheric processing (transformation and fate) was simulated by CMAQ (v5.2.1) using
the CB6r3 and the aerosol model with non-volatile primary organic carbon (AE6nvPOA). The
CMAQ model also included the on-line windblown dust emission sources (excluding agricultural
land), which are not always included in the regional platform but are important for large-scale
transport of dust.
6.3.4 CMAQ Evaluation
The CMAQ predictions for ozone, fine particulate matter, sulfate, nitrate, ammonium, organic
carbon, elemental carbon, nitrogen deposition, and specific air toxics (formaldehyde,
acetaldehyde, benzene and naphthalene) from the 2016 base scenario were compared to
measured concentrations in order to evaluate the ability of the modeling platform to replicate
observed concentrations. This evaluation was comprised of statistical and graphical comparisons
of paired modeled and observed data. Details on the model performance evaluation, including a
description of the methodology, the model performance statistics, and results, are provided in the
Air Quality Modeling TSD for this proposed rulemaking (AQM TSD).22
6.3.5 Model Simulation Scenarios
As part of our analysis for this rulemaking, the hourly CMAQ outputs were used to calculate
8-hour ozone design value concentrations, daily and annual PM2.5 design value concentrations,
annual NO2 concentrations, annual CO concentrations, annual and seasonal (summer and winter)
air toxics concentrations, visibility levels and annual total nitrogen deposition for each of the
following scenarios:
- 2016 base year
- 2045 reference
- 2045 control
Air quality modeling was done for the future year 2045 when the program will be fully
implemented and when most of the regulated fleet will have turned over. We use the predictions
from the air quality model in a relative sense by combining the 2016 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 for the May 1 - Sept 30 ozone
season, daily and annual PM2.5 concentrations, and visibility impairment for each of the 2045
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., 2014-2018).23 Additional predictions from the
CMAQ model are used in the demographic analysis (Chapter 6.4.9) and in the benefits analysis
described in Chapter 8.3.1 of the RIA. The CO, NO2, annual and seasonal formaldehyde,
acetaldehyde, benzene, naphthalene, and annual nitrate deposition projections were not predicted
in a relative sense due to the limited observational data available.
292
-------
The projected daily and annual PM2.5 design values were calculated using the Speciated
Modeled Attainment Test (SMAT) approach. 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)."24 Several updated
datasets and techniques were used for this analysis. These changes are fully described within the
technical support document for the Final Transport Rule Air Quality Modeling Technical
Support Document.25 The projected 8-hour ozone design values were calculated using the
approach identified in EPA's guidance on air quality modeling attainment demonstrations.26
6.4 Air Quality Modeling Results of the Proposed Rule
This section describes the results of the air quality modeling analysis done for the proposed
rule. The "reference" scenario represents projected 2045 air quality without the proposed rule
and the "control" scenario represents projected 2045 air quality with the proposed rule. This
section presents modeled changes in ambient concentrations of air pollutants when comparing
the "reference" and "control" scenarios. Decreases in concentration mean that the "control"
scenario decreases the pollutant concentration compared to the "reference" scenario.
Everything in the reference and control scenarios was held constant except the onroad
inventories, which reflected the application of the proposed standards at the time we conducted
the modeling. This includes the meteorological data (reflecting calendar year 2016 conditions)
and the emissions for all other sources, including boundary conditions and initial conditions used
in the air quality modeling methodology.
The reference and control scenarios include projections of existing control programs that EPA
had already adopted for mobile source emissions, as well as other federal, state and local
programs which are expected to reduce concentrations of pollutants in the ambient air in the
future. These control programs include (but are not limited to) the Tier 3 Motor Vehicle
Emission and Fuel Standards (79 FR 23414, April 28, 2014), the 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), and the Heavy-Duty Engine and
Vehicle Standards and Highway Diesel Fuel Sulfur Control Requirements (66 FR 5002, January
18,2001).
Not included in the reference or control scenarios for the air quality modeling are additional
federal or state programs that were not finalized at the time that the air quality modeling analysis
was initiated. For example, the CARB Heavy-Duty Low NOx Omnibus rule and the CA
Advanced Clean Trucks (ACT) rule were not final, so the emission reductions associated with
these rulemakings are not included in the air quality modeling analysis for the proposed rule.' XI
L Additional information on the CARB Omnibus program is available in Section I.D of the preamble for the
proposed rule. Additional discussion on the CARB ACT program is available in Sections I.D, VI.D, and XI of the
preamble for the proposed rule.
M The draft RIA Chapter 5 Appendix 6 presents a sensitivity analysis of the estimated emission inventory impacts
from nationwide adoption of the Omnibus rule; the draft RIA was made available with the proposed rule and is
available on the EPA website fortius rulemaking: https://www.epa.gov/regulations-emissions-vehicles-and-
engines/proposed-rule-and-related-materials-control-air-1.
293
-------
Since we did not include these rules in either the reference or control scenarios, our modeling for
this rule appropriately reflects the expected air quality improvements from this action.
6.4.1 Ozone Design Value Impacts of Proposed Rulemaking
This section summarizes the ozone air quality impacts of the proposed rule in 2045, based on
our CMAQ modeling. Our modeling indicates that ozone design value concentrations will
decrease dramatically in many areas of the country as a result of the proposed rule.
Figure 6-5 presents the changes in 8-hour ozone design value concentrations in 2045.N
Figure 6-5: Projected Change in 8-hour Ozone Design Values in 2045 due to Proposed Rule
As shown in Figure 6-5, the majority of the design value decreases in 2045 are greater than
1.5 ppb. There are also 82 counties with projected 8-hour ozone design value decreases of more
than 2.5 ppb; many of the counties with the largest design value decreases are in California, and
in the Atlanta and St. Louis urban areas. The maximum projected decrease in an 8-hour ozone
design value in 2045 is 5.1 ppb in Riverside County, California. Not all counties have monitor
N An 8-hour ozone design value is the concentration that determines whether a monitoring site meets the NAAQS
for ozone. The full details involved in calculating an 8-hour ozone design value are given in appendix I of 40 CFR
part 50.
294
-------
data that meets the requirements to calculate a design value concentration; counties without a
calculated design value are left white.
Table 6-1 shows the average projected change, due to the proposed rule, in 2045 8-hour ozone
design values for: (1) all modeled counties (with 2016 base case design values), (2) counties with
2016 base case design values that are above the level of the 2015 ozone NAAQS, (3) counties
with 2016 base case design values that are equal to or within 10% below the level of the 2015
NAAQS, (4) counties with 2045 reference scenario design values that are above the level of the
2015 ozone NAAQS, (5) counties with 2045 reference scenario design values that are equal to or
within 10% below the level of the 2015 ozone NAAQS, (6) counties with 2045 control scenario
design values that are above the level of the 2015 ozone NAAQS, and (7) counties with 2045
control scenario design values that are equal to or within 10% below the level of the 2015 ozone
NAAQS. Counties within 10 percent of the level of the NAAQS are intended to reflect counties
that although not violating the standards, would also be impacted by changes in ambient levels of
ozone as they work to ensure long-term attainment or maintenance of the ozone NAAQS. On a
population-weighted basis, the average modeled future-year 8-hour ozone design value is
projected to decrease by over 2 ppb in 2045 due to the proposed rule.
Table 6-1: Average Change in Projected 8-hour Ozone Design Values in 2045 due to Proposed Rule
Projected Design Value Category
Number of
Counties
2045
Population3
Average
Change
in 2045
Design
Value
(ppb)
Population-
Weighted
Average
Change in
Design
Value (ppb)
all modeled counties
457
246,949,949
-1.87
-2.23
counties with 2016 base year design values above the level
of the 2015 8-hour ozone standard
118
125,319,158
-2.12
-2.43
counties with 2016 base year design values within 10% of
the 2015 8-hour ozone standard
245
93,417,097
-1.83
-2.10
counties with 2045 reference design values above the level
of the 2015 8-hour ozone standard
15
37,758,488
-2.26
-3.03
counties with 2045 reference design values within 10% of
the 2015 8-hour ozone standard
56
39,302,665
-1.78
-2.02
counties with 2045 control design values above the level of
the 2015 8-hour ozone standard
10
27,930,138
-2.36
-3.34
counties with 2045 control design values within 10% of the
2015 8-hour ozone standard
42
31,395,617
-1.69
-1.77
aPopulation numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2015). Complete
Demographic Database. Washington, DC. http://www.woodsandpoole.com/index.php.
These modeling results project that there would be 15 counties with 8-hour ozone design
values above the level of the 2015 ozone NAAQS in 2045 without the proposed rule or any other
additional standards in place. Table 6-2 below presents the changes in design values for these
counties.
295
-------
Table 6-2: Change in 8-hour Ozone Design Values for Counties Projected to be Above the Level of the 2015 8-
hour Ozone NAAQS in 2045
County Name, State
Population in
2045a
Change in 2045
projected 8-hour
Ozone Design
Value (DV) (ppb)
2045 Reference
Ozone Design
Value (ppb)
2045 Control
Ozone Design
Value (ppb)
San Bernardino, California
3,191,663
-4.6
98.0
93.4
Los Angeles, California
11,755,545
-3.3
92.3
89.0
Riverside, California
3,926,478
-5.1
83.3
78.2
Fairfield, Connecticut
1,050,293
-1.4
79.4
78.0
Imperial, California
296,070
-0.3
76.6
76.3
Kern, California
1,251,350
-2.2
76.5
74.3
San Diego, California
4,452,722
-2.5
75.2
72.7
Fresno, California
1,371,355
-2.8
74.9
72.1
Richmond, New York
614,033
-0.8
73.7
72.9
Mariposa, California
20,630
-0.6
72.0
71.4
Salt Lake, Utah
1,387,960
-1.9
71.9
70.0
Tulare, California
601,851
-2.4
71.5
69.1
Sheboygan, Wisconsin
124,284
-1.9
71.4
69.5
Davis, Utah
556,296
-1.9
71.3
69.4
Harris, Texas
7,157,959
-2.2
71.2
69.0
a Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2015). Complete
Demographic Database. Washington, DC. http://www.woodsandpoole.com/index.php.
Our modeling predicts that the proposed rule would reduce ozone design values in some
counties from above the level of the standard to below it. While the number of counties with
projected design values above the level of the NAAQS is less certain than the average projected
changes in design values, our modeling projects that in 2045 ozone design values in five counties
(Salt Lake and Davis Counties in Utah, Tulare County in California, Sheboygan County in
Wisconsin, and Harris County in Texas) will change from being above the level of the standard
in the reference scenario to being below the level of the standard in the control scenario. The
projected population in these five counties in 2045 is almost 10 million people.
As described in Chapter 4 of this RIA, the science of ozone formation, transport, and
accumulation is complex. The air quality modeling projects ozone design value decreases as a
result of emissions changes from the proposed standards in the vast majority of counties. This
change in ozone results from interactions between photochemistry, background concentrations of
ozone, VOC and NOx, local emissions and meteorology. However, there is one county in 2045
that is projected to have no change in modeled ozone design value concentration (Skagit County,
Washington).
6.4,2 Annual PMxs Design Value Impacts of Proposed Rulemaking
This section summarizes the annual average PM2.5 air quality impacts of the proposed rule in
2045, based on our CMAQ modeling. Our modeling indicates that annual PM2.5 design values
will decrease due to the proposed rule. The decreases in annual PM2.5 design values are due to
the projected reductions in NOx, primary PM2.5, and VOC emissions. We expect this rule's
reductions in directly emitted PM2.5 will also contribute to reductions in PM2.5 concentrations
near roadways, although our air quality modeling is not of sufficient resolution to capture that
impact.
296
-------
Figure 6-6 presents the changes in annual PM2.5 design values in 2045.°
Figure 6-6: Projected Change in Annual PM2.5 Design Values in 2045 due to Proposed Rule
As shown in Figure 6-6, we project that in 2045 most counties will have design value
decreases of between 0.01 ug/rrr and 0.05 |ig/m3. There are also 15 counties with projected
annual PM2.5 design value decreases of more than 0 .1 |ig/mJ; these counties are in California and
Utah. The maximum projected decrease in a 2045 annual PM2.5 design value is 0.21 ug/'m3 in
Tulare County, California. Not all counties have monitor data that meets the requirements to
calculate a design value concentration; counties without a calculated design value are left white.
Table 6-3 presents the average projected change, due to the proposed rule, in 2045 annual
PM2.5 design values for: (1) all modeled counties (with 2016 base case design values), (2)
counties with 2016 base case design values that are above the level of the 2012 annual PM2.5
standard, (3) counties with 2016 base case design values that are equal to or within 10% below
the level of the 2012 standard, (4) counties with 2045 reference scenario design values that are
above the level of the 2012 annual PM2.5 standard, (5) counties with 2045 reference scenario
design values that are equal to or within 10% below the level of the 2012 annual PM2.5 standard,
An annual PM2j 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.
297
-------
(6) counties with 2045 control scenario design values that are above the level of the 2012 annual
PM2.5 standard and (7) counties with 2045 control scenario design values that are equal to or
within 10% below the level of the 2012 annual PM2.5 standard. Counties within 10 percent of the
level of the standard are intended to reflect counties that although not violating the standards,
would also be impacted by changes in ambient levels of PM2.5 as they work to ensure long-term
attainment or maintenance of the 2012 annual PM2.5 NAAQS. On a population-weighted basis,
the average modeled future year annual PM2.5 design value is projected to decrease by 0.04
|ig/m3 due to the proposed rule.
Table 6-3: Average Change in Projected Annual PM2.5 Design Values in 2045 due to Proposed Rule
Projected Design Value Category
Number
of
Counties
2045
Population3
Average
Change in
2045
Design
Value
(ug/m3)
Population-
Weighted
Average
Change in
Design Value
(ug/m3)
all modeled counties
568
273,604,437
-0.04
-0.04
counties with 2016 base year design values above
the level of the 2012 annual PM2.5 standard
17
26,726,354
-0.09
-0.05
counties with 2016 base year design values
within 10% of the 2012 annual PM2.5 standard
5
4,009,527
-0.06
-0.06
counties with 2045 reference design values above
the level of the 2012 annual PM2.5 standard
12
25,015,974
-0.10
-0.05
counties with 2045 reference design values
within 10% of the 2012 annual PM2.5 standard
6
1,721,445
-0.06
-0.06
counties with 2045 control design values above
the level of the 2012 annual PM2.5 standard
10
23,320,070
-0.10
-0.05
counties with 2045 control design values within
10% of the 2012 annual PM2 5 standard
8
3,417,349
-0.08
-0.09
a Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2015). Complete
Demographic Database. Washington, DC. http://www.woodsandpoole.com/index.php.
There are 12 counties, mostly in California, that are projected to have annual PM2.5 design
values above the level of the NAAQS in 2045 without the proposed rule or any other additional
standards in place. Table 6-4 below presents the changes in design values for these counties.
298
-------
Table 6-4: Change in Annual PM2.5 Design Values for Counties Projected to be Above the Level of the Annual
PM2.5 NAAQS in 2045
County Name, State
Population
in 2045a
Change in 2045
Projected Annual
PM2.5 Design Value
(DV) (ug/m3)
2045 Reference
Design Value
(fig/m3)
2045 Control
Design Value
(jig/m3)
Kern, California
1,251,350
-0.15
15.78
15.63
Kings, California
185,866
-0.19
14.82
14.63
San Bernardino, California
3,191,663
-0.02
14.21
14.19
Plumas, California
21,297
-0.05
14.12
14.06
Tulare, California
601,851
-0.21
14.05
13.84
Riverside, California
3,926,478
-0.04
13.43
13.39
Imperial, California
296,070
-0.02
12.93
12.91
Fresno, California
1,371,355
-0.16
12.87
12.71
Los Angeles, California
11,755,545
-0.02
12.26
12.24
Pinal, Arizona
718,595
-0.11
12.24
12.13
Stanislaus, California
716,019
-0.16
12.17
12.02
San Joaquin, California
979,885
-0.11
12.07
11.96
a Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2015). Complete
Demographic Database. Washington, DC. http://www.woodsandpoole.com/index.php.
Our modeling predicts that the proposed rule would reduce annual PM2.5 design values in
some counties from above the level of the standard to below it. While the number of counties
with projected design values above the level of the NAAQS is less certain than the average
changes in design values, annual PM2.5 design values in two counties (Stanislaus County,
California and San Joaquin County, California) are projected to change from being above the
level of the standard in the reference scenario to being below the level of the standard in the
control scenario. The projected population in these two counties in 2045 is over 1.5 million
people.
6,4.3 24-hour PM2.5 Design Value Impacts of Proposed Rulemaking
This section summarizes the 24-hour PM2.5 air quality impacts of the proposed rule in 2045,
based on our CMAQ modeling. Our modeling indicates that most 24-hour PM2.5 design values
would decrease due to the proposed rule. The decreases in 24-hour PM2.5 design values are due
to the projected reductions in NOx, primary PM2.5, and VOC emissions. We expect this rule's
reductions in directly emitted PM2.5 will also contribute to reductions in PM2.5 concentrations
near roadways, although our air quality modeling is not of sufficient resolution to capture that
impact.
Figure 6-7 presents the changes in 24-hour PM2.5 design values in 2045.p
p A 24-hour PM2 5 design value is the concentration that determines whether a monitoring site meets the 24-hour
NAAQS for PM2 5. The Ml details involved in calculating a 24-hour PM2 5 design value are given in appendix N of
40 CFR part 50.
299
-------
. ?.*
, h 'r
,U
i J
ri -
A¦" "
;. '-> » W-- i »
I i' • '* t i '
fr^-x'y
T
r
#/
Legend
hj-n-ja- */Q_».ii.«
I <- -0.5 u&'"n3
| => -{! 5 ba c=-CI v5
I > -0.251n c= -U 15
1fl
2f:
-35
-3 151C «=-0Q5 275
| >-3 0510 =D.061d<0.15 1
| >=0.1 S to <0.25 5
| >= D 751n ^ G 5 C
I *= 0 6 $
,'f
7
H
">A
•. •
•t. •-
Courtly Site Difference MDaiVy PM2.SOV — 2W5ffi_ctf_C7Tm>mrs 104Stlt jef_CTI
Figure 6-7: Projected Change in 24-hour PM2.5 Design Values in 2045 due to Proposed Rule
As shown in Figure 6-7, in 2045 there are 170 counties with projected 24-hour PM2.5 design
value decreases greater than 0.15 ug/m3. These counties are in mainly in the midwest, southeast
and western United States. The maximum projected decrease in a 2045 24-hour PM2.5 design
value is 1.79 ug/rrr in Tulare County, California. Not all counties have monitor data that meets
the requirements to calculate a design value concentration; counties without a calculated design
value are left white.
Table 6-5 shows the average projected change, due to the proposed rule, in 2045 24-hour
PM2.5 design values for: (1) all modeled counties (with 2016 base case design values), (2)
counties with 2016 base case design values that are above the level of the 2006 24-hour PM2.5
standard, (3) counties with 2016 base case design values that are equal to or within 10% below
the level of the 2006 24-hour PM2.5 standard, (4) counties with 2045 reference scenario design
values that are above the level of the 2006 24-hour PM2.5 standard, (5) counties with 2045
reference scenario design values that are equal to or within 10% below the level of the 2006 24-
hour PM2.5 standard, (6) counties with 2045 control scenario design values that are above the
level of the 2006 24-hour PM2.5 standard, and (7) counties with 2045 control scenario design
values that are equal to or within 10% below the level of the 2006 24-hour PM2.5 NAAQS.
Counties within 10 percent of the level of the standard are intended to reflect counties that
although not violating the standards, would also be impacted by changes in ambient levels of
PM2.5 as they work to ensure long-term attainment or maintenance of the 2006 24-hour PM2.5
300
-------
NAAQS. On a population-weighted basis, the average modeled future-year 24-hour PM2.5
design value is projected to decrease by 0.17 |ig/m3 in 2045 due to the proposed rule.
Table 6-5: Average Change in Projected 24-hour PM2.5 Design Values in 2045 due to Proposed Rule
Projected Design Value Category
Number
of
Counties
2045
Population3
Average
Change
in 2045
Design
Value
(ug/m3)
Population-
Weighted
Average
Change in
Design Value
(ug/m3)
all modeled counties
568
272,852,777
-0.12
-0.17
counties with 2016 base year design values above the
level of the 2006 daily PM2.5 standard
33
28,394,253
-0.40
-0.67
counties with 2016 base year design values within 10%
of the 2006 daily PM2.5 standard
15
13,937,416
-0.18
-0.27
counties with 2045 reference design values above the
level of the 2006 daily PM2.5 standard
29
14,447,443
-0.38
-0.55
counties with 2045 reference design values within 10%
of the 2006 daily PM2.5 standard
12
22,900,297
-0.30
-0.59
counties with 2045 control design values above the
level of the 2006 daily PM2.5 standard
29
14,447,443
-0.38
-0.55
counties with 2045 control design values within 10%
of the 2006 daily PM2.5 standard
10
19,766,216
-0.26
-0.60
a Population numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2015). Complete
Demographic Database. Washington, DC. http://www.woodsandpoole.com/index.php.
There are 29 counties that are projected to have 24-hour PM2.5 design values above the level
of the NAAQS in 2045 without the proposed rule or any other additional controls in place. Table
6-6 below presents the changes in design values for these counties.
301
-------
Table 6-6: Change in 24-hour PM2.5 Design Values for Counties Projected to be Above the 24-hour PM2.5
NAAQS in 2045
County Name, State
Population in
2045a
Change in 24-
hour PM2.5
Design Value
Oig/m3)
2045 Reference
Design Value
(fig/m3)
2045 Control
Design Value
(fig/m3)
Okanogan, Washington
47,922
-0.24
57.02
56.79
Ravalli, Montana
52,336
-0.03
56.93
56.90
Kern, California
1,251,350
-0.49
55.78
55.29
Fresno, California
1,371,355
-0.68
49.86
49.17
Jackson, Oregon
281,974
-0.28
49.22
48.94
Kings, California
185,866
-1.42
47.23
45.81
Plumas, California
21,297
-0.16
46.30
46.14
Klamath, Oregon
71,950
-0.16
44.39
44.24
Siskiyou, California
46,491
-0.03
44.04
44.02
Lincoln, Montana
19,924
-0.14
43.05
42.91
Tulare, California
601,851
-1.79
42.63
40.84
Missoula, Montana
139,759
-0.13
42.49
42.36
Lemhi, Idaho
8,830
-0.08
42.46
42.39
Lewis and Clark, Montana
95,256
-0.06
41.17
41.11
Flathead, Montana
150,424
-0.12
40.75
40.64
Yakima, Washington
289,388
-0.22
40.64
40.42
Lake, Oregon
8,605
-0.10
40.43
40.33
Stanislaus, California
716,019
-1.26
39.54
38.28
Lane, Oregon
440,599
-0.13
39.53
39.39
Josephine, Oregon
106,207
-0.27
39.46
39.19
Alameda, California
1,936,700
-0.30
38.81
38.51
Madera, California
208,957
-0.62
38.49
37.87
San Joaquin, California
979,885
-1.22
38.15
36.93
Kittitas, Washington
53,927
-0.17
37.82
37.65
Riverside, California
3,926,478
-0.51
37.35
36.84
Shoshone, Idaho
11,064
-0.16
37.07
36.91
Crook, Oregon
24,645
-0.19
37.00
36.81
Benewah, Idaho
10,426
-0.13
36.72
36.59
Salt Lake, Utah
1,387,960
0.02
36.04
36.06
aPopulation numbers based on Woods & Poole data. Woods & Poole Economics, Inc. (2015). Complete
Demographic Database. Washington, DC. http://www.woodsandpoole.com/index.php.
While the count of modeled nonattainment counties is much less certain than the average
changes in air quality, in 2045, there are no 24-hour PM2.5 design values that are projected to
change from being above the level of the standard in the reference case to being below the level
of the standard in the proposed control case.
As described in Chapter 4 of this RIA, PM2.5 in the atmosphere can be primary or secondary
and its composition, transport, and accumulation is complex. The air quality modeling projects
24-hour PM2.5 design value decreases as a result of emissions changes from the proposed rule in
the vast majority of counties. However, there are a handful of counties where 24-hour PM2.5
design values are projected to increase. These increases are likely due to elevated secondary
PM2.5 formation rates resulting from increased oxidant levels which occur during stagnant cold
weather due to reductions in NOx.
302
-------
6.4.4 Nitrogen Dioxide Concentration Impacts of Proposed Rulemaking
This section summarizes the annual average NO2 air quality impacts of the proposed rule in
2045, based on our CMAQ modeling. Our modeling indicates that annual average NO2
concentrations would decrease as a result of the proposed rule, if finalized as proposed. Figure
6-8 presents the changes in annual NO2 concentrations in 2045.
As shown in Figure 6-8, our modeling indicates that by 2045 annual NO2 concentrations in
the majority of the country would decrease between 0.01 and 0.1 ppb due to the proposed rule.
However, decreases in annual NO2 concentrations would be greater than 0.2 ppb along many
highway corridors and greater than 0.3 ppb in most urban areas. The absolute reductions
correspond to reductions of greater than 5 percent in annual NO2 concentrations across much of
the country, see Figure 6-9. Although we didn't model changes in 1-hour concentrations, the
proposed rule would also likely decrease 1-hour NO2 concentrations and help any potential
nonattainment areas attain and maintenance areas maintain the NO2 standard.*2 We expect this
rule will also contribute to reductions in NO2 concentrations near roadways, although our air
quality modeling is not of sufficient resolution to capture that impact.R
Q As noted in Chapter 6.1.3, there are currently no nonattainment areas for the N02 NAAQS.
R The 12 km grid resolution of the air quality modeling domain does not allow us to analyze the concentration
gradients of NO2 and other pollutants which are likely to occur within a few hundred meters near roads.
303
-------
ppc-
M -- -g .30
mm -0.30 to -0.2c
B -0.20 Li} -&.LC
0.20 ID 0.01
0.U1 Ml 0.U1
u.tn nam
0,10:»(J.20
mm o 30 -0 0 30
^m 0.30
Figure 6-9: Percent Change in Annual Ambient NO2 Concentrations in 2045
6.4.5 Carbon Monoxide Concentration Impacts of Proposed Rulemaking
This section summarizes the annual average CO air quality impacts of the proposed rule in
2045, based on our CMAQ modeling. Our modeling indicates that annual average CO
concentrations would decrease as a result of the proposed rule. Figure 6-10 and Figure 6-11
present the absolute and percent changes in annual CO concentrations in 2045.
Figure 6-8: Projected Absolute Change in Annual Ambient N02 Concentrations in 2045
•: -30 .U
-SD.OIXi-25 C-
-25.OlQ-IO.C-
-2D.0 tD -5.0
-5.0 ju -2.5
2.5 ¦do 1.0
¦ .U m 1.0
1 -:i tn 7 r:
2.5 b.D
5.0 tD 10 0
10.0 X} 25.0
25.0 itj 50.0
> 5D.0
304
-------
As shown in Figure 6-10, our modeling indicates that by 2045 annual CO concentrations in
the majority of the country would decrease between 0.02 and 0.5 ppb due to the proposed
rulemaking. However, decreases in annual CO concentrations would be greater than 1.5 ppb in
some urban areas. The absolute reductions correspond to percent changes of less than 1 percent
across the country, except for the Phoenix area, where there are some larger decreases between 1
and 2 percent. Although we didn't model changes in 8-hour or 1-hour concentrations, the
proposed standards would also likely decrease 1-hour and 8-hour CO concentrations and help
any potential nonattainment areas attain and maintenance areas maintain the CO standard^
| ; ! I ¦ vr-'l
i ! - - - v~j X. ppt-
• I. , \ "" ' ¦¦ < -3 DO
, ;.'v / ^ -3.00 to-2.5C
- _ V' H -2.50Lo-2.0C
) /y'"\ f ¦ -2.00tD-l.5C
! I ¦' .A ¦¦ 1.50 to 1.00
-y;- h .1 .oo to
fc* ; i"""1 •, - - y" -o.txDto-o.o2
. - ' . • ' -D O? to 0.0?
J 5 ^
HgcDCCBI Win; 3065 >. '
Figure 6-10: Absolute Change in Annual Ambient CO Concentrations in 2045
s As noted in Chapter 6.1.4, there are currently no nonattainment areas forthe CO NAAQS.
305
-------
Mw;DQ105Mn; ? r 'v :* -a '•"! >
Figure 6-11: Percent Change in Annual Ambient CO Concentrations in 2045
6.4.6 Air Toxics Impacts of Proposed Rulemaking
This section summarizes the changes in annual average air toxic (acetaldehyde, benzene,
formaldehyde and naphthalene) concentrations in 2045 due to the proposed rule. Our modeling
indicates that the proposed rule would have relatively little impact on national average ambient
concentrations of the modeled air toxics in 2045. Annual percent changes are less than 1% for air
toxics across most of the country. Annual absolute changes in ambient concentrations are
generally less than 0.001 |ig/m3 for benzene and naphthalene (Figure 6-12 and Figure 6-13
below). There are small increases in acetaldehyde across the country, see Figure 6-14. The
increases in acetaldehyde likely occur because species that lead to production or recycling of
acetaldehyde increase as their reactions with nitrogen oxides decrease. For formaldehyde there
are decreases across most of the country and a few areas with increases, see Figure 6-15. The
increases in formaldehyde concentration due to the proposed rule are likely related to higher
concentrations of OH radicals in areas where ozone increases due to NOx emissions reductions
(see Chapter 4.1.1.1).
306
-------
¦¦ 0.200 CO 0.300 M >5o.o
¦¦ >0.300
Figure 6-12: Changes in Ambient Benzene Concentrations in 2045 due to Proposed Rule: Absolute Changes
in iLig/m1 (left) and Percent Changes (right)
m*L
Figure 6-13: Changes in Ambient Naphthalene Concentrations in 2045 due to Proposed Rule: Absolute
Changes in fig/m3 (left) and Percent Changes (right)
307
-------
if*=v-Lj-!-r" "f—"#r I KD. I ,
b i i 1\ I .V ¦ r - I V ; >
I < 0 300
I -0.300 to -0.200
I -0.200 to-0.100
I 0.100 to-0.010 1
-0.010 to-0.001 -
-0.001 to 0.001
0.001 to 0.010
0 010 to 0.100
i 0.100 to 0.200
I 0 200 to 0 300
I >0.300
1 1/ ') K
ug/m3 | \ «- ¦ - \
VsJ
V'
i , - \if & ¦>
Figure 6-14: Changes in Ambient Acetaldehyde Concentrations in 2045 due to Proposed Rule: Absolute
Changes in (ug/m1 (left) and Percent Changes (right)
Figure 6-15: Changes in Ambient Formaldehyde Concentrations in 2045 due to Proposed Rule: Absolute
Changes in jiig/m1 (left) and Percent Changes (right)
6.4.7 Visibility Impacts of Proposed Rulemaking
Air quality modeling was used to project visibility conditions in 145 Mandatory Class I
Federal areas across the U.S. with and without the proposed rule in 2045. The results show that
in 2045, the proposed rule would improve projected visibility on the 20% most impaired days in
all modeled areas 1 The average visibility on the 20 percent most impaired days at all modeled
T 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
308
-------
Mandatory Class I Federal areas is projected to improve by 0.04 deciviews, or 0.37 percent, in
2045. The greatest improvement in visibility would occur in San Gorgonio and San Jacinto
Wilderness Areas in California, where visibility is projected to improve by 1.56 percent (0.21
deciviews) in 2045 due to the proposed rule. The AQM TSD contains the full visibility results
from 2045 for the 145 analyzed areas.22
6.4,8 Deposition Impacts of Proposed Rulemaking
Our air quality modeling projects decreases in nitrogen deposition due to the proposed rule.
Figure 6-16 shows that by 2045 the proposed rule would result in decreases in nitrogen
deposition over much of the eastern US and in urban areas of the western US, with the largest
decreases in Atlanta and Los Angeles. Figure 6-17 indicates those decreases correspond to
annual percent decreases of more than one percent over much of the country, with some
localized decreases of over 4 percent. As discussed in Chapter 4.2.3.1, there is considerable
evidence that nitrogen deposition adversely affects terrestrial, wetland, freshwater, and estuarine
ecosystems. The reductions in nitrogen deposition expected from this rule, along with other
actions to reduce NOx emissions, will reduce acidification and nitrogen enrichment across the
US, including the Chesapeake Bay and other water bodies. This will lead to improved
ecosystem functions, reduced coastal eutrophication, increased recreational demand, and other
beneficial effects.
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.
309
-------
Figure 6-16: Absolute Change in Annual Deposition of Nitrogen in 2045
¦ j?
¦r
*
%
< -10.0
10.3 Us 5.C
5.0 lf> 4.0
-¦a.-:? to -3-D
-3.0 to -1.0
-1.0 to-0.5
0.5 Id 0.5
0.5- to 1.0
1 0 to 3.0
3.0 10 4.0
>4.0
\
y \
V 4
m*.- or-u 'cm-
Figure 6-17: Percent Change in Annual Deposition of Nitrogen in 2045
6.4.9 Demographic Analysis of Air Quality
When feasible, EPA's Office of Transportation and Air Quality conducts full-scale
photochemical air quality modeling to demonstrate how its national mobile source regulatory
actions affect ambient concentrations of regional pollutants throughout the United States. As
described in Chapter 6.2, the air quality modeling we conducted for the proposal also supports
310
-------
our analysis of future projections of PM2.5 and ozone concentrations in a "baseline" scenario
absent the rule and in a "control" scenario that assumes the rule is in place.u These baseline and
control scenarios are also used as inputs to the health benefits analysis. As demonstrated in
Chapter 6.4 and Chapter 8, the ozone and PM2.5 improvements that are projected to result from
the rule, and the health benefits associated with those pollutant reductions, will be substantial.
This air quality modeling data can also be used to conduct an analysis of how human
exposure to future air quality varies with sociodemographic characteristics relevant to potential
environmental justice concerns in scenarios with and without the rule in place. Although the
spatial resolution of the air quality modeling is not sufficient to capture very local heterogeneity
of human exposures, particularly the pollution concentration gradients near roads, the analysis
does allow estimates of demographic trends at a national scale. We developed this approach by
considering the purpose and specific characteristics of this rulemaking, as well as the nature of
known and potential exposures to the air pollutants controlled by the standards. The heavy-duty
standards apply nationally and will be implemented consistently across roadways throughout the
U.S. The pollutant predominantly controlled by the standard is NOx. Reducing emissions of
NOx will reduce formation of ozone and secondarily formed PM2.5, which will reduce human
exposures to regional concentrations of ambient ozone and PM2.5 These reductions will be
geographically widespread. Taking these factors into consideration, this demographic analysis
evaluates the exposure outcome distributions that will result from this rule at the national scale
with a focus on locations that are projected to have the highest baseline concentrations of PM2.5
and ozone.
To analyze trends in exposure outcomes, we sorted projected 2045 baseline air quality
concentrations from highest to lowest concentration and created two groups: areas within the
contiguous U.S. with the worst air quality (grid cells with the highest 5 percent of
concentrations) and the rest of the country (remaining 95 percent of grid cells). This approach
can then answer two principal questions to determine disparity of air quality on the basis of race
and ethnicity:
1. What is the racial and ethnic composition of areas with the worst baseline air quality in
2045?
2. Are those with the worst air quality benefitting more from the heavy-duty vehicle and
engine standards?
We found that in the 2045 baseline, nearly double the number of people of color live within
areas with the worst ozone and PM2.5 air pollution compared to non-Hispanic Whites (NH-
Whites).v We also found that (in absolute terms) the largest predicted improvements in both
u Air quality modeling was performed for the proposed rule, which used emission reductions that are very similar to
the emission reductions projected for the final rule. Given the similar structure of the proposed and final programs,
we expect consistent geographic distribution of emissions reductions and modeled improvements in air quality, and
that the air quality modeling conducted at the time of proposal adequately represents the final rule. Specifically, we
expect this rule will decrease ambient concentrations of air pollutants, including significant improvements in ozone
concentrations in 2045 as demonstrated in the air quality modeling analysis.
v The demographic analysis uses air quality modeling that has a contiguous U.S. domain. The analysis does not
characterize distributional trends in areas of the U.S. that fall outside of this domain.
311
-------
ozone and PM2.5 are estimated to occur in areas with the worst baseline air quality, and a larger
number of people of color are projected to reside in these areas.
6.4.9.1 Data and Methods
We began with projected 2045 baseline and control scenarios of modeled PM2.5 and ozone
concentration data (described in RIA Chapter 8.2.1). Ambient air quality concentration data
(annual average |ig/m3 for PM2.5 and May-September daily maximum 8-hour average ppb for
ozone) was estimated at a standard grid resolution of 12km x 12km across the contiguous United
States (CONUS).w Using 2045 baseline air quality data as our reference scenario, we sorted
baseline air quality concentrations from highest to lowest concentration and compared two air
quality concentration groups - grid cells in the highest 5 percent of the distribution of baseline
concentrations and grid cells in the remaining 95 percent.x
Figure 6-18The maps in Figure 6-18 display the spatial distribution of grid cells with baseline
concentrations in the highest 5 percent for both PM2.5 and ozone concentrations. We retain this
distinction throughout the analysis in order to track how air quality is distributed by air quality
concentration group in the baseline and how the rule impacts air quality in these same grid cells
with the standards in place.
The analysis also used population projections stratified by race/ethnicity, age, and sex are
based on proprietary economic forecasting models developed by Woods and Poole in 2015.27
The Woods and Poole database contains county-level projections of population by age, sex, and
race out to 2050, relative to a baseline using the 2010 Census data. Population projections for
each county are determined simultaneously with every other county in the U.S to consider
patterns of economic growth and migration.Y The projected population for 2045 was extracted
from the Environmental Benefits Mapping and Analysis Program - Community Edition
(BenMAP-CE)z at the same 12km x 12km grid resolution as the air quality data. Race and
ethnicity of individuals projected to live in a given area were compiled into two broad categories,
"people of color" and "Non-Hispanic White (NH-White).',AA We chose to aggregate race and
ethnicity categories in this way to address the uncertainty present with population projections far
into the future - it is difficult to predict with precision patterns of economic growth and
migration (see Section 6.4.9.3 for more discussion about uncertainty). In 2045, there are 409
million people projected to be living in the contiguous United States; 208 million are projected to
w Note that the ambient PM2 5 and ozone air quality concentration data used in this analysis are different than the
PM2.5 and ozone design value metrics presented in RIA Chapter 6. Design values are pollutant concentrations that
determine whether a monitoring site meets the NAAQS for a given pollutant.
x Using higher and lower percentiles to compare risks, exposures and outcomes has been applied by EPA's Office of
Air and Radiation in previous distributional analyses of regulatory air quality modeling (see MATS, CSAPR, PM
NAAQS) and is consistent with EPA's EJ Technical Guidance.
Y More information about the population projections can be found in the Technical Support Document (TSD) for the
Final Revised Cross-State Air Pollution Rule Update for the 2008 Ozone Season NAAQS:
https://www.epa.gOv/sites/default/files/2021-03/documents/estimating_pm2.5-_and_ozone-
attributable_health_benefits_tsd.pdf
z More information about BenMAP-CE can be found here: https://www.epa.gov/benmap. Additional information
about the population projections used in this analysis can be found in Appendix J of the BenMAP-CE User's
Manual: https://www.epa.gov/benmap/benmap-ce-manual-and-appendices.
^ "People of color" includes Black, Asian, Native American, Hawaiian/Pacific Islander, and Hispanic populations.
312
-------
be NH-White and 201 million are projected to be people of color. To put these projections into
perspective, 2010 populations for the contiguous United States were 201 million for NH-White
and 106 million for people of color.
Additionally, this analysis looked at the distribution of poverty status within the same air
quality concentration groups - 12km x 12km grid cells in the highest 5 percent of the distribution
of baseline concentrations and grid cells in the remaining 95 percent. We applied county-level
poverty status derived from the Census' American Community Survey (ACS) 5-year estimates
from 2015 to 2019, which represents the fraction of county-level population below and above
200% of the poverty line.BB We note that measures of "current" pover ty are not necessarily
predictors of future poverty status; poverty status in the 2045 population could be different in
terms of both scale and geographic location. However, for the purposes of this analysis, we
believe applying a "current" measure of those who live above and below 200% of the poverty
line is illustrative.
For each pollutant and air quality concentration group (i.e., highest 5% of concentration or
remaining 95% of grid cells), we calculated the average baseline, control, and reduction in
concentrations. We then summed the population by group (people of color or NH-White;
populations above or below 200% of the poverty line) for each air quality concentration group.
P *1>A ¦'ffX __/77W f \ '1 1 IS Jfl
ZX'- Vi •• .Jrr—k % i -W
\ ¦ nzm % is Ft r/rr^
NIT v\ vf \\
V 4 V
(a) (b)
Figure 6-18: Distributional maps of populated 12km grid cells across the contiguous United States in 2045.
Darker areas represent the location of grid cells within the highest 5 percent of baseline concentrations for (a)
PM2.5 and (b) ozone
EPA received comments related to the methods the Agency used to analyze the distribution of
impacts of the heavy-duty vehicle and engine standards. After consideration of comments, we
have retained the demographic analysis from the proposal based upon the data and methods
described above. However, in response to comments that the Agency consider the disparate
impacts of the rule through the analysis of race/ethnicity-stratified impacts, we have added an
analysis of the demographic composition of population-weighted national average air quality
impacts. For scenarios with and without the rule in place, we present national average air quality
®B County-level poverty status was mapped to the 12km x 12km grid cell domain using spatial weighting in
BenMAP-CE.
313
-------
concentrations that are weighted by specific race and ethnicity populations. Using the same air
quality and population data described above, we sum the product of each projected CMAQ grid-
cell population and its corresponding CMAQ grid-cell air quality concentration and then divide
by the total population by race/ethnicity. As described in Section 6.4.9.3, we caution that the
population projection data by race and ethnicity is uncertain and that the spatial resolution of the
air quality modeling is not sufficient to capture local heterogeneity of human exposures.
6.4.9.2 Results
Of the approximately 48,000 populated CMAQ grid cells that encompass the contiguous
United States, nearly 2,400 are in the highest 5 percent of the baseline distribution. For PM2.5, the
concentration at the 95th percentile is 7.76 |ig/m3 (median: 5.18 |ig/m3), and for ozone it is 49.91
ppb (median: 38.34 ppb). In 2045, 144 million people are projected to live within the highest 5
percent of grid cells for PM2.5 and 39 million are projected to live in areas with the highest
concentrations of ozone (Figure 6-18).
As shown in Table 6-7, in 2045, the number of people of color projected to live within the
grid cells with the highest baseline concentrations of ozone (26 million) is nearly double that of
NH-Whites (14 million). Thirteen percent of people of color are projected to live in areas with
the worst baseline ozone, compared to seven percent of NH-Whites. The rule will reduce human
exposures to ambient ozone for all population groups, but the 39 million people living in areas
with the worst air quality will experience a greater reduction in ozone than the 370 million
people living in the remaining 95 percent of grid cells.
314
-------
Table 6-7: Demographic Analysis of Projected 2045 Ozone Concentrations (ppb), Sorted by Average Baseline
Ozone Air Quality: NH-White and People of Color
Seasonal Average Ozone Concentrations in ppb
(5% to 95% Range)
2045 Population - millions
Baseline
Control
Reduction
All 12km x
12km Grid Cells
in CONUS
(n-47,795)
Total Population in
CONUS - All Grid Cells
409
39.18
(29.91 -49.91)
38.71
(29.42 -49.50)
0.47
(0.16-0.90)
Non-Hispanic White
Population in CONUS
208
People of Color
Population in CONUS
201
Highest 5% of
Baseline Ozone
Concentrations
(n=2,391)
Population in Highest 5%
(% of Total in CONUS)
39
(10%)
52.59
(50.01 -58.20)
51.95
(49.57 - 57.22)
0.64
(0.27 - 1.44)
Non-Hispanic White
(% of NH-W in CONUS)
14
(7%)
People of Color
(% of POC in CONUS)
26
(13%)
Remaining 95%
of Baseline
Ozone
Concentrations
(n-45,404)
Population in Remaining 95%
(% of Total in CONUS)
370
(90%)
38.47
(29.82 - 48.70)
38.01
(29.34 -48.31)
0.46
(0.16-0.87)
Non-Hispanic White
(% of NH-W in CONUS)
194
(93%)
People of Color
(% of POC in CONUS)
176
(86%)
PM2.5 results have a similar pattern to what we observe for ozone. As shown in Table 6-8, in
2045, the number of people of color projected to live within the grid cells with the highest
baseline concentrations of PM2.5 (93 million) is nearly double that of NH-Whites (51 million).
Forty-six percent of people of color are projected to live in areas with the worst baseline PM2.5,
compared to 25 percent of NH-Whites. Those in areas with the worst air quality will experience
a greater reduction in PM2.5 than those in the remaining 95 percent of grid cells.
315
-------
Table 6-8: Demographic Analysis of Projected 2045 PM2.5 Concentrations (jig/m3), Sorted by Average PM2.5
Baseline Air Quality: NH-White and People of Color
Annual Average PM2.5 Concentrations in jig/m3
(5% to 95% Range)
2045 Population - millions
Baseline
Control
Reduction
All 12km x
12km Grid
Cells in
CONUS
(n=47,795)
Total Population in
CONUS - AU Grid Cells
409
5.23
(2.65 - 7.76)
5.21
(2.65 - 7.72)
0.022
(0.003 - 0.052)
Non-Hispanic White
Population in CONUS
208
People of Color
Population in CONUS
201
Highest 5% of
Baseline
PM2.5
Concentration
s (n=2,391)
Population in Highest 5%
(% of Total in CONUS)
144
(35%)
9.03
(7.80 - 12.07)
8.99
(7.76 - 12.01)
0.044
(0.008 -0.097)
Non-Hispanic White
(% of NH-W in CONUS)
51
(25%)
People of Color
(% of POC in CONUS)
93
(46%)
Remaining
95% of
Baseline
PM2.5
Concentration
s (n=45,404)
Population in Remaining 95%
(% of Total in CONUS)
265
(65%)
5.03
(2.62 - 7.24)
5.01
(2.62 - 7.20)
0.020
(0.003 - 0.049)
Non-Hispanic White
(% of NH-W in CONUS)
156
(75%)
People of Color
(% of POC CONUS)
108
(54%)
In Table 6-9 and Table 6-10, we looked at populations above and below 200% of the federal
poverty line. Using 2045 population estimates, 126 million people are projected to live below
200% of the poverty line, 13 million (10 percent) of whom will also be living in an area with the
worst baseline concentrations of ozone. Similarly, 10 percent of people projected to live above
200% of the poverty line will also be living in an area with the worst baseline concentrations of
ozone. For PM2.5, 37 percent of those living below 200% of the poverty line will also be living in
areas with the worst baseline concentrations of PM2.5, compared to 35 percent of the population
above 200% of the poverty line projected to live in those same areas. While some disparity exists
for PM2.5, overall, the demographic results for those living above and below 200% of the poverty
line are not as pronounced in the areas with the worst air quality as they are for race and
ethnicity.
316
-------
Table 6-9: Demographic Analysis of Projected 2045 Ozone Concentrations (ppb), Sorted by Average Baseline
Ozone Air Quality: Poverty Status
Seasonal Average Ozone Concentrations in ppb
(5% to 95% Range)
2045 Population - millions
Baseline
Control
Reduction
All 12km x
12km Grid
Cells in
CONUS
(n=47,795)
Total Population in
CONUS - AU Grid Cells
409
39.18
(29.91 -49.91)
38.71
(29.42 -49.50)
0.47
(0.16-0.90)
Population Below 200% of
the Poverty Line in CONUSa
126
Population Above 200% of
the Poverty Line in CONUS
283
Highest 5% of
Baseline
Ozone
Concentrations
(n=2,391)
Population in Highest 5%
(% of Total in CONUS)
39
(10%)
52.59
(50.01 -58.20)
51.95
(49.57 - 57.22)
0.64
(0.27 - 1.44)
Population Below 200%
of the Poverty Line
(% Below 200% in CONUS)
13
(10%)
Population Above 200%
of the Poverty Line
(% Below 200% in CONUS)
26
(10%)
Remaining
95% of
Baseline
Ozone
Concentrations
(n=45,404)
Population in Lowest 95% (%
of Total in CONUS)
370
(90%)
38.47
(29.82 - 48.70)
38.01
(29.34 -48.31)
0.46
(0.16-0.87)
Population Below 200%
of the Poverty Line
(% Below 200% in CONUS)
112
(90%)
Population Above 200%
of the Poverty Line
(% Below 200% in CONUS)
257
(90%)
a Note that the poverty measure used here is based on ACS 5-year estimates from 2015 to 2019 at the county
level representing the fraction of county-level population below and above 200% of the poverty line. Counts of 2045
population reflect projections based on 2010 Census Data and population growth factors estimated by Woods &
Poole (2015). Measures of "current" poverty are not necessarily predictors of future poverty status.
317
-------
Table 6-10: Demographic Analysis of Projected 2045 PM2.5 Concentrations (jig/m3), Sorted by Average PM2.5
Baseline Air Quality: Poverty Status
Annual Average PM2.5 Concentrations in jig/m3
(5% to 95% Range)
2045 Population - millions
Baseline
Control
Reduction
All 12km x 12km
Grid Cells in
CONUS
(n=47,795)
Total Population in
CONUS - AU Grid Cells
409
5.23
(2.65 - 7.76)
5.21
(2.65 - 7.72)
0.022
(0.003 - 0.052)
Population Below 200% of
the Poverty Line in CONUSa
126
Population Above 200% of
the Poverty Line in CONUS
283
Highest 5% of
Baseline PM25
Concentrations
(n=2,391)
Population in Highest 5%
(% of Total in CONUS)
144
(35%)
9.03
(7.80 - 12.07)
8.99
(7.76 - 12.01)
0.044
(0.008 -0.097)
Population Below 200%
of the Poverty Line
(% Below 200% in CONUS)
46
(37%)
Population Above 200%
of the Poverty Line
(% Below 200% in CONUS)
98
(35%)
Remaining 95%
of Baseline PM25
Concentrations
(n=45,404)
Population in Lowest 95% (%
of Total in CONUS)
265
(65%)
5.03
(2.62 - 7.24)
5.01
(2.62 - 7.20)
0.020
(0.003 - 0.049)
Population Below 200%
of the Poverty Line
(% Below 200% in CONUS)
79
(63%)
Population Above 200%
of the Poverty Line
(% Below 200% in CONUS)
185
(65%)
a Note that the poverty measure used here is based on ACS 5-year estimates from 2015 to 2019 at the county
level representing the fraction of county-level population below and above 200% of the poverty line. Counts of 2045
population reflect projections based on 2010 Census Data and population growth factors estimated by Woods &
Poole (2015). Measures of "current" poverty are not necessarily predictors of future poverty status.
318
-------
While the demographic analyses in Table 6-7 and Table 6-8 demonstrate the possible
disparity that exists between NH-Whites and people of color, we have expanded the analysis of
air quality impacts experienced by specific race and ethnic groups. In Table 6-11, we present the
national population-weighted average ozone concentrations for each specific race and ethnicity
category in scenarios without (baseline) and with (control) the rule in place. We also present the
reduction in ozone (from baseline to control) for each race and ethnicity category along with the
relative reduction from baseline expressed as a percentage. To highlight the changes in each
category, results are color-coded by air quality (concentrations increase from light blue to dark
blue) and by air quality improvements (reductions increase from light green to dark green).
On a population-weighted basis, all race and ethnicity population categories experience
reductions in exposure to ozone as a result of the rule. NH-Black populations experience the
lowest concentrations of ozone (both with and without the rule in place), while also experiencing
the greatest reductions. NH-Native Americans are projected to live in areas with the highest
ozone concentrations while also experiencing slightly smaller reductions from this rule compared
to other race and ethnicity populations. In relative terms, the percent reduction in ozone
experienced by each race and ethnicity category ranges from a 1.33% reduction from baseline
(NH-Native American) to a 1.92% reduction from baseline (NH-Black).
In Table 6-12, we present the national population-weighted average PM2.5 concentrations for
each specific race and ethnicity category using the same color-coding scheme described for
Table 6-11. On a population-weighted basis, all race and ethnicity population categories
experience reductions in exposure to PM2.5 as a result of the rule. The largest reductions in PM2.5
are projected to occur in areas where NH-Black populations reside. NH-Native Americans
experience the lowest concentrations of PM2.5 (both with and without the rule in place) and are
projected to receive slightly smaller reductions from this rule compared to other race and
ethnicity populations. Hispanic populations are projected to experience the highest PM2.5
concentrations in both the baseline and control scenarios. In relative terms, the percent reduction
in PM2.5 experienced by each race and ethnicity category ranges from a 0.40% reduction from
baseline (Hispanic, NH-Asian, NH-Native American) to a 0.52% reduction from baseline (NH-
Black).
319
-------
Table 6-11: Demographic Analysis of National Average 2045 Ozone Concentrations (ppb), by Race/Ethnicity"
2045 Population
(million)
Baseline Ozone
Concentration
Control Ozone
Concentration
Reduction in
Ozone
% Reduction
in Ozone
All Race/Ethnicity
409
39.30
38.66
0.64
1.64%
NH-White
208
38.61
37.94
0.67
1.72%
Hispanic
108
41.10
40.51
0.60
1.45%
NH-Black
55
37.55
36.83
0.72
1.92%
NH-Asian
35
40.46
39.91
0.56
1.38%
NH-Native American
3
41.34
40.78
0.55
1.33%
a National averages are weighted by population. We sum the product of each projected CMAQ grid-cell population
and its corresponding CMAQ grid-cell air quality concentration and then divide by the total population.
Table 6-12: Demographic Analysis of National Average 2045 PM2.5 Concentrations (jig/m3), by
Race/Ethnicity"
2045 Population
(million)
Baseline PM25
Concentration
Control PM25
Concentration
Reduction in
PM25
% Reduction
in PM25
All Race/Ethnicity
409
7.26
7.23
0.034
0.47%
NH-White
208
6.83
6.79
0.035
0.51%
Hispanic
108
7.90
7.87
0.031
0.40%
NH-Black
55
7.50
7.41
0.039
0.52%
NH-Asian
35
7.64
7.61
0.030
0.40%
NH-Native American
3
6.18
6.15
0.025
0.40%
a National averages are weighted by population. We sum the product of each projected CMAQ grid-cell population
and its corresponding CMAQ grid-cell air quality concentration and then divide by the total population.
6.4.9.3 Uncertainty in the Demographic Analysis
The results of this demographic analysis are dependent on the available input data and its
associated uncertainty. As we note in both the air quality modeling and health benefits chapters,
uncertainties exist along the entire pathway from emissions to air quality to population
projections and exposure. The demographic analysis (including poverty status) is subject to these
same sources of uncertainty.
A limitation of this analysis is the 12km x 12km horizontal grid spacing of the air quality
modeling domain. Such resolution is unable to capture the heterogeneity of human exposures to
pollutants within that area, especially pollutant concentration gradients that exist near roads. EPA
is considering how to better estimate the near-roadway air quality impacts of its regulatory
actions and how those impacts are distributed across populations. Because the heavy-duty
320
-------
standards apply nationally and will be implemented consistently across roadways throughout the
U.S., we can still make useful observations of demographic trends at a national scale using the
air quality modeling data at a 12km x 12km resolution.
Another key source of uncertainty is the accuracy of the projected baseline concentrations of
PM2.5 and ozone because we use modeled 2045 baseline air quality as the basis for our
comparisons. Assumptions that influence projections of future air quality include emissions in
the future baseline (stationary source emissions are only projected out to year 2028 and held
constant out to 2045) and the meteorology used to model air quality (2016 conditions). With this
uncertainty in mind, we prefer to examine the air quality impacts of the standards by comparing
the baseline scenario to the control scenario in order to highlight incremental changes in air
quality due to the standards. By looking at the incremental change, any underlying uncertainty
present in both the modeled baseline and control air quality data is largely offset. However, when
we rank grid cells from dirtiest to cleanest using 2045 baseline concentrations, the uncertainties
associated with the baseline take on greater importance when interpreting the results of the
analysis.
There is also inherent uncertainty in the Woods & Poole-based populations projected out to
2045. As mentioned above, the population projections are based on proprietary economic
forecasting models developed by Woods and Poole in 2015 and are relative to a baseline using
the 2010 Census data. Underlying the population projections are forecasted variables such as
income, employment, and population. Each of these forecasts require many assumptions:
economy-wide modeling to project income and employment, net migration rates based on
employment opportunities and taking into account fertility and mortality, and the estimation of
age/sex/race distributions at the county-level based on historical rates of mortality, fertility, and
migration. To the extent these patterns and assumptions have changed since the population
projections were estimated, and to the extent that these patterns and assumptions may change in
the future, we would expect the projections of future population would be different than those
used in this analysis.
For the analysis of exposure trends organized by baseline concentration, we attempted to
address some of this population projection uncertainty by compiling race and ethnicity into two
broad categories, "people of color" and "NH-White." Such broad groupings help avoid overly
precise interpretations of inherently uncertain projections of population and demographics,
especially when looking at areas of worst air quality against the remaining areas across the
contiguous United States. EPA continues to investigate how best to incorporate population
projections into our analyses to disaggregate populations of concern by relevant socioeconomic
variables, and to identify the interactions between demographic changes and air quality changes.
In response to this commitment, while acknowledging the uncertainty in future population
projections by specific race and ethnicity category, we have expanded the demographic analysis
of air quality impacts to include specific race and ethnicity groups. Tables 6-11 and 6-12 report
air quality exposure trends by race and ethnicity at the national level.
The measure of poverty status used in this analysis is based on data from the American
Community Survey representing the rate of poverty between 2015-2019 and is not projected to
reflect poverty status in the future. This assumption is inherently uncertain since measures of
"current" poverty are not necessarily predictors of future poverty status.
321
-------
We intend to continue to refine demographic analyses in future rulemakings including
potentially assessing how much of the results may or may not be driven by emissions changes
compared to projected changes in demographics.
Finally, we note that the air quality scenario we modeled to support the air quality, benefits,
and demographic analyses for this rulemaking is based on modeling conducted for the proposal.
Despite the differences between the modeled scenario and the final standards, the emission
reductions used in the air quality modeling analysis for the proposed rule are similar to those
associated with the final standards, see Chapter 5.4. Both scenarios result in reductions in
emissions of VOC and PM2.5 and large reductions in emissions of NOx, and we expect that the
final rule will also lead to substantial improvements in air quality.
Chapter 6 References
1 https://www.epa.gov/ground-level-ozone-pollution/ozone-naaqs-timelines.
2 https://www.epa.gov/pm-pollution/national-ambient-air-quality-standards-naaqs-pm
3 U.S. EPA (2022) Technical Support Document EPA Air Toxics Screening Assessment.
2017AirToxScreen TSD. https://www.epa.gov/system/files/documents/2022-
03/airtoxscreen_2017tsd.pdf
4 U.S. Environmental Protection Agency (2007). Control of Hazardous Air Pollutants from
Mobile Sources; Final Rule. 72 FR 8434, February 26, 2007.
5 U.S. EPA. (2022) 2018 Air Toxics Screening Assessment.
https://www.epa.gov/AirToxScreen/2018-airtoxscreen-assessment-results
6 Rich Cook, Sharon Phillips, Madeleine Strum, Alison Eyth & James Thurman (2020):
Contribution of mobile sources to secondary formation of carbonyl compounds, Journal of the
Air & Waste Management Association, DOI: 10.1080/10962247.2020.1813839
7U.S. EPA. Integrated Science Assessment (ISA) for Particulate Matter (Final Report, 2019).
U.S. Environmental Protection Agency, Washington, DC, EPA/600/R-19/188, 2019.
8 EPA's Report on the Environment: Regional Haze,
http s:// cfpub. epa.gov/roe/indicator. cfm?i=21
9 Regional Haze Storymap, accessed in 2020 from epa.gov/visibility.
https://epa.maps.arcgis.com/apps/Cascade/index.html?appid=e4dbe2263elf49fb849aflc73a04e2
f2
10 EPA Report on the Environment, technical documentation.
https://cfpub.epa.gov/roe/technical-documentation.cfm?i=l&pvw=.
11 Trends data comes from the EPA Report on the Environment. Accessed in 2020,
https://cfpub.epa.gov/roe/indicator.cfm?i=l#4 Based on data from the NADP/National Trends
Network, 2018.
322
-------
12 US EPA, 2021. Technical Support Document (TSD) Air Quality Modeling for the HD 2027
Proposal.
13 The Sixth External Peer Review of the Community Multiscale Air Quality (CMAQ) Modeling
System. Available online at: https://www.epa.gov/sites/production/files/2019-
08/documents/sixth_cmaq_peer_review_comment_report_6.19.19.pdf
14 National Emissions Inventory Collaborative (2019). 2016vl Emissions Modeling Platform.
Retrieved from http://views.cira.colostate.edu/wiki/wiki/10202.
15 Skamarock, W.C., et al. (2008) A Description of the Advanced Research WRF Version 3.
https://opensky.ucar.edU/islandora/object/technotes:500.
16USEPA (2019). Meteorological Model Performance for Annual 2016 Simulation WRF v3.8
https://www.epa.gov/sites/default/files/2020-10/documents/met_model_performance-
2016_wrf.pdf. EPA-454/R-19-010.
17 https://www.mmm.ucar.edu/weather-research-and-forecasting-model
18Byun, 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: https://www.cmascenter.org/.
19 Henderson, B., et al. (2018) Hemispheric-CMAQ Application and Evaluation for
2016, Presented at 2019 CMAS Conference, available
https://cmascenter.Org/conference//2018/slides/0850_henderson_hemispheric-
cmaq_application_2018 .pptx.
20 Mathur, R., et al. (2017) Extending the Community Multiscale Air Quality (CMAQ) modeling
system to hemispheric scales: overview of process considerations and initial applications, Atmos.
Chem. Phys., 17, 12449-12474, https://doi.org/10.5194/acp-17-12449-2017.
21 USEPA (2019). Technical Support Document: Preparation of Emissions Inventories for the
Version 7.1 2016 Hemispheric Emissions Modeling Platform. Office of Air Quality Planning and
Standards.
22 US EPA, 2021. Technical Support Document (TSD) Air Quality Modeling for the HD 2027
Proposal.
23 SMAT-CE (Software for the Modeled Attainment Test - Community Edition) vl.6:
https://www.epa.gov/scram/photochemical-modeling-tools; 2014-2018 FRM data:
ftp://newftp.epa.gov/Air/aqmg/SMAT/Ambient_Data/2018/
24U.S. EPA, 2004, Procedures for Estimating Future PM2.5 Values for the CAIR Final Rule by
Application of the (Revised) Speciated Modeled Attainment Test (SMAT)- Updated 11/8/04.
25 U.S. EPA, 2011, Final Cross State Air Pollution Rule Air Quality Modeling TSD.
26 U.S. EPA, 2018. Modeling Guidance For Demonstrating Air Quality Goals for Ozone, PM2.5,
and Regional Haze; EPA-454/R-18-009; Research Triangle Park, NC; November 2018.
27 Woods & Poole (2015). Complete Demographic Database.
323
-------
324
-------
Chapter 7 Program Costs
In this chapter, EPA presents estimates of the costs associated with the emissions-reduction
technologies that manufacturers could add in response to the final standards. We present these
not only in terms of the upfront technology costs per engine as presented in Chapter 3 of this
RIA, but also how those costs will change in the years following implementation. We also
present the costs associated with the final regulatory useful life provisions that correspond to
those standards, as well as costs associated with the warranty provisions. These technology costs
are presented in terms of direct manufacturing costs and associated indirect costs—i.e., research
and development (R&D), administrative costs, marketing, and other costs of running a company.
We term the sum of these direct and indirect costs "technology costs" or "technology package
costs." They represent the costs incurred by manufacturers—i.e., regulated entities—to comply
with the final program.A The analysis also includes estimates of the possible operating costs
associated with the final program. These operating costs represent estimated costs incurred by
users of MY 2027 and later heavy-duty vehicles.6 All costs are presented in 2017 dollars unless
noted otherwise.
The costs presented here are grouped into three main categories, as described below:
• Technology Package Costs: these are the direct costs of new or modified technology—
that EPA projects manufacturers will add—and the associated indirect costs that will
be involved with bringing those technologies to market (research, development,
warranty, etc.). In our analysis, these costs are expected to be incurred by
manufacturers of new HD engines and vehicles. "Direct" costs represent the direct
manufacturing costs of the technologies we expect to be used to comply with the final
standards over the final useful lives. We use those costs to estimate the year-over-year
manufacturing costs going forward from the first year of implementation. "Indirect"
costs include the indirect costs of the technologies we expect to be used to comply
with the final standards, in part due to the useful life provisions. Indirect costs also
include costs expected under the final program due to the warranty provisions.
• Operating Costs: these are the costs associated with the truck and bus operation that
are projected to be impacted by the final program. For example, costs associated with
tire replacement are not included since the final standards are not expected to impact
tire replacement, but costs associated with repair of the more costly emission-related
components are included. These costs are estimated to be incurred by
purchasers/owners of new MY 2027 HD vehicles.
A More precisely, these technology costs represent costs that manufacturers are expected to attempt to recapture via
new vehicle sales. As such, profits are included in the indirect cost calculation. Clearly, profits are not a "cost" of
compliance~EPA is not imposing new regulations to force manufacturers to make a profit. However, profits are
necessary for manufacturers in the heavy-duty industry, a competitive for-profit industry, to sustain their operations.
As such, manufacturers are expected to make a profit on the compliant vehicles they sell, and we include those
profits in estimating technology costs.
B Importantly, the final standards, useful lives, and warranty periods apply only to new, MY 2027 and later heavy-
duty vehicles. The legacy fleet is not subject to the new requirements and, therefore, users of prior model year
vehicles will not incur the operating costs we estimate.
325
-------
• Program Costs: these are the new technology package costs and operating costs
combined (the sum of numbers 1 and 2, above). These costs represent our best
estimate of the costs to society. As such, any taxes (e.g., fuel taxes) are excluded since
taxes represent a transfer payment from one member of society to another with no net
cost to society. Total program costs under the final program are presented in terms of
calendar year 2045 costs, present value costs, and annualized costs (see Table 7-51
and Table 7-52).c
The cost analysis is done using a tool written in Python and contained in the docket. The
Python tool along with some documentation is contained in the docket to this rule and on our
website.1
7.1 Technology Package Costs
As noted, individual technology piece costs were presented in Chapter 3. Those costs are, in
general, the direct manufacturing costs (DMC) estimated for the first year of implementation of
the final standards for the final useful lives. Those costs are used here as a starting point in
estimating program costs. Following the year in which costs are first incurred for each phase, we
have applied a learning effect to represent the cost reductions expected to occur via the "learning
by doing" phenomenon.2 This provides a year-over-year cost for each technology as applied to
new engine sales. We have then applied industry standard "retail price equivalent (RPE)" markup
factors industry-wide, with adjustments discussed below, to estimate indirect costs. Both the
learning effects applied to direct costs and the application of markup factors to estimate indirect
costs are consistent with the cost estimation approaches used in EPA's past transportation-related
regulatory programs.3 The sum of the direct and indirect costs represents our estimate of
technology costs per vehicle on a year-over-year basis. These technology costs multiplied by
estimated sales then represent the total technology costs associated with the final program.
This cost calculation approach presumes that the expected technologies will be purchased by
original equipment manufacturers (OEMs) from their suppliers. So, while the DMC estimates
include the indirect costs and profits incurred by the supplier, the indirect cost markups we apply
cover the indirect costs incurred by OEMs to incorporate the new technologies into their vehicles
and to cover profit margins typical of the heavy-duty truck industry. We discuss the indirect
costs markups in more detail in Section 7.1.2.
These technology package costs (both direct and indirect), while first incurred by
manufacturers of new engines, are presumed to be passed on to the consumers of those engines
(i.e., heavy-duty truck makers and, ultimately, their purchasers/owners).
Note that, throughout this discussion of costs we use the term regulatory class which is
roughly equivalent to a service class; we use the term regulatory class for consistency with our
MOVES model and its classification system so that our costs align with our inventory estimates
and the associated benefits discussed in Chapters 5 and 8.
c The costs presented in Table 7-98 and Table 7-99 are presented again in Table ES-2,which summarizes the net
benefits of the final standards.
326
-------
7.1.1 Direct Manufacturing Costs
To produce a unit of output, manufacturers incur direct and indirect costs. Direct costs include
cost of materials and labor costs to manufacture that unit. Indirect costs are discussed in the
following section. The direct manufacturing costs presented here include individual technology
costs for emission-related engine components and for exhaust aftertreatment systems (EAS).
Notably, for this analysis we include not only the marginal increased costs associated with the
final program, but also the emission control system costs for the "no action" baseline case (Table
7-5 and Table 7-6).D Throughout this discussion we refer to baseline technology costs, or
baseline costs, which are meant to reflect our cost estimate of engine systems—that portion that is
emission-related—and the exhaust aftertreatment costs absent the impacts of final program. This
inclusion of baseline system costs contrasts with EPA's approach in recent Greenhouse Gas rules
or the light-duty Tier 3 criteria pollutant rule where we estimated costs relative to a "no action"
baseline case, which obviated the need to estimate baseline costs. We have included baseline
costs in this analysis because the final emissions warranty and regulatory useful life provisions
are expected to have some impact on not only the new technology added to comply with the final
program, but also on any existing emission control systems (see Chapter 2 for more details on
the final Emissions Warranty and Regulatory Useful Life). The new warranty and useful life
provisions will increase costs not only for the new technology added in response to the new
standards, but also for the technology already in place (to which the new technology is added)
because the new warranty and useful life provisions will apply to the entire emission-control
system, not just the new technology added in response to the new standards. The baseline direct
manufacturing costs detailed below are thus meant to reflect that portion of baseline case engine
hardware and aftertreatment systems for which new indirect costs will be incurred due to the
final warranty and useful life provisions, even absent any changes in the level of emission
standards.
We have estimated the baseline engine costs based on recently completed studies by the
International Council on Clean Technology (ICCT) as discussed in more detail below. The
baseline EAS costs were presented in Chapter 3 of this RIA. The estimated marginal technology
costs associated with the final standards were also presented in Chapter 3 of this RIA.
As noted, the costs shown in Table 7-5 and Table 7-6 include costs for the baseline case.E For
the baseline diesel engine-related costs associated with emission control (i.e., a portion of the
fuel system, the EGR system, etc.), we have relied on a white paper done by the International
Council on Clean Transportation (ICCT) entitled, "Costs of Emission Reduction Technologies
for Heavy-Duty Diesel Vehicles."4 In Table 14 of that paper, ICCT presented technology costs
to meet U.S. standards at different stages for a 12L engine. The different stages of U.S. standards
were the 1998, 2004, 2007 and 2010 standards. Relevant portions of ICCT's Table 14—those
portions associated with engine-related technologies—are shown in Table 7-1. For the fuel
system and the turbo charger, ICCT shows only 50 percent of the total cost and states in the text
D See Chapter 5 for more information about the baseline and how that baseline is characterized. For this cost
analysis, the baseline, or no action, case consists of MY 2019 engines and emission control systems. See also
Section VI for more information about the emission inventory baseline and how that baseline is characterized. Why
we include costs for the no action case is described in this section.
E See RIA Chapters 1.1 and 1.2 for more information on emission control technologies available on current, or
baseline, engines.
327
-------
that, for components that serve other purposes in addition to emission control (e.g., the fuel
system delivers fuel to power the engine and the turbo charger serves to increase engine power in
addition to their emission control functions), only 50% of the cost is considered in their
analysis.5 ICCT notes that their costs are likely conservative since they do not consider learning
effects applied to the cost estimates associated with each regulatory stage. Lastly, ICCT notes
that their cost estimates are stated in 2015 dollars.
Table 7-1: ICCT Cost Estimates of 12L Diesel Engine-Related Emission Control Costs Associated with Past
US Emission Standards (2015 dollars)
Air/fuel control and engine out emissions
US 1998
US2004
US2007
US2010
Fuel system—50% of total cost
-
376
38
41
Variable Geometry Turbo (extra cost)—50% of total cost
-
-
185
-
Exhaust Gas Recirculation (EGR) system
-
439
-
-
EGR cooling
-
108
-
-
Total for air/fuel control and engine out emissions
-
923
223
41
In this analysis, EPA has made use of these ICCT cost estimates by first doubling the fuel
system and turbo charger costs to get the full cost of those systems (i.e., to undo the halving of
those costs done by ICCT). We then added to that result the EGR costs to get a total cost of
$1,827. We have then scaled them based on engine displacement in a manner consistent with our
approach to estimating exhaust aftertreatment costs (EAS, see Chapter 3 of this RIA). The engine
displacements used in our EAS cost estimates were 7, 8 and 13 liters for light, medium and
heavy heavy-duty engines, respectively. We have estimated the class 2b and 3 engines as
equivalent to the light heavy-duty (7L) and the urban bus engines as equivalent to the medium-
duty (8L) engines. We then adjusted the costs from the ICCT study's 2015 dollars to 2017
dollars consistent with the FRM analysis. The resultant diesel engine-related costs used in this
analysis are shown in Table 7-2.
Table 7-2: Diesel Engine-Related Emission Control System Costs in the "No Action" Baseline*
Class 2b3
Light HDE
Medium HDE
Heavy HDE
Urban bus
Engine displacement
7
7
8
13
8
Displacement based scalar
7/12=0.58
7/12=0.58
8/12=0.67
13/12=1.08
8/12=0.67
Baseline cost, 2015 dollars
(1,827 times Displacement based
scalar)
$1,066
$1,066
$1,218
$1,979
$1,218
Baseline cost, 2017 dollars
(1.03 GPD deflator Baseline cost
in 2015 dollars)*
$1,097
$1,097
$1,254
$2,038
$1,254
* See Table 7-8 and associated text for information on the GDP deflators used in this analysis; costs shown are by
MOVES regulatory class; there are a small number of diesel engines used in LHD2b3 that are engine (rather than
chassis) certified and are, therefore, expected to incur costs associated with the final rule.
For the baseline gasoline engine-related costs associated with emission control (i.e., a portion
of the fuel system, etc.), we have relied on a white paper done by ICCT entitled, "Estimated Cost
of Emission Reduction Technologies for Light-Duty Vehicles."6 In Table 4-10 of that paper,
ICCT presented technology costs to meet U.S. light-duty Tier 2, Bin 5 for a 4.5L engine. The
ICCT estimate shown was $306. ICCT notes that their cost estimates are stated in 2011 dollars.
328
-------
In this analysis, EPA has made use of this ICCT cost estimate by scaling them based on
engine displacement from the 4.5L light-duty displacement assumed by ICCT to a more typical
gasoline HD engine displacement of 7L. The resultant gasoline engine-related costs used in this
analysis are shown in Table 7-3.
Table 7-3: Gasoline Engine-Related Emission Control System Costs in the "No Action" Baseline*
Light HDE
Medium HDE
Heavy HDE
Engine displacement
7
7
7
Displacement based scalar
7/4.5=1.56
7/4.5=1.56
7/4.5=1.56
Baseline cost, 2011 dollars
(306 times Displacement based scalar)
$476
$476
$476
Baseline cost, 2017 dollars
(1.099 GPD deflator Baseline cost in 2011 dollars)*
$523
$523
$523
* See Table 7-8 and associated text for information on the GDP deflators used in this analysis; note that there are
no engine certified LHD2b3 gasoline engines and, therefore, none are expected to incur costs associated with this
final rule.
For the baseline CNG engine-related costs associated with emission control (a portion of the
fuel system, etc.), we have relied on the ICCT baseline gasoline costs presented in Table 7-3, but
have scaled those costs based on more typical diesel engine displacements because CNG engines
tend to be converted diesel engines but with fuel systems more typical of gasoline engines. The
diesel engine displacements used for scaling gasoline costs were presented in Table 7-2. The
resultant baseline engine-related CNG costs are shown in Table 7-4.
Table 7-4: CNG Engine-Related Emission Control System Costs in the "No Action" Baseline*
Heavy HDE
Urban bus
Engine displacement
12
9
Displacement based scalar
12/4.5=2.67
8/4.5=2.0
Baseline cost, 2011 dollars
(306 times Displacement based scalar)
$816
$612
Baseline cost, 2017 dollars
(1.099 GPD deflator * Baseline cost in 2011 dollars)*
$896
$672
* See Table 7-8 and associated text for information on the GDP deflators used in this
analysis.
For cylinder deactivation costs under the final standards, we have used FEV-conducted
teardown-based cylinder deactivation costs as presented in Chapter 3 of this RIA.7 The marginal
technology costs for exhaust aftertreatment components—also detailed and presented in Chapter
3 of this RIA—are updated relative to the proposal. In the proposal, we used an ICCT
methodology with extensive revision by EPA. In this final analysis, the exhaust aftertreatment
costs are based on FEV-conducted teardown-based costs.8
The cost basis (the year dollars) for many of these costs were also presented in a mixed set of
cost basis years. For the cost analysis presented here, we use 2017 dollars throughout the
analysis for consistency with the proposal which used 2017 dollars. The costs presented in
Chapter 3 are repeated in Table 7-5 for diesel regulatory classes, in Table 7-6 for gasoline
regulatory classes, and in Error! Reference source not found, for CNG regulatory classes with
329
-------
the exception that all values presented here are updated to a consistent 2017 dollar basis.F See
Chapter 5.2 for a discussion of regulatory classes. Table 7-8 shows the gross-domestic product
price deflators used to adjust to 2017 dollars. Note that we have estimated costs for regulatory
classes that exist in our MOVES runs (see Chapter 5 of this RIA) to remain consistent with the
inventory impacts we have estimated. Note also that, throughout this section, we use several
acronyms, including heavy-duty engine (HDE, exhaust aftertreatment system (EAS), and
compressed natural gas (CNG).
Table 7-5: Diesel Technology and Package Direct Manufacturing Costs per Engine by Regulatory Class for
the Baseline and Final Program, 2017 dollars
MOVES
Regulatory Class
Technology
Baseline
Final Program
(MY2027 increment to Baseline)
Package
3,681
1,920
Engine hardware
1,097
0
Class 2b3
Closed crankcase
0
0
Cylinder deactivation
0
196
EAS
2,585
1,724
Package
3,699
1,957
Engine hardware
1,097
0
Light HDE
Closed crankcase
18
37
Cylinder deactivation
0
196
EAS
2,585
1,724
Package
3,808
1,817
Engine hardware
1,254
0
Medium HDE
Closed crankcase
18
37
Cylinder deactivation
0
147
EAS
2,536
1,634
Package
5,816
2,316
Engine hardware
2,037
0
Heavy HDE
Closed crankcase
18
37
Cylinder deactivation
0
206
EAS
3,761
2,074
Package
3,884
1,850
Engine hardware
1,254
0
Urban bus
Closed crankcase
18
37
Cylinder deactivation
0
147
EAS
2,613
1,666
F The MY 2019 engine and aftertreatment costs estimates presented in RIA Chapters 3.1.5 and 3.2.3 are used as the
MY 2027 baseline cost in the tables in this RIA Chapter 7.1.1.
330
-------
Table 7-6: Gasoline Technology and Package Direct Manufacturing Costs per Engine by Regulatory Class for
the Baseline and Final Program, 2017 dollars
MOVES Regulatory Class
Technology
Baseline
Final Program
(MY2027 increment to Baseline)
Package
2,681
688
Light HDE
Engine hardware
522
0
Aftertreatment
2,158
664
ORVR
0
24
Package
2,681
688
Medium HDE
Engine hardware
522
0
Aftertreatment
2,158
664
ORVR
0
24
Package
2,681
688
Heavy HDE
Engine hardware
522
0
Aftertreatment
2,158
664
ORVR
0
24
Table 7-7: CNG Technology and Package Direct Manufacturing Costs per Engine by Regulatory Class for
the Baseline and Final Program, 2017 dollars
MOVES Regulatory Class
Technology
Baseline
Final Program
(MY2027 increment to Baseline)
Heavy HDE
Package
8,585
25
Engine hardware
896
0
Aftertreatment
7,689
25
Urban bus
Package
6,438
19
Engine hardware
672
0
Aftertreatment
5,766
19
Table 7-8: GDP Price Deflators* Used to Adjust Costs to 2017 Dollars
Cost Basis Year
Conversion Factor
2011
1.098
2015
1.029
2017
1.000
2018
0.977
2019
0.960
* Based on the National Income and Product Accounts, Table 1.1.9 Implicit Price Deflators for Gross Domestic
Product, Bureau of Economic Analysis, U.S. Department of Commerce, April 28, 2022.
The direct costs are then adjusted to account for learning effects going forward from the first
year of implementation. To make that adjustment, the following equation is used.9
/xti\b
y^={—)
Where,
yt = cost (or price) of a given item at time t
xt = cumulative production of a given item at time t
xt+i = cumulative production of a given item at time t+1
331
-------
b = the learning rate
In this cost analysis, EPA makes an adjustment to this standard formula by inserting a "seed
volume factor" meant to represent the number of years of sales of a technology leading to
learning prior to the year for which the technology's cost estimate is intended (model year 2027
as indicated in Table 7-5, Table 7-6, and Error! Reference source not found.). In other words,
manufacturers may sell some of the systems expected for compliance with the final standards in
years prior to the first year of the new standards, thereby learning and realizing cost reductions in
that time. The value of this seed volume factor might be 0 to represent a new, unsold technology,
or any value >0 to represent a relatively new technology having had some sales in prior years. A
seed volume factor of 0 places the technology at the beginning of the learning curve and,
therefore, subsequent learning effects (i.e., cost reductions) will be most rapid in the years
immediately following the first year of implementation. An increasing seed volume factor serves
to move the effects of learning further along the curve making subsequent learning effects less
dramatic in the years immediately following the first year of the analysis. Figure 7-1 shows the
effect of the seed volume factor on the levels of learning applied to the direct manufacturing
costs assuming constant sales year-over-year and a learning rate of -0.245.G
1200
o
u
k_
to
CD
>
1,000
0.800
o 0,600
cu
>
cu
cc
o
u
0,400
0,200
0.000
yearl year2 year3 year4 year5 year6 year7 yearS year9
SeedVolumeFactor=0
•SeedVolumeFactor=3
Figure 7-1: Costs Relative to First Year Costs using Different Seed Volume Factors
In the end, the learning effects are calculated using the following formula.
G In effect, the "seed volume factof' sets the cumulative number of units produced by an organization, which is a
required data point to estimate the conventional form of the learning curve. Throughout this analysis, we have used a
learning rate of -0.245 as developed for EPA by ICF and a foremost Subject Matter Expert, see "Cost Reduction
through Learning in Manufacturing Industries and in the Manufacture of Mobile Sources, Final Report and Peer
Review Report," EPA-420-R-16-018, November 2016.
332
-------
Salest=0 + (Salest=0 * SeedVolumeFactor)
xt + (Salest=0 * SeedV olumeF actor)
Where,
b = the learning rate (-0.245 in this analysis)
yt=o = estimated direct cost in the first year of implementation (e.g., MY2027 or MY2030)
Salest=o = sales in the first year of implementation (e.g., MY2027 or MY2030)
SeedVolumeFactor = 0 or greater to represent the number of years of learning already having occurred on a
technology
xt = the cumulative sales of vehicles complying with the new standard (equal to first year sales in the first
year of implementation, first year sales plus second year sales in the second year of implementation, etc.)
To illustrate the seed volume factor, if we assume the learning rate, b, is -0.245 (the value
used in this analysis), the direct cost in the first year of implementation, yt=o, is $100, the sales in
the first year of implementation, Salest=o, is 1000 engines, and the seed volume factor is 0 (i.e.,
no learning having occurred prior to the first year of the analysis), then the cost, yt, would be
$100. This is because the cumulative production in year t, xt, equals Salest=o in the first year of
implementation leaving the formula as:
If the sales in the following year were an additional 1000 engines, the cost would decrease to
$84 since xt would now be 2000 and the formula would be:
If the seed volume factor is set to 3 (i.e., to approximate 3 years of learning prior to the first
year of the analysis, then the cost in 2027 would again be $100 since the formula would be
With 3 years of learning estimated to have already occurred, the costs going forward from the
first year reduce at a slower pace than in the previous example. Therefore, increasing the seed
volume factor results in less rapid learning effects going forward. The seed volume factors used
in this analysis are shown in Table 7-9. A factor of 10 has been used for the baseline
technologies since those technologies will have undergone considerable learning by MY 2027.
We have used a factor of 3 for the final standards to reflect at least 3 years of sales with
technologies very similar to those expected under the final standards thereby resulting in
conservative learning-based cost reductions moving forward from MY 2027.
(1)— 0.245 *$100 =$100
* $100 = $84
However, in 2028, the formula would be
333
-------
Table 7-9: Seed volume factors used in this analysis
Fuel
Regulatory Class
Baseline
Final Program
(MY 2027 increment to Baseline)
Class 2b3
10
3
Light HDE
10
3
Diesel
Medium HDE
10
3
Heavy HDE
10
3
Urban Bus
10
3
Light HDE
10
3
Gasoline
Medium HDE
10
3
Heavy HDE
10
3
CNG
Heavy HDE
10
3
Urban Bus
10
3
Learning factors were applied on a technology package cost basis, and MOVES projected
sales volumes were used to determine first year sales (Salest=o) and cumulative sales (xt). The
resultant direct manufacturing costs and how those costs reduce over time are presented in both
Chapter 7.1.3 (on a per vehicle basis) and in Chapter 7.3.1 (on a total cost basis). The resultant
learning effects are shown in Table 7-10 for diesel HDE, Table 7-11 for gasoline HDE, and
Table 7-12 for CNG HDE.
Table 7-10: Learning Effects Applied to Direct Manufacturing Costs for Diesel HDE
Model Year
Class 2b3
Light HDE
Medium HDE
Heavy HDE
Urban bus
Baseline
FRM
Baseline
FRM
Baseline
FRM
Baseline
FRM
Baseline
FRM
2027
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
2028
0.979
0.947
0.979
0.946
0.979
0.946
0.979
0.947
0.979
0.947
2029
0.960
0.906
0.959
0.904
0.959
0.904
0.960
0.906
0.960
0.905
2030
0.943
0.873
0.941
0.869
0.942
0.870
0.942
0.872
0.942
0.871
2031
0.927
0.845
0.924
0.840
0.925
0.840
0.926
0.843
0.926
0.842
2032
0.913
0.821
0.909
0.814
0.909
0.815
0.911
0.818
0.910
0.817
2033
0.900
0.800
0.894
0.792
0.895
0.793
0.897
0.796
0.896
0.795
2034
0.887
0.781
0.880
0.771
0.881
0.773
0.884
0.777
0.883
0.775
2035
0.875
0.765
0.867
0.753
0.868
0.755
0.871
0.760
0.870
0.758
2036
0.864
0.750
0.854
0.737
0.856
0.738
0.860
0.744
0.858
0.742
2037
0.854
0.736
0.843
0.722
0.844
0.724
0.849
0.730
0.847
0.727
2038
0.844
0.724
0.831
0.708
0.833
0.710
0.838
0.716
0.836
0.714
2039
0.835
0.712
0.821
0.695
0.822
0.697
0.828
0.704
0.826
0.701
2040
0.826
0.701
0.810
0.683
0.812
0.685
0.819
0.693
0.816
0.690
2041
0.817
0.691
0.800
0.672
0.802
0.674
0.809
0.682
0.807
0.679
2042
0.809
0.682
0.791
0.661
0.793
0.664
0.801
0.672
0.798
0.669
2043
0.801
0.673
0.782
0.651
0.784
0.654
0.792
0.663
0.789
0.659
2044
0.794
0.664
0.773
0.642
0.776
0.645
0.784
0.654
0.781
0.650
2045
0.787
0.656
0.765
0.633
0.767
0.636
0.776
0.645
0.773
0.641
334
-------
Table 7-11: Learning Effects Applied to Direct Manufacturing Costs for Gasoline HDE
Model Year
Light HDE
Medium HDE
Heavy HDE
Baseline
FRM
Baseline
FRM
Baseline
FRM
2027
1.000
1.000
1.000
1.000
1.000
1.000
2028
0.979
0.946
0.979
0.946
0.979
0.946
2029
0.959
0.904
0.959
0.904
0.959
0.904
2030
0.941
0.869
0.941
0.869
0.941
0.869
2031
0.924
0.840
0.924
0.840
0.924
0.840
2032
0.909
0.814
0.909
0.814
0.909
0.814
2033
0.894
0.792
0.894
0.792
0.894
0.792
2034
0.880
0.772
0.880
0.771
0.880
0.771
2035
0.867
0.753
0.867
0.753
0.867
0.753
2036
0.855
0.737
0.854
0.737
0.854
0.737
2037
0.843
0.722
0.843
0.722
0.843
0.722
2038
0.832
0.708
0.831
0.708
0.831
0.708
2039
0.821
0.695
0.820
0.695
0.820
0.695
2040
0.811
0.683
0.810
0.683
0.810
0.683
2041
0.801
0.672
0.800
0.672
0.800
0.672
2042
0.791
0.662
0.791
0.661
0.791
0.661
2043
0.782
0.652
0.782
0.651
0.782
0.651
2044
0.774
0.642
0.773
0.642
0.773
0.642
2045
0.765
0.633
0.765
0.633
0.765
0.633
Table 7-12: Learning Effects Applied to Direct Manufacturing Costs for CNG HDE
Model Year
Heavy HDE
Urban bus
Baseline
FRM
Baseline
FRM
2027
1.000
1.000
1.000
1.000
2028
0.979
0.947
0.979
0.947
2029
0.960
0.905
0.960
0.905
2030
0.942
0.870
0.942
0.871
2031
0.925
0.841
0.926
0.842
2032
0.910
0.816
0.910
0.817
2033
0.895
0.794
0.896
0.795
2034
0.882
0.774
0.883
0.775
2035
0.869
0.756
0.870
0.758
2036
0.857
0.740
0.858
0.742
2037
0.845
0.725
0.847
0.727
2038
0.834
0.711
0.836
0.714
2039
0.824
0.699
0.826
0.701
2040
0.814
0.687
0.816
0.690
2041
0.804
0.676
0.807
0.679
2042
0.795
0.666
0.798
0.669
2043
0.786
0.656
0.789
0.659
2044
0.778
0.647
0.781
0.650
2045
0.770
0.638
0.773
0.641
7.1.2 Indirect Costs
The indirect costs presented here are all the costs estimated to be incurred by manufacturers of
new heavy-duty engines and vehicles associated with producing the unit of output that are not
335
-------
direct costs. For example, they 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 good sold. Although it is
possible to account for direct costs allocated to each unit of goods sold, it is more challenging to
account for indirect costs allocated to a unit of goods sold. To ensure that regulatory analysis
capture the changes in indirect costs, markup factors, which relate total indirect costs to total
direct costs, have been developed and used by EPA and other stakeholders. These factors are
often referred to as retail price equivalent (RPE) multipliers. RPE multipliers provide, at an
aggregate level, the relative shares of revenues, where:
Revenue = Direct Costs + Indirect Costs
so that:
Revenue/Direct Costs = 1 + Indirect Costs/Direct Costs = RPE
and,
Indirect Costs = Direct Costs x (RPE - 1).
If the relationship between revenues and direct costs (i.e., RPE) can be shown to equal an
average value over time, then an estimate of direct costs can be multiplied by that average value
to estimate revenues, or total costs. Further, that difference between estimated revenues, or total
costs, and estimated direct costs can be taken as the indirect costs. EPA has frequently used these
multipliers10 to predict the resultant impact on costs associated with manufacturers' responses to
regulatory requirements and we are using that approach in this analysis.
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. In the past, EPA has expressed a concern with using the RPE multiplier for all
technologies, because it is not likely that the indirect costs of vehicle modifications are the same
for all technologies.11 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. EPA developed an alternative indirect cost methodology—the Indirect
Cost Multiplier (ICM)—to address those concerns.12
The cost of different technologies was an important distinction in modeling efforts supporting
EPA's greenhouse gas rulemakings (GHG) since a variety of GHG technologies were available
to manufacturers and EPA wanted to project which combinations of technologies were most cost
effective toward achieving compliance. For this final rule, we do not have that consideration
since we are projecting the same technologies for all vehicles within a given regulatory class and
are not attempting to project from a variety of technologies which are most cost effective toward
achieving compliance. For that reason, EPA is using the RPE approach to estimate indirect costs
in this analysis.
336
-------
RPEs themselves are also inherently difficult to estimate because the accounting statements of
manufacturers do not neatly categorize all cost elements as either direct or indirect costs. Hence,
each researcher developing an RPE estimate must apply a certain amount of judgment to the
allocation of the costs.13
This cost analysis estimates indirect costs by applying markup factors used in past
rulemakings setting new greenhouse gas standards for heavy-duty trucks.14 The markup factors
are based on financial filings with the Securities and Exchange Commission for several engine
and engine/truck manufacturers in the heavy-duty industry as detailed in a study done by RTI
International for EPA.15 The RPE factors developed by RTI for HD engine manufacturers, HD
truck manufacturers and for the HD truck industry as a whole are shown in Table 7-13.H Also
shown in Table 7-13 are the RPE factors developed by RTI for light-duty vehicle
manufacturers.16
Table 7-13: Retail Price Equivalent Factors in the Heavy-Duty and Light-Duty Industries
Cost Contributor
HD Engine
HD Truck
HD Truck
LD Vehicle
Manufacturer
Manufacturer
Industry
Industry
Direct manufacturing cost
1.00
1.00
1.00
1.00
Warranty
0.02
0.04
0.03
0.03
R&D
0.04
0.05
0.05
0.05
Other (admin, retirement, health, etc.)
0.17
0.22
0.29
0.36
Profit (cost of capital)
0.05
0.05
0.05
0.06
RPE
1.28
1.36
1.42
1.50
For this analysis, EPA has based cost estimates for diesel and CNG regulatory classes on the
HD Truck Industry values shown in Table 7-13.1 Because most of the changes apply to engines,
we first considered using the HD Engine Manufacturer values. However, the industry is
becoming more vertically integrated and the costs we are trying to estimate are those that occur
at the end purchaser, or retail, level. For that reason, we believe that the truck industry values
represent the factors of most interest to this analysis. For gasoline regulatory classes, we have
used the LD Vehicle Industry values shown in Table 7-13. We have chosen those values since
they more closely represent the cost structure of manufacturers in that industry—Ford, General
Motors, and Fiat Chrysler.
Of the cost contributors listed in Table 7-13, Warranty and R&D are the elements of indirect
costs that the final requirements are expected to impact. As discussed in Chapter 2 of this RIA,
EPA is lengthening the required warranty period, which we expect to increase the contribution of
warranty costs to indirect costs. EPA is also lengthening the regulatory useful life, which we
expect to result in increased R&D expenses as systems are developed to deal with the longer life
H The engine manufacturers included were Hino and Cummins; the truck manufacturers included were PACCAR,
Navistar, Daimler and Volvo. Where gaps existed such as specific line items not reported by these companies due to
differing accounting practices, data from the Heavy Duty Truck Manufacturers Industry Report by Supplier
Relations LLC (2009) and Census (2009) data for Other Engine Equipment Manufacturing Industry (NAICS
333618) and Heavy Duty Truck Manufacturing Industry (NAICS 336120) were used to fill the gaps. This is detailed
in the study report at Appendix A. 1.
1 Note that the report used the term "HD Truck" while EPA generally uses the term "HD vehicle;" they are
equivalent when referring to this report.
337
-------
during which compliance with standards will be required. We expect that the minor OBD-related
R&D efforts as discussed in Section IV.C of the preamble will be done given the R&D estimated
here. Profit is listed to highlight that profit is being considered and included in the analysis. All
other indirect cost elements—those encapsulated by the "Other" category, including General and
Administrative Costs, Retirement Costs, Healthcare Costs, and other overhead costs—as well as
Profits, are expected to scale according to their historic levels of contribution.
As mentioned, Warranty and R&D are the elements of indirect costs that the final
requirements are expected to impact. Warranty expenses are the costs that a business expects to
incur, or has already incurred, for the repair or replacement of goods that it has sold. The total
amount of warranty expense is limited by the warranty period that a business typically allows.
After the warranty period for a product has expired, a business no longer incurs a warranty
liability; thus, a longer warranty period results in a longer period of liability for a product. At the
time of sale, companies are expected to set aside money in a warranty liability account to cover
any potential future warranty claims. If and when warranty claims are made by customers, the
warranty liability account is debited and a warranty claims account is credited to cover warranty
claim expenses.
In the proposed analysis, to address the expected increased indirect cost contributions
associated with warranty (increased funding of the warranty liability account) due to the
proposed longer warranty requirements, we applied scaling factors commensurate with the
changes in proposed Option 1 or Option 2 to the number of miles included in the warranty period
(i.e., VMT-based scaling factors). Industry commenters took exception to this approach, arguing
that it resulted in underestimated costs associated with warranty. To support their comments, the
Truck and Engine Manufacturers Association (EMA) submitted data that showed costs
associated with actual warranty claims for roughly 250,000 heavy heavy-duty vehicles. The chart
included in the EMA comments is shown in Figure 7-2 and is in the public docket for this rule.
338
-------
< e
o n
$15,000
Mull HHOCK mudrt Mnadf iM. 1 Of Mi (motflfy lyr/VX*
3
g
<
0 2 4 6 8 10
Total Warranty Coverage tn Year*
Figure 7-2: Warranty Costs Submitted as Part of the Comments from the Truck and Engine Manufacturers
Association; see EPA-HQ-OAR-2019-0055-1203-A1, page 151
EPA considers this EMA comment and supporting information to be persuasive, not only
because it represents data, but also because it represents data from three manufacturers and over
250,000 vehicles. However, the data are for heavy HDE, so it is not possible to determine an
appropriate cost per year for light or medium HDE from the data directly. Also, the data
represent actual warranty claims without any mention of the warranty claims rate (i.e., the share
of engines sold that are making the warranty claims represented in the data). This latter issue
makes it difficult to determine the costs that might be imposed on all new engines sold to cover
the future warranty claims for the relatively smaller fraction of engines that incur warranty
repair. In other words, if all heavy HDE purchases are helping to fund a warranty liability
account, it is unclear if the $1,000 per year per engine is the right amount or if $1,000 per year is
needed on only that percent of engines that will incur warranty repair. In the end, warranty costs
imposed on new engine sales should be largely recouped by purchasers of those engines in the
form of reduced emission repair expenses. EPA believes it is highly unlikely that any
manufacturer would use their warranty program as a profit generator under the $ 1,000 per engine
approach, especially in a market as competitive as the HD engine and vehicle industry. The
possibility exists that the costs associated with the longer warranty coverage required by this rule
will (1) converge towards those of the better performing OEMs; and (2) drop over time via
something analogous to the learning by doing phenomenon described earlier. If true, we have
probably overestimated the costs estimated here as attributable to this nile.
Thus, after careful consideration of these comments regarding warranty, and the engineering
judgement of EPA subject matter experts, we revised our approach to estimating warranty costs,
and for the final rule we have estimated warranty costs assuming a cost of $1,000 (2018 dollars
or $977 in 2017 dollars) per estimated number of years of warranty coverage for a heavy heavy-
339
-------
duty diesel engine or heavy-duty vehicle equipped with such an engine. For other regulatory
(engine) classes, we have scaled that value by the ratio of their estimated baseline emission-
control system direct cost to the estimated emission-control system direct cost of the baseline
heavy heavy-duty diesel engine. We use the baseline heavy heavy-duty diesel engine direct cost
here because it should be consistent with the data behind the $1,000 per year value. The resulting
warranty costs per year for a MY2027 HDE are as shown in Table 7-14. Importantly, these are
emission-related warranty costs.
Table 7-14: Warranty Costs per Year of Estimated Warranty Coverage (2017 dollars)*
MOVES Regulatory Class
Scaling Approach
Diesel
Gasoline
CNG
Class 2b3
Base 2b3 DMC /
Base Diesel Heavy HDE DMC
618
Light HDE
Base Light HDE DMC /
Base Diesel Heavy HDE DMC
621
450
Medium HDE
Base Medium HDE DMC /
Base Diesel Heavy HDE DMC
639
449
Heavy HDE
Base Heavy HDE DMC /
Base Diesel Heavy HDE DMC
977
448
1,442
Urban bus
Base Urban bus DMC /
Base Diesel Heavy HDE DMC
652
1,081
* The Base Diesel HDE DMC would be the $5,816 value shown in Table 7-5.
As noted, we have used the estimated number of years of warranty coverage, not the regulated
number of years. In other words, a long-haul tractor accumulating over 100,000 miles per year
will reach any regulated warranty mileage prior to a refuse truck accumulating under 40,000
miles per year, assuming both are in the same regulatory class and, therefore, have the same
warranty provisions. In all cases, we estimate the number of years of warranty coverage by
determining the minimum number of years to reach either the regulated number of years, the
regulated number of miles, or the regulated number of hours of operation. In making this
estimation, whether for warranty or for useful life, we start with the required age/miles/hours and
a typical number of miles driven per year. The typical miles driven is calculated as the average
number of miles driven during the first 7 years of operation, according to our MOVES model.
Using that value, and the average speeds for each vehicle according to our MOVES model, we
can calculate the age at which the required miles and required hours (if applicable) will be
reached. The ages at which warranty and useful life are estimated to be reached are then
determined as the minimum of the required age, the calculated age based on miles per year, and
the calculated age based on hours per year (if applicable). The results for both warranty and
useful life are shown in Table 7-15.
340
-------
Table 7-15: Ages when Warranty and Useful Life are Estimated to be Reached, MY2027 Diesel HD Vehicles
Baseline
FRM Control
Warranty
Useful Life
Warranty
Useful Life
Class 2b3
Light Commercial Trucks
3.3
7.2
10.0
15.0
Long-Haul Single Unit Trucks
1.4
3.2
6.0
7.7
Passenger Trucks
3.3
7.2
10.0
15.0
Short-Haul Single Unit Trucks
2.3
5.0
9.5
12.2
Light HDE
Long-Haul Single Unit Trucks
1.4
3.2
6.0
7.7
Other Buses
1.3
2.9
5.5
7.1
School Buses
3.8
8.5
10.0
15.0
Short-Haul Single Unit Trucks
2.3
5.0
9.5
12.2
Transit Buses
1.3
2.9
5.5
7.1
Medium HDE
Long-Haul Single Unit Trucks
3.6
5.3
8.0
10.0
Motor Homes
5.0
10.0
10.0
12.0
Other Buses
3.3
4.9
7.4
9.2
Refuse Trucks
4.3
6.3
9.6
12.0
School Buses
5.0
10.0
10.0
12.0
Short-Haul Combination Trucks
1.7
2.6
3.9
4.9
Short-Haul Single Unit Trucks
5.0
8.4
10.0
12.0
Transit Buses
3.3
4.9
7.4
9.2
Heavy HDE
Long-Haul Combination Trucks
1.8
3.1
3.2
4.7
Long-Haul Single Unit Trucks
5.0
10.0
10.0
11.0
Motor Homes
5.0
10.0
10.0
11.0
Other Buses
5.0
10.0
10.0
11.0
Refuse Trucks
5.0
10.0
10.0
11.0
School Buses
5.0
10.0
10.0
11.0
Short-Haul Combination Trucks
3.5
6.0
6.2
9.0
Short-Haul Single Unit Trucks
5.0
10.0
10.0
11.0
Urban Bus
Transit Buses
2.6
10.0
10.0
11.0
341
-------
Table 7-16: Ages when Warranty and Useful Life are Estimated to be Reached, MY2027 Gasoline HD
Vehicles
Baseline
FRM Control
Warranty
Useful Life
Warranty
Useful Life
Light HDE
Long-Haul Single Unit Trucks
1.4
3.2
4.6
5.7
Motor Homes
5.0
10.0
10.0
15.0
Other Buses
1.3
2.9
4.2
5.3
School Buses
3.8
8.5
10.0
15.0
Short-Haul Single Unit Trucks
2.3
5.0
7.3
9.1
Transit Buses
1.3
2.9
4.2
5.3
Medium HDE
Long-Haul Single Unit Trucks
1.4
3.2
4.6
5.7
Motor Homes
5.0
10.0
10.0
15.0
Short-Haul Single Unit Trucks
2.3
5.0
7.3
9.1
Heavy HDE
Long-Haul Single Unit Trucks
1.4
3.2
4.6
5.7
Motor Homes
5.0
10.0
10.0
15.0
Short-Haul Single Unit Trucks
2.3
5.0
7.3
9.1
Table 7-17: Ages when Warranty and Useful Life are Estimated to be Reached, MY2027 CNG HD Vehicles
Baseline
FRM Control
Warranty
Useful Life
Warranty
Useful Life
Heavy HDE
Long-Haul Single Unit Trucks
5.0
10.0
10.0
11.0
Other Buses
5.0
10.0
10.0
11.0
Refuse Trucks
5.0
10.0
10.0
11.0
School Buses
5.0
10.0
10.0
11.0
Short-Haul Combination Trucks
3.5
6.0
6.2
9.0
Short-Haul Single Unit Trucks
5.0
10.0
10.0
11.0
Urban Bus
Transit Buses
2.6
10.0
10.0
11.0
Lastly, with respect to warranty, we have estimated that many of the regulated products are
sold today with a warranty period longer than the required warranty period. In the proposal, we
calculated baseline warranty costs only for the required warranty periods. In the final analysis,
we calculate baseline warranty costs for the warranty periods with which most are actually sold.
For diesel and CNG heavy HDE, we assume all are sold with warranties covering 250,000 miles,
and for diesel and CNG medium HDE, we assume half are sold with warranties covering
150,000 miles. For all other engines and associated fuel types, we have not estimated any use of
extended warranties in the baseline.
We carry these annual warranty costs for both the baseline and the final standards despite the
addition of new technology. We believe this is reasonable for two reasons: (1) the source data
mentioned above included several years of data during which there must have been new
technology introductions, yet annual costs appear to have remained generally steady; and, (2) the
R&D we expect to be done, discussed next, is expected to improve overall durability, which
should serve to help maintain historical annual costs.
342
-------
For R&D, we have maintained the approach used in the proposal, although it is applied using
the final useful life provisions. For R&D on a Class 8 truck, the final standards would extend
regulatory useful life from 10 years, 22,000 hours, or 435,000 miles to 11 years, 32,000 hours, or
650,000 miles. We have applied a scaling factor of 1.49 (650/435) to the 0.05 R&D contribution
factor for MYs 2027 and later. We apply this same methodology to estimating R&D for other
vehicle categories. We estimate that once the development efforts into longer useful life are
complete, increased expenditures will return to their normal levels of contribution. Therefore,
we have implemented R&D scalars for three years (MY 2027 through MY 2029). In MY 2030
and later, the R&D scaling factors are no longer applied.
The VMT-based scaling factors applied to R&D cost contributors used in our cost analysis of
final standards are shown in Table 7-18 for diesel and CNG regulatory classes and in Table 7-19
for gasoline regulatory classes.
Table 7-18: Scaling Factors Applied to RPE R&D Contribution Factors to Reflect Changes in their
Contributions, Diesel & CNG Regulatory Classes
Scenario
MOVES Regulatory Class
MY2027 through MY2029
MY2030+
Baseline
Class 2b3
1.00
1.00
Light HDE
1.00
1.00
Medium HDE
1.00
1.00
Heavy HDE
1.00
1.00
Urban Bus
1.00
1.00
Final Program
Class 2b3
2.45
1.00
Light HDE
2.45
1.00
Medium HDE
1.89
1.00
Heavy HDE
1.49
1.00
Urban Bus
1.49
1.00
Table 7-19: Scaling Factors Applied to RPE Contribution Factors to Reflect Changes in their Contributions,
Gasoline Regulatory Classes
Scenario
MOVES Regulatory Class
MY2027 through MY2029
MY2030+
Baseline
Light HDE
1.00
1.00
Medium HDE
1.00
1.00
Heavy HDE
1.00
1.00
Final Program
Light HDE
1.82
1.00
Medium HDE
1.82
1.00
Heavy HDE
1.82
1.00
Lastly, as mentioned in section 7.1.1, the markups for estimating indirect costs are applied to
our estimates of the absolute direct manufacturing costs for emission-control technology shown
in Table 7-5, Table 7-6 and Table 7-7, not just the incremental costs associated with the final
program (i.e., the Baseline+Final costs).J Table 7-20 provides an illustrative example using a
baseline technology cost of $5000, an incremental cost of $1000, and an indirect cost R&D
contribution of 0.05 with a simple scalar of 1.5 associated with a longer useful life period. In this
1 Increased indirect costs are included for the baseline technology because the final warranty and useful life
provisions will impact those technologies too, not just the new, incremental technology.
343
-------
case, the costs could be calculated according to two approaches as shown. By including the
baseline costs, we are estimating considerable new R&D costs in the final analysis as illustrated
by the example where including baseline costs results in R&D costs of $450 while excluding
baseline costs results in R&D costs of just $75.K
Table 7-20: Simplified Example of Indirect Warranty Costs Calculated on an Incremental vs. Absolute
Technology Package Cost (values are not from the analysis and are for presentation only)
Using Baseline Costs Only
Using Absolute Costs
Baseline direct manufacturing cost (DMC)
$5000
$5000
Action case DMC
$1000
$5000+ $1000 = $6000
Indirect R&D Costs
$1000x0.05 x 1.5 = $75
$6000x0.05x1.5 = $450
Incremental DMC + R&D
$1000+ $75 = $1075
$1000+ $450 =$1450
7,1.3 Technology Costs per Vehicle
The following tables present the technology costs estimated for the final program on a per-
vehicle basis for MY 2027. Reflected in these tables are learning effects on direct manufacturing
costs and scaling effects associated with final program requirements. The sum is also shown and
reflects the direct plus indirect cost per vehicle in the specific model year where direct costs refer
to estimated direct manufacturing costs and indirect costs refer to estimated costs such as
research and development, warranty, and administrative costs incurred by manufacturers in
achieving compliance. Note that the indirect costs shown include warranty, R&D, "other," and
profit, the latter two which scale with direct costs via the indirect cost contribution factor. While
direct costs do not change across the different vehicle types (i.e., long-haul versus short-haul
combination), the indirect costs do vary because differing miles driven and operating hours
between types of vehicles result in different warranty and useful life estimates in actual use.
These differences impact the estimated warranty and R&D costs.
Note that, while we show costs per vehicle here, it is important to remember that these are
costs and not prices. We make no effort at estimating how manufacturers will price their
products. Manufacturers may pass costs along to purchasers via price increases in a manner
consistent with what we show here. However, manufacturers may also price certain products
higher than what we show while pricing others lower—the higher-priced products thereby
subsidizing the lower-priced products. This is true in any market, not just the heavy-duty
highway industry. This is perhaps especially true with respect to the indirect costs we have
estimated because, for example, R&D done to improve emission durability can readily transfer
across different engines but the hardware added to an engine is uniquely tied to that engine.
Importantly, we present costs here for MY 2027 vehicles, but these costs continue for every
model year going forward from there. Consistent with the learning impacts described in section
7.1.1, the costs per vehicle decrease slightly over time, but only the increased R&D costs are
expected to decrease significantly. Increased R&D is estimated to occur for three years following
K As noted earlier, we have included baseline costs in this analysis because the final emissions warranty and
regulatory useful life provisions will be expected to have some impact on not only the new technology added to
comply with the final program, but also on any existing emission control systems (See Chapter 2 for more details on
proposed Emissions Warranty and Regulatory Useful Life).
344
-------
and including MY 2027, after which time its contribution to indirect costs is as shown in Table
7-13.
Table 7-21: MY2027 Diesel Class 2b3 Technology Costs per Vehicle Associated with the Final Program, 2017
dollars
Direct Costs
Indirect Costs
Costs per Vehicle
FRM Baseline
Light Commercial Trucks
3,681
3,448
7,130
Long-Haul Single Unit Trucks
3,681
2,321
6,003
Passenger Trucks
3,681
3,448
7,130
Short-Haul Single Unit Trucks
3,681
2,837
6,518
FRM Baseline+Final Program
Light Commercial Trucks
5,601
8,775
14,376
Long-Haul Single Unit Trucks
5,601
6,310
11,912
Passenger Trucks
5,601
8,775
14,376
Short-Haul Single Unit Trucks
5,601
8,477
14,078
Increased Cost of the Final Program
Light Commercial Trucks
1,920
5,326
7,246
Long-Haul Single Unit Trucks
1,920
3,989
5,909
Passenger Trucks
1,920
5,326
7,246
Short-Haul Single Unit Trucks
1,920
5,640
7,560
Table 7-22: MY2027 Diesel Light HDE Technology Costs per Vehicle Associated with the Final Program,
2017 dollars
Direct Costs
Indirect Costs
Costs per Vehicle
FRM Baseline
Long-Haul Single Unit Trucks
3,699
2,332
6,031
Other Buses
3,699
2,263
5,962
School Buses
3,699
3,829
7,528
Short-Haul Single Unit Trucks
3,699
2,851
6,550
Transit Buses
3,699
2,263
5,962
FRM Baseline + Final Program
Long-Haul Single Unit Trucks
5,656
6,353
12,009
Other Buses
5,656
6,064
11,720
School Buses
5,656
8,830
14,485
Short-Haul Single Unit Trucks
5,656
8,530
14,186
Transit Buses
5,656
6,064
11,720
Increased Cost of the Final Program
Long-Haul Single Unit Trucks
1,957
4,021
5,978
Other Buses
1,957
3,800
5,757
School Buses
1,957
5,001
6,957
Short-Haul Single Unit Trucks
1,957
5,680
7,636
Transit Buses
1,957
3,800
5,757
345
-------
Table 7-23: MY2027 Diesel Medium HDE Technology Costs per Vehicle Associated with the Final Program,
2017 dollars
Direct Costs
Indirect Costs
Costs per Vehicle
FRM Baseline
Long-Haul Single Unit Trucks
3,808
3,774
7,582
Motor Homes
3,808
4,682
8,490
Other Buses
3,808
3,597
7,404
Refuse Trucks
3,808
4,217
8,025
School Buses
3,808
4,682
8,490
Short-Haul Combination Trucks
3,808
2,595
6,402
Short-Haul Single Unit Trucks
3,808
4,682
8,490
Transit Buses
3,808
3,597
7,404
FRM Baseline + Final Program
Long-Haul Single Unit Trucks
5,625
7,572
13,197
Motor Homes
5,625
8,839
14,464
Other Buses
5,625
7,175
12,799
Refuse Trucks
5,625
8,564
14,189
School Buses
5,625
8,839
14,464
Short-Haul Combination Trucks
5,625
4,930
10,555
Short-Haul Single Unit Trucks
5,625
8,839
14,464
Transit Buses
5,625
7,175
12,799
Increased Cost of the Final Program
Long-Haul Single Unit Trucks
1,817
3,798
5,615
Motor Homes
1,817
4,157
5,974
Other Buses
1,817
3,578
5,395
Refuse Trucks
1,817
4,347
6,164
School Buses
1,817
4,157
5,974
Short-Haul Combination Trucks
1,817
2,335
4,153
Short-Haul Single Unit Trucks
1,817
4,157
5,974
Transit Buses
1,817
3,578
5,395
346
-------
Table 7-24: MY2027 Diesel Heavy HDE Technology Costs per Vehicle Associated with the Final Program,
2017 dollars
Direct Costs
Indirect Costs
Costs per Vehicle
FRM Baseline
Long-Haul Combination Trucks
5,816
4,025
9,841
Long-Haul Single Unit Trucks
5,816
7,151
12,967
Motor Homes
5,816
7,151
12,967
Other Buses
5,816
7,151
12,967
Refuse Trucks
5,816
7,151
12,967
School Buses
5,816
7,151
12,967
Short-Haul Combination Trucks
5,816
5,658
11,473
Short-Haul Single Unit Trucks
5,816
7,151
12,967
FRM Baseline + Final Program
Long-Haul Combination Trucks
8,132
6,535
14,667
Long-Haul Single Unit Trucks
8,132
13,139
21,271
Motor Homes
8,132
13,139
21,271
Other Buses
8,132
13,139
21,271
Refuse Trucks
8,132
13,139
21,271
School Buses
8,132
13,139
21,271
Short-Haul Combination Trucks
8,132
9,474
17,606
Short-Haul Single Unit Trucks
8,132
13,139
21,271
Increased Cost of the Final Program
Long-Haul Combination Trucks
2,316
2,510
4,827
Long-Haul Single Unit Trucks
2,316
5,988
8,304
Motor Homes
2,316
5,988
8,304
Other Buses
2,316
5,988
8,304
Refuse Trucks
2,316
5,988
8,304
School Buses
2,316
5,988
8,304
Short-Haul Combination Trucks
2,316
3,816
6,132
Short-Haul Single Unit Trucks
2,316
5,988
8,304
Table 7-25: MY2027 Diesel Urban Bus Technology Costs per Vehicle Associated with the Final Program,
2017 dollars
Direct Costs
Indirect Costs
Costs per Vehicle
FRM Baseline
3,884
3,238
7,122
FRM Baseline+Final Program
5,734
8,901
14,635
Increased Cost of the Final Program
1,850
5,663
7,512
347
-------
Table 7-26: MY2027 Gasoline HDE Technology Costs per Vehicle Associated with the Final Program, 2017
dollars
Direct Costs
Indirect Costs
Costs per Vehicle
FRM Baseline
Long-Haul Single Unit Trucks
2,681
1,905
4,585
Motor Homes
2,681
3,511
6,192
Other Buses
2,681
1,855
4,535
School Buses
2,681
2,989
5,670
Short-Haul Single Unit Trucks
2,681
2,280
4,961
Transit Buses
2,681
1,855
4,535
FRM Baseline+Final Program
Long-Haul Single Unit Trucks
3,369
3,784
7,153
Motor Homes
3,369
6,223
9,592
Other Buses
3,369
3,624
6,993
School Buses
3,369
6,223
9,592
Short-Haul Single Unit Trucks
3,369
4,986
8,355
Transit Buses
3,369
3,624
6,993
Increased Cost of the Final Program
Long-Haul Single Unit Trucks
688
1,880
2,568
Motor Homes
688
2,712
3,401
Other Buses
688
1,770
2,458
School Buses
688
3,234
3,923
Short-Haul Single Unit Trucks
688
2,706
3,394
Transit Buses
688
1,770
2,458
Table 7-27: MY2027 CNG Heavy HDE Technology Costs per Vehicle Associated with the Final Program,
2017 dollars
Direct Costs
Indirect Costs
Costs per Vehicle
FRM Baseline
Long-Haul Single Unit Trucks
8,585
10,556
19,141
Other Buses
8,585
10,556
19,141
Refuse Trucks
8,585
10,556
19,141
School Buses
8,585
10,556
19,141
Short-Haul Combination Trucks
8,585
8,351
16,936
Short-Haul Single Unit Trucks
8,585
10,556
19,141
FRM Baseline+Final Program
Long-Haul Single Unit Trucks
8,610
17,988
26,598
Other Buses
8,610
17,988
26,598
Refuse Trucks
8,610
17,988
26,598
School Buses
8,610
17,988
26,598
Short-Haul Combination Trucks
8,610
12,577
21,187
Short-Haul Single Unit Trucks
8,610
17,988
26,598
Increased Cost of the Final Program
Long-Haul Single Unit Trucks
25
7,431
7,457
Other Buses
25
7,431
7,457
Refuse Trucks
25
7,431
7,457
School Buses
25
7,431
7,457
Short-Haul Combination Trucks
25
4,225
4,251
Short-Haul Single Unit Trucks
25
7,431
7,457
348
-------
Table 7-28: MY2027 CNG Urban Bus Technology Costs per Vehicle Associated with the Final Program, 2017
dollars
Direct Costs
Indirect Costs
Costs per Vehicle
FRM Baseline
6,438
5,367
11,806
FRM Baseline+Final Program
6,457
13,490
19,948
Increased Cost of the Final Program
19
8,123
8,142
7.2 Operating Costs
We have estimated three impacts on operating costs expected to be incurred by users of new
MY 2027 and later heavy-duty vehicles: increased diesel exhaust fluid (DEF) consumption by
diesel vehicles due to increased DEF dose rates to enable compliance with more stringent NOx
standards; decreased fuel costs by gasoline vehicles due to new onboard refueling vapor recovery
systems that allow burning (in engine) of otherwise evaporated hydrocarbon emissions; and,
emission repair impacts. For the repair impacts, we expect that the longer duration warranty
period will result in lower owner/operator-incurred repair costs due to fewer repairs being paid
for by owners/operators since more costs will be borne by the manufacturer, and that the longer
duration useful life periods will result in increased emission control system durability. We have
estimated the net effect on repair costs and describe our approach, along with increased DEF
consumption and reduced gasoline consumption below.
As noted in the introductory text to this chapter, the operating costs we estimate here are for
the heavy-duty truck operation impacted by the final program (e.g., repair of emission-related
components). These costs (and savings) are incurred by heavy-duty truck purchasers/owners.
7.2.1 Costs Associated with Increased Diesel Exhaust Fluid (DEF) Consumption in Diesel
Engines
To estimate baseline case DEF consumption in heavy-duty vehicles with diesel engines, this
analysis uses the relationship, shown below, of DEF dose rate relative to the reduction in NOx
over the SCR catalyst.L> 17
NOx reduction = —73.679x + 0.0149
where x is equal to the DEF dose rate. This relationship was developed giving consideration to
FTP emissions. By estimating the FTP NOx reduction across the SCR catalyst, the DEF dose
rate can be calculated. NOx reduction is estimated from the difference between estimated engine-
out and FTP tailpipe NOx emissions; these variables along with the calculated DEF dose rate for
the baseline case are shown in Table 7-29.
L The relationship between DEF dose rate and NOx reduction across the SCR catalyst is based on methodology
presented in the Technical Support Document to the 2012 Non-conformance Penalty rule (the NCP Technical
Support Document, or NCP TSD).
349
-------
Table 7-29: Diesel Exhaust Fluid Consumption Rates for Diesel Vehicles in the Baseline Case
Value
Engine-out NOX
(FTP g/hp-hr)
4.0
Tailpipe NOX
(FTP mg/hp-hr)
200
DEF Dose Rate
(% of fuel consumed)
5.18%
To estimate DEF consumption impacts under the final program, which involves changes to
not only the new FTP emission standards but also the new SET and LLC standards along with
new off-cycle standards, we developed a new approach to estimate DEF consumption. For this
analysis, we scaled DEF consumption with the NOx reductions achieved under each of the
alternatives. To do this, we considered the molar mass of NOx, the molar mass of urea, the molar
ratio of NO to NO2, the mass concentration of urea in DEF along with the density of DEF to
estimate the theoretical gallons of DEF consumed per ton of NOx reduced at 442 gallons/ton.
The theoretical DEF dosing rates was then compared to the data collected from the CARB Stage
3 test program for the hot FTP, SET and LLC (see Chapter 3 of the RIA). The data from this
testing showed that the NOx specific DEF dosing was 536, 478, and 568 gallons/ton for the hot
FTP, SET and LLC, respectively. Since this data takes into account any over dosing that occurs
for part of the cycle, and NO2/NO ratio being greater than 1 for parts of the cycle, we have
adjusted the theoretical 442 gallons/ton NOx to the average of the hot FTP, SET and LLC, which
is 527 gallons/ton. These values are shown in Table 7-30.
Table 7-30: Derivation of DEF Consumption per Ton of NOx Reduced
Value
Molar mass of NOx (g/mol)
46.0055
Molar mass of urea (g/mol)
60.07
Molar ratio of NO to NO2
1
Mass concentration of urea in DEF
0.325
Density of DEF (g/mL)
1.09
Theoretical gallons DEF/ton NOx reduced
442
Proposed gallons DEF/ton NOx reduced
527
The final calculation of DEF consumption in each option is to multiply the gallons of diesel
fuel consumed by 5.18%, to account for the baseline DEF consumption and add to that amount
527 gallons of DEF for each ton of NOx reduced from the baseline. Both the gallons of diesel
fuel consumed and the tons of NOx reduced are taken directly from the year-over-year MOVES
results.
The gallons of DEF consumed are then multiplied by the estimated price of DEF per gallon.
This analysis uses the DEF prices presented in the NCP Technical Support Document with
growth beyond 2042 projected at the same 1.3 percent rate as noted in the NCP TSD. Note that
the DEF prices presented in Table 7-31 update the NCP TSD's 2011 prices to 2017 dollars using
the GDP deflator presented in Table 7-8.
350
-------
Table 7-31: Diesel Exhaust Fluid Price per Gallon (2017 dollars)
Calendar Year
DEF Price/Gallon
2027
3.25
2028
3.30
2029
3.33
2030
3.37
2031
3.42
2032
3.46
2033
3.52
2034
3.56
2035
3.60
2036
3.65
2037
3.69
2038
3.75
2039
3.79
2040
3.85
2041
3.89
2042
3.94
2043
4.00
2044
4.04
2045
4.10
The impacts on DEF costs are shown in Table 7-32. Note that the impacts of the final
program are the increased costs shown in Table 7-32, the baseline and final program costs are
shown to provide a sense of scale for the increased costs. Because these are operating costs
which occur over time, we present them at both 3 and 7 percent discount rates.
351
-------
Table 7-32: MY2027 Lifetime DEF Costs per Diesel Vehicle Associated with Final NOx Standards, 2017
dollars
3% Discount Rate
7% Discount Rate
Light
HDE
Medium
HDE
Heavy
HDE
Urban
Bus
Light
HDE
Medium
HDE
Heavy
HDE
Urban
Bus
FRM Baseline
Long-Haul Combination Trucks
34,009
25,768
Long-Haul Single Unit Trucks
3,759
5,686
6,823
2,937
4,443
5,331
Motor Homes
1,489
1,764
1,068
1,265
Other Buses
9,118
11,285
11,688
6,695
8,286
8,582
Refuse Trucks
8,435
8,787
6,317
6,581
School Buses
2,331
3,030
3,187
1,712
2,225
2,340
Short-Haul Combination Trucks
16,323
17,154
12,735
13,384
Short-Haul Single Unit Trucks
2,733
4,144
4,975
2,100
3,184
3,823
Transit Buses
9,192
11,254
11,742
6,750
8,263
8,622
FRM Baseline+Final Program
Long-Haul Combination Trucks
37,621
28,580
Long-Haul Single Unit Trucks
4,011
6,215
7,916
3,136
4,865
6,200
Motor Homes
1,617
2,016
1,162
1,450
Other Buses
9,805
12,277
13,594
7,209
9,040
10,011
Refuse Trucks
9,182
10,246
6,895
7,696
School Buses
2,501
3,293
3,671
1,839
2,424
2,702
Short-Haul Combination Trucks
17,575
19,378
13,727
15,154
Short-Haul Single Unit Trucks
2,949
4,573
5,864
2,268
3,522
4,517
Transit Buses
9,867
12,149
13,410
7,253
8,945
9,863
Increased Cost of the Final Program
Long-Haul Combination Trucks
3,612
2,812
Long-Haul Single Unit Trucks
252
529
1,094
199
422
869
Motor Homes
128
253
94
185
Other Buses
687
992
1,906
514
754
1,428
Refuse Trucks
747
1,459
579
1,115
School Buses
170
263
484
127
199
362
Short-Haul Combination Trucks
1,251
2,224
992
1,771
Short-Haul Single Unit Trucks
216
429
889
168
337
694
Transit Buses
675
896
1,669
504
681
1,241
7.2.2 Costs Associated with Changes in Fuel Consumption on Gasoline Engines
This analysis estimates a small decrease in fuel costs, i.e., fuel savings, by vehicles equipped
with gasoline engines because the final ORVR system will be expected to capture previously
evaporated fuel and then burn that fuel in the engine (see Table 7-6 for our estimated ORVR
direct manufacturing cost). To estimate these impacts, we first converted grams of hydrocarbon
captured by the ORVR system to milliliters of gasoline that will ultimately be burned in the
engine. Based on that relationship, we estimate that 1.48 milliliters of gasoline will be consumed
for each gram of hydrocarbon emissions reduced under the final program. We estimated this
value, 1.48, by assuming that the ORVR system will exchange captured butane for gasoline on
an energy basis and Tier 3 certification fuel has a density of 0.7482 g/ml, or 2832 g/gal at 60
degrees F.18 We then used a butane energy density of 45.8 MJ/kg, or 19752 Btu/lb,19 and the Tier
3 certification fuel energy density of 17890 Btu/lb,20 giving a ratio of 1.117 grams of gasoline
displaced for each gram of butane burned. Using the density of Tier 3 certification fuel, we get
352
-------
1.117/ 2832 = 0.0003943 gallons, or 1.48 ml, of gasoline saved for each gram of butane
captured since the owner/operator is no longer paying for evaporated fuel as it will be burned in
the engine. Using AEO 2019 reference case gasoline prices, the impacts on fuel costs for
MY2027 gasoline HD vehicles are shown in Table 7-33 (note that negative values indicate lower
fuel costs, or fuel savings). In the aggregate, we estimate that the ORVR requirements in the
final program will result in an annual reduction of approximately 0.3 million (calendar year
2027) to 4.9 million (calendar year 2045) gallons of gasoline, representing roughly 0.1 percent of
gasoline consumption from impacted vehicles.
Table 7-33: MY2027 Lifetime Fuel Costs per Gasoline Vehicle Associated with ORVR Requirements, 2017
dollars
3% Discount Rate
7% Discount Rate
Light
HDE
Medium
HDE
Heavy
HDE
Light
HDE
Medium
HDE
Heavy
HDE
FRM Baseline
Long-Haul Single Unit Trucks
120,876
150,530
192,727
94,841
118,108
151,216
Motor Homes
30,329
38,339
48,887
21,905
27,691
35,309
Other Buses
273,223
201,982
School Buses
69,242
51,188
Short-Haul Single Unit Trucks
86,494
109,427
139,754
66,791
84,501
107,918
Transit Buses
269,797
199,449
FRM Baseline+Final Program
Long-Haul Single Unit Trucks
120,744
150,349
192,470
94,739
117,969
151,019
Motor Homes
30,271
38,260
48,781
21,864
27,635
35,233
Other Buses
272,656
201,570
School Buses
69,110
51,092
Short-Haul Single Unit Trucks
86,397
109,292
139,566
66,717
84,399
107,777
Transit Buses
269,245
199,047
Increased Cost of the Final Program
Long-Haul Single Unit Trucks
-132
-181
-257
-102
-139
-197
Motor Homes
-58
-79
-106
-41
-56
-75
Other Buses
-567
-412
School Buses
-132
-96
Short-Haul Single Unit Trucks
-97
-135
-187
-74
-102
-141
Transit Buses
-552
-402
7.2,3 Emission-Related Repair Cost Impacts Associated with the Final Program
The final extended warranty and useful life requirements will have an impact on emission-
related repair costs incurred by heavy-duty vehicle owners. Researchers have noted the
relationships among quality, reliability, and warranty for a variety of goods.21 Wu, for instance,
examines how analyzing warranty data can provide "early warnings" on product problems that
can then be used for design modifications.22 Guajardo et al. describe one of the motives for
warranties to be "incentives for the seller to improve product quality;" specifically for light-duty
vehicles, they find that buyers consider warranties to substitute for product quality, and to
complement service quality.23 The other rationales are protection for consumers against failures,
provision of product quality information to consumers, and a means to distinguish consumers
according to their risk preferences. Murthy and Jack, for new products, and Saidi-Mehrabad et al.
353
-------
for second-hand products, consider the role of warranties in improving a buyer's confidence in
quality of the good.24'25
On the one hand, we expect owner-incurred emission repair costs to decrease due to the final
program because the longer emission warranty requirements will result in more repair costs
covered by the OEMs. Further, we expect improved serviceability in an effort by OEMs to
decrease repair costs they will incur. We also expect that the longer useful life periods in the
final program will result in more durable parts to ensure regulatory compliance over the longer
timeframe. On the other hand, we also expect that the more costly emission control systems
required by the final program may result in higher repair costs which might increase owner-
incurred costs outside the warranty and useful life periods.
As discussed in Chapter 7.1.2, we have estimated increased OEM costs associated with
increased warranty liability (i.e., longer warranty periods), and for more durable parts resulting
from the longer useful life periods. These costs are accounted for via increased warranty costs
and increased research and development (R&D) costs. We also included additional
aftertreatment costs in the direct manufacturing costs to address the increased useful life
requirements (e.g., larger catalyst volume; see Chapters 2 and 3 of the RIA for detailed
discussions). We estimate that these efforts will help to reduce emission repair costs during the
emission warranty and regulatory useful life periods, and possibly beyond.
In the proposal, to estimate impacts on emission repair costs, we began with an emission
repair cost curve derived from an industry white paper. 26 Some commenters took exception to
our approach, preferring instead that we use what they consider to be a more established repair
and maintenance cost estimate from the American Transportation Research Institute.27 Given the
duration of the ATRI study and the amount of data behind it, we have moved to using that study
in this final analysis (the November 2021 study).
In the ATRI study, 10 years of repair and maintenance costs are presented. Note that the costs
presented by ATRI are for all repair and maintenance, not just emission-related repair and
maintenance and not just emission-related repair—the real focus of our analysis since our
emission-related warranty and useful life provisions are geared only toward emission-related
systems. Also, in the ATRI study, they do not provide a dollar basis for their results. In looking
at a prior 2019 ATRI study,28 we found identical values presented in all applicable years as were
presented in the 2021 study (see Table 8 of the 2021 study and Table 9 of the 2019 study). This
led us to believe that the costs presented had not been updated to a consistent dollar basis but
instead were reported in nominal terms in each year. We then converted the ATRI costs to 2017
dollars using the deflators presented in Table 7-8 and calculated the average value over the 10
years of data to arrive at the 0.158 dollars per mile and 6.31 dollars per hour values used as a
starting point in our analysis.
354
-------
Table 7-34: Repair and Maintenance Costs per Mile
Calendar Year
ATRI Study,
Dollars per mile
EPA Assumed Dollar Basis
ATRI Cost in 2017 Dollars,
Dollars per mile
2011
0.152
2011
0.167
2012
0.138
2012
0.149
2013
0.148
2013
0.157
2014
0.158
2014
0.164
2015
0.156
2015
0.161
2016
0.166
2016
0.169
2017
0.167
2017
0.167
2018
0.171
2018
0.167
2019
0.149
2019
0.143
2020
0.148
2020
0.140
Average
0.155
0.158
Table 7-35: Repair and Maintenance Costs per Hour of Operation
Calendar Year
ATRI Study,
Dollars per hour
EPA Assumed Dollar Basis
ATRI Cost in 2017 Dollars,
Dollars per hour
2011
6.07
2011
6.66
2012
5.52
2012
5.95
2013
5.92
2013
6.27
2014
6.31
2014
6.56
2015
6.23
2015
6.41
2016
6.65
2016
6.78
2017
6.58
2017
6.58
2018
6.72
2018
6.56
2019
5.87
2019
5.63
2020
6.00
2020
5.69
Average
6.19
6.31
The ATRI study is a comprehensive study of tractor-trailer fleet freight operators. As such,
the costs presented in Table 7-34 and Table 7-35 are taken as representative of diesel heavy HDE
equipped vehicles. Given the different emission control system costs of light and medium HDEs,
we considered it necessary to adjust the 0.158 and 6.31 dollar values to be more applicable to
light and medium HD vehicles. To do this, we used the same approach as described above in
scaling diesel HDE engine warranty costs for engines of other sizes and other fuels. We also
wanted to adjust the repair and maintenance to reflect only emission-related repair, eliminating
not only non-emission-related repairs but also all maintenance. To do so, we used the approach
used for the same purpose and described in the proposal. To estimate the emission repair portion
of these costs, we used the figure on page 3 of the Fleet Advantage whitepaper which showed the
percent of total repair and maintenance costs attributable to different systems on the vehicle.29
The details of that chart are recreated below in Table 7-36 along with EPA's estimates for what
portion of the repair and maintenance costs could be considered to be emission repairs.
As shown, our analysis estimates that 10.8 percent of repair and maintenance costs are
emission repairs that our warranty and useful life provisions are meant to impact. In general, the
maintenance/repair shares are estimated to be 50/50 repair/maintenance, with the exception of
"Preventive maintenance" and "Exhaust system." Preventive maintenance, by definition, is not
355
-------
repair, and thus 100 percent is considered maintenance. For the exhaust system, we estimate that
80 percent of those costs are repair costs, with maintenance costs limited to DPF cleaning. The
share of emission-related vs. non-emission-related items was broken down, in general, by those
that are clearly emission-related, where we attribute 100 percent to the emission-related category,
versus those that are clearly not emission-related where 100 percent is attributed to the non-
emission-related category. Not shown are the non-emission repair share which totals 35.0
percent, the emission maintenance share which is 7.2 percent and the non-emission maintenance
share which is 47.0 percent.
Table 7-36: Percentage of Total Repair & Maintenance Costs Attributable to Different Vehicle Systems30
Fleet Advantage
EPA Estimates
System
Percent of Total
Repair &
Maintenance
Cost
Maintenance
Share
Repair
Share
Non-Emission-
Related Share
Emission-
Related
Share
Emission
Repair
Share
(EPA
Estimate)
Tires, tubes,
liners & valves
43%
50%
50%
100%
0%
0.0%
Preventive
maintenance
12%
100%
0%
100%
0%
0%
Brakes
9%
50%
50%
100%
0%
0%
Expendable
items
8%
50%
50%
100%
0%
0%
Lighting
5%
50%
50%
100%
0%
0%
Cranking
5%
50%
50%
100%
0%
0%
Power plant
3%
50%
50%
0%
100%
1.5%
Exhaust system
6%
20%
80%
0%
100%
4.8%
Fuel system
6%
50%
50%
0%
100%
3%
Engine/motor
3%
50%
50%
0%
100%
1.5%
Total
100%
54.2%
45.8%
82.0%
18.0%
10.8%
The end results are shown in Table 7-37 and Table 7-38 which show the dollar per mile and
dollar per hour values, respectively, used in this analysis for emission-related repairs. Note that
the costs shown are scaled upward as described in the text (i.e., package direct cost divided by
the baseline package direct cost within each regulatory class and fuel type) for emission-related
repairs done beyond the useful life. Which set of repair cost values we used, dollar per mile or
dollar per hour of operation, is shown in Table 7-39 along with the average speeds used to
estimate the number of hours of operation per year (MOVES vehicle miles travelled divided by
average speed). We have applied the dollar per hour value for most vocational vehicles and the
dollar per mile value for others.
356
-------
Table 7-37: Emission-Related Repair Costs per Mile, 2017 dollars per mile *
Repair & Maintenance
Emission-Related Repair
(10.8% of Repair & Maintenance)
Scaling Approach
Diesel
Gasoline
CNG
Diesel
Gasoline
CNG
Class 2b3
Base Class 2b3 DMC /
Base Diesel Heavy HDE DMC
0.100
0.011
Light HDE
Base Light HDE DMC /
Base Diesel Heavy HDE DMC
0.101
0.073
0.011
0.008
Medium HDE
Base Medium HDE DMC /
Base Diesel Heavy HDE DMC
0.103
0.073
0.011
0.008
Heavy HDE
Base Heavy HDE DMC /
Base Diesel Heavy HDE DMC
0.158
0.073
0.232
0.017
0.008
0.025
Urban Bus
Base Urban bus DMC /
Base Diesel Heavy HDE DMC
0.098
0.162
0.011
0.018
* The Base Diesel Heavy HDE DMC would be the $5,816 value shown in Table 7-5.
Table 7-38: Emission-Related Repair Costs per Hour of Operation, 2017 dollars per hour *
Repair & Maintenance
Emission-Related Repair
(10.8% of Repair & Maintenance)
Scaling Approach
Diesel
Gasoline
CNG
Diesel
Gasoline
CNG
Class 2b3
Base Class 2b3 DMC /
Base Diesel Heavy HDE DMC
3.99
0.431
Light HDE
Base Light HDE DMC /
Base Diesel Heavy HDE DMC
4.03
2.91
0.435
0.314
Medium HDE
Base Medium HDE DMC /
Base Diesel Heavy HDE DMC
4.11
2.91
0.444
0.314
Heavy HDE
Base Heavy HDE DMC /
Base Diesel Heavy HDE DMC
6.31
2.91
9.27
0.714
0.314
1.00
Urban Bus
Base Urban bus DMC /
Base Diesel Heavy HDE DMC
3.91
6.47
0.422
0.699
* The Base Diesel Heavy HDE DMC would be the $5,816 value shown in Table 7-5.
357
-------
Table 7-39: Repair Cost per Unit-Attribute and Average Speeds used in Calculating Emission-Related Repair
Costs
MOVES Sourcetype
Repair cost per unit attribute
Average Speed
(miles per hour) *
Passenger Trucks
Dollars per mile
42.9
Light Commercial Trucks
Dollars per mile
41.2
Other Buses
Dollars per hour
42.5
Transit Buses
Dollars per hour
43.0
School Buses
Dollars per hour
43.0
Refuse Trucks
Dollars per hour
44.0
Short-Haul Single Unit Trucks
Dollars per hour
43.9
Long-Haul Single Unit Trucks
Dollars per mile
44.5
Motor Homes
Dollars per mile
43.9
Short-Haul Combination Trucks
Dollars per mile
47.7
Long-Haul Combination Trucks
Dollars per mile
51.5
* Sonntag, Darrell. Population and Activity of Onroad Vehicles in MOVESCTINPRM. Attachment to
Memorandum to Docket EPA-HQ-OAR-2019-0055: "Updates to MOVES for Emissions Analysis of the HD 2027
NPRM." May 2021.
As noted above, given that future engines and vehicles will be equipped with new, more
costly technology, it is possible that the annual repair costs for vehicles under the final program
will be higher than the annual repair costs in the baseline. We have included such an increase for
the period beyond useful life by scaling the emission-related repair costs shown in Table 7-37
and Table 7-38 by the direct manufacturing costs of the applicable regulatory class divided by its
direct manufacturing costs in the baseline. In other words, if the final program direct
manufacturing cost for a diesel HDE is $8,455 and its baseline cost is $5,816, the 0.017
emission-related repair cost per mile value shown in Table 7-37 would become 0.025 (8455
divided by 5816 times 0.017). This is perhaps conservative because it seems reasonable to
assume that the R&D used to improve durability during the useful life period would also
improve durability beyond it. Nonetheless, we also think it is reasonable to include an increase in
repair costs, relative to the baseline case, because the period beyond useful life is of marginally
less concern to manufacturers^ Table 7-40, Table 7-41 and Table 7-42 show the emission-
related repair cost per mile or cost per hour values used in the analysis for the period beyond the
estimated useful life.
M This is not meant to suggest that manufacturers no longer care about their products beyond their regulatory useful
life, but rather to support the expectation that regulatory pressures~i.e., regulatory compliance during the useful life-
-tend to focus resources.
358
-------
Table 7-40: Diesel Emission-Related Repair Cost per Mile or Cost per Hour used for the Period Beyond
Useful Life
Baseline
Proposal
Scaling Approach
Cents per
mile
Cents per
hour
Cents per
mile
Cents per
hour
Class 2b3
Light Commercial Trucks
Final Class 2b3 DMC /
Base Class 2b3 DMC
1.08
1.74
Long-Haul Single Unit Trucks
1.08
1.74
Passenger Trucks
1.08
1.74
Short-Haul Single Unit Trucks
43.1
69.4
Light HDE
Long-Haul Single Unit Trucks
Final Light HDE DMC /
Base Light HDE DMC
1.09
1.75
Other Buses
43.3
70.1
School Buses
43.3
70.1
Short-Haul Single Unit Trucks
43.3
70.1
Transit Buses
43.3
70.1
Medium HDE
Long-Haul Single Unit Trucks
Final Medium HDE DMC /
Base Medium HDE DMC
1.12
1.75
Motor Homes
1.12
1.75
Other Buses
44.6
69.7
Refuse Trucks
44.6
69.7
School Buses
44.6
69.7
Short-Haul Combination Trucks
1.12
1.75
Short-Haul Single Unit Trucks
44.6
69.7
Transit Buses
44.6
69.7
Heavy HDE
Long-Haul Combination Trucks
Final Heavy HDE DMC /
Base Heavy HDE DMC
1.71
2.48
Long-Haul Single Unit Trucks
1.71
2.48
Motor Homes
1.71
2.48
Other Buses
68.1
99.1
Refuse Trucks
68.1
99.1
School Buses
68.1
99.1
Short-Haul Combination Trucks
1.71
2.48
Short-Haul Single Unit Trucks
68.1
99.1
Urban Bus
Transit Buses
Final Urban bus HDE DMC /
Base Urban bus HDE DMC
45.5
71.0
359
-------
Table 7-41: Gasoline Emission-Related Repair Cost per Mile or Cost per Hour used for the Period Beyond
Useful Life
Baseline
Proposal
Scaling Approach
Cents per
mile
Cents per
hour
Scaling
Approach
Cents per
mile
Light HDE
Long-Haul Single Unit Trucks
Final Light HDE DMC /
Base Light HDE DMC
0.79
0.99
Motor Homes
0.79
0.99
Other Buses
31.4
39.5
School Buses
31.4
39.5
Short-Haul Single Unit Trucks
31.4
39.5
Transit Buses
31.4
39.5
Medium HDE
Long-Haul Single Unit Trucks
Final Medium HDE DMC /
Base Medium HDE DMC
0.79
0.99
Motor Homes
0.79
0.99
Short-Haul Single Unit Trucks
31.4
39.5
Heavy HDE
Long-Haul Single Unit Trucks
Final Heavy HDE DMC /
Base Heavy HDE DMC
0.79
0.99
Motor Homes
0.79
0.99
Short-Haul Single Unit Trucks
31.4
39.5
Table 7-42: CNG Emission-Related Repair Cost per Mile or Cost per Hour used for the Period Beyond
Useful Life
Baseline
Proposal
Scaling Approach
Cents per
mile
Cents per
hour
Scaling
Approach
Cents per
mile
Heavy HDE
Long-Haul Single Unit Trucks
Final Heavy HDE DMC /
Base Heavy HDE DMC
2.52
2.53
Other Buses
100.6
100.9
Refuse Trucks
100.6
100.9
School Buses
100.6
100.9
Short-Haul Combination
Trucks
2.52
2.53
Short-Haul Single Unit Trucks
100.6
100.9
Urban Bus
Transit Buses
Final Urban bus HDE DMC /
Base Urban bus HDE DMC
75.4
75.7
As done for warranty costs, we have used estimated ages for when warranty and useful life
are reached, using the required miles, ages and hours along with the estimated miles driven and
hours of operation for each specific type of vehicle (refer to Table 7-15, Table 7-16, and Table
7-17 to see how many years of warranty and useful life are actually estimated for each type of
diesel, gasoline and CNG vehicle, respectively). As noted above, this means that warranty and
useful life ages are reached in different years for a long-haul combination truck driving over
100,000 miles per year or over 2,000 hours per year and a refuse truck driven around 40,000
miles per year or operating less than 1,000 hours per year. The resultant MY 2027 lifetime
emission-related repair costs are shown in Table 7-43 for diesel HD vehicles, Table 7-44 for
360
-------
gasoline HD vehicles and Table 7-45 for CNG HD vehicles. Since these costs occur over time,
we present them using both a 3 percent and a 7 percent discount rate.
Note that these costs assume that all emission-related repair costs are paid by manufacturers
during the warranty period, and beyond the warranty period the emission-related repair costs are
incurred by owners/operators.
361
-------
Table 7-43: MY2027 Lifetime Emission-Related Repair Costs per Diesel Vehicle, 2017 dollars
3% Discount Rate
7% Discount Rate
Class
2b3
Light
HDE
Medium
HDE
Heavy
HDE
Urban
bus
Class
2b3
Light
HDE
Medium
HDE
Heavy
HDE
Urban
bus
FRM Baseline
Light Commercial Trucks
1,502
1,030
Long-Haul
Combination Trucks
22,041
16,138
Long-Haul
Single Unit Trucks
3,192
3,208
2,493
3,060
2,429
2,440
1,790
2,109
Motor Homes
613
936
394
602
Other Buses
4,292
3,668
4,719
3,083
2,499
3,074
Passenger Trucks
1,502
1,030
Refuse Trucks
2,222
3,110
1,506
2,065
School Buses
1,148
1,050
1,604
771
684
1,045
Short-Haul
Combination Trucks
6,635
8,088
5,003
5,823
Short-Haul
Single Unit Trucks
1,790
1,799
1,292
1,973
1,311
1,318
876
1,338
Transit Buses
4,242
3,625
3,941
3,047
2,469
2,732
FRM Baseline+Final Program
Light Commercial Trucks
790
449
Long-Haul
Combination Trucks
25,070
17,497
Long-Haul
Single Unit Trucks
2,264
2,284
1,531
1,524
1,497
1,509
956
906
Motor Homes
480
728
272
415
Other Buses
4,090
3,261
3,454
2,598
1,978
1,979
Passenger Trucks
790
449
Refuse Trucks
1,408
2,038
819
1,180
School Buses
667
772
1,174
378
439
673
Short-Haul
Combination Trucks
7,029
6,436
4,960
4,225
Short-Haul
Single Unit Trucks
758
764
721
1,115
447
451
421
655
Transit Buses
4,042
3,224
2,394
2,567
1,955
1,370
Increased Cost of the Final Program
Light Commercial Trucks
-712
-581
Long-Haul
Combination Trucks
3,028
1,359
Long-Haul
Single Unit Trucks
-929
-924
-962
-1,536
-932
-931
-834
-1,203
Motor Homes
-132
-207
-122
-187
Other Buses
-203
-406
-1,265
-486
-520
-1,095
Passenger Trucks
-712
-581
Refuse Trucks
-814
-1,072
-687
-885
School Buses
-481
-278
-430
-393
-245
-372
Short-Haul
Combination Trucks
394
-1,651
-43
-1,598
Short-Haul
Single Unit Trucks
-1,032
-1,035
-570
-857
-864
-867
-455
-684
Transit Buses
-200
-402
-1,547
-480
-514
-1,362
362
-------
Table 7-44: MY2027 Lifetime Emission-Related Repair Costs per Gasoline Vehicle, 2017 dollars
3% Discount Rate
7% Discount Rate
Light
HDE
Medium
HDE
Heavy
HDE
Light
HDE
Medium
HDE
Heavy
HDE
FRM Baseline
Long-Haul Single Unit Trucks
2,324
2,324
2,324
1,768
1,768
1,768
Motor Homes
431
431
431
278
278
278
Other Buses
3,111
2,234
School Buses
832
559
Short-Haul Single Unit Trucks
1,304
1,304
1,304
955
955
955
Transit Buses
3,074
2,208
FRM Baseline+Final Program
Long-Haul Single Unit Trucks
1,831
1,831
1,831
1,271
1,271
1,271
Motor Homes
275
275
275
156
156
156
Other Buses
2,898
1,917
School Buses
442
252
Short-Haul Single Unit Trucks
764
764
764
483
483
483
Transit Buses
2,865
1,895
Increased Cost of the Final Program
Long-Haul Single Unit Trucks
-493
-493
-493
-497
-497
-497
Motor Homes
-156
-156
-156
-122
-122
-122
Other Buses
-212
-317
School Buses
-390
-306
Short-Haul Single Unit Trucks
-540
-540
-540
-471
-471
-471
Transit Buses
-210
-313
363
-------
Table 7-45: MY2027 Lifetime Emission-Related Repair Costs per CNG Vehicle, 2017 dollars
3% Discount Rate
7% Discount Rate
Heavy HDE
Urban Bus
Heavy HDE
Urban Bus
FRM Baseline
Long-Haul Single Unit Trucks
4,517
3,113
Other Buses
6,966
4,537
Refuse Trucks
4,590
3,048
School Buses
2,368
1,542
Short-Haul Combination Trucks
11,938
8,595
Short-Haul Single Unit Trucks
2,912
1,975
Transit Buses
6,532
4,529
FRM Baseline+Final Program
Long-Haul Single Unit Trucks
1,720
1,029
Other Buses
3,807
2,194
Refuse Trucks
2,260
1,317
School Buses
1,294
746
Short-Haul Combination Trucks
7,723
5,143
Short-Haul Single Unit Trucks
1,248
737
Transit Buses
2,822
1,626
Increased Cost of the Final Program
Long-Haul Single Unit Trucks
-2,797
-2,084
Other Buses
-3,158
-2,344
Refuse Trucks
-2,330
-1,732
School Buses
-1,074
-797
Short-Haul Combination Trucks
-4,215
-3,452
Short-Haul Single Unit Trucks
-1,664
-1,238
Transit Buses
-3,710
-2,903
7.3 Program Costs
Using the cost elements outlined above, we have estimated the costs associated with the final
program as presented in the following tables. Costs are broken into two main categories:
Technology Costs and Operating Costs. Technology costs are broken further into direct costs and
the indirect costs elements (warranty costs, R&D costs, other indirect costs and profits) to arrive
at total technology costs incurred by manufacturers (i.e., regulated entities). Operating costs are
broken into urea/DEF costs (diesel only), fuel savings (gasoline only) and repair costs to arrive at
total operating costs incurred by owner/operators of new MY2027 and later HD vehicles. Section
7.3.1 presents the total technology costs for the final program and the updated costs for proposed
Option 2, both relative to the updated baseline case costs. Section 7.3.2 presents the operating
costs, similarly grouped. Section 7.3.3 presents the total program costs relative to the baseline
case for the final program and the updated proposed Option 2. Costs are presented in 2017
dollars in undiscounted annual values along with present and equivalent annualized values (PV
and EAV, respectively) at both 3 and 7 percent discount rates with discounted values discounted
to the 2027 calendar year.
7.3.1 Total Technology Costs
The tables shown here show direct manufacturing, warranty, R&D, profits, other indirect
costs and total technology costs incurred by manufacturers. Values shown for a given calendar
year are undiscounted values while discounted present values (PV) and equivalent annualized
364
-------
values (EAV) are presented at both 3 and 7 percent discount rates with values discounted to
2027. All values are shown in 2017 dollars.
Table 7-46: Technology Cost Impacts of the Final Program Relative to the Baseline Case, Millions of 2017
dollars *
Calendar Year
Direct
Manufacturing
Costs
Warranty
Costs
R&D
Costs
Other
Indirect
Costs
Profits
Total
Technology
Costs
2027
1,100
2,100
210
340
58
3,800
2028
1,100
2,100
200
320
55
3,700
2029
1,000
2,100
190
310
53
3,700
2030
1,000
2,100
51
300
52
3,500
2031
1,000
2,200
50
300
51
3,600
2032
990
2,200
49
290
50
3,600
2033
980
2,200
49
290
50
3,600
2034
980
2,300
49
290
49
3,600
2035
960
2,300
48
280
49
3,700
2036
950
2,300
48
280
48
3,700
2037
950
2,400
48
280
48
3,700
2038
950
2,400
48
280
48
3,700
2039
950
2,500
47
280
48
3,800
2040
950
2,500
47
280
48
3,800
2041
950
2,500
47
280
48
3,900
2042
950
2,600
47
280
48
3,900
2043
950
2,600
47
280
48
3,900
2044
950
2,700
48
280
48
4,000
2045
950
2,700
48
280
48
4,100
PV, 3%
14,000
33,000
1,100
4,200
720
53,000
PV, 7%
10,000
24,000
900
3,000
520
38,000
EAV, 3%
990
2,300
78
290
50
3,700
EAV, 7%
1,000
2,300
87
290
51
3,700
* Values show 2 significant digits. Note that the Information Collection Request costs addressed in Section
XII of the preamble would fall within the "Other" indirect costs shown here.
365
-------
Table 7-47: Technology Cost Impacts of the Updated Proposed Option 2 Relative to the Baseline Case,
Millions of 2017 dollars *
Calendar Year
Direct
Manufacturing
Costs
Warranty
Costs
R&D
Costs
Other
Indirect
Costs
Profits
Total
Technology
Costs
2027
1,100
370
190
340
58
2,100
2028
1,100
370
180
320
55
2,000
2029
1,000
370
180
310
53
1,900
2030
1,000
370
51
300
52
1,800
2031
1,000
370
50
300
51
1,800
2032
990
380
49
290
50
1,800
2033
980
380
49
290
50
1,700
2034
980
380
49
290
49
1,700
2035
960
390
48
280
49
1,700
2036
950
390
48
280
48
1,700
2037
950
390
48
280
48
1,700
2038
950
400
48
280
48
1,700
2039
950
400
47
280
48
1,700
2040
950
410
47
280
48
1,700
2041
950
410
47
280
48
1,700
2042
950
420
47
280
48
1,700
2043
950
420
47
280
48
1,700
2044
950
430
48
280
48
1,800
2045
950
440
48
280
48
1,800
PV, 3%
14,000
5,600
1,100
4,200
720
26,000
PV, 7%
10,000
4,000
850
3,000
520
19,000
EAV, 3%
990
390
75
290
50
1,800
EAV, 7%
1,000
390
82
290
51
1,800
* Values show 2 significant digits. Note that the Information Collection Request costs addressed in Section
XII of the preamble would fall within the "Other" indirect costs shown here.
73.2 Total Operating Costs
The tables shown here show emission repair costs, urea costs, pre-tax fuel costs and total
operating costs incurred by owner/operators of new MY 2027 and later HD vehicles. Values
shown for a given calendar year are undiscounted values while discounted values are presented
at both 3 and 7 percent discount rates. All values are shown in 2017 dollars.
Note that some values are shown as negative costs, or savings. This is expected with respect
to the fuel costs since the ORVR requirements are expected to reduce gasoline consumption due
to the capture of previously evaporated emissions. With respect to the emission repair costs, new
MY 2027 and later HD vehicles will have no emission-related repair costs in the early years
(MYs 2027 through 2030) since all of those vehicles will be covered under warranty. However,
by MY 2031 and later, those early compliant vehicles will begin to experience emission-related
repairs thereby countering the lack of emission-related repairs on new MY 2031 and later
vehicles. This pattern will continue, with older vehicles eventually experiencing emission-related
repairs that outweigh the comparatively smaller number of new vehicles covered under warranty.
This explains the lower magnitude of emission repair savings in the later years shown in the
tables.
366
-------
Table 7-48: Operating Cost Impacts of the Final Program Relative to the Baseline Case, Millions of 2017
dollars *
Calendar Year
Emission Repair Costs
Urea Costs
Pre-tax Fuel Costs
Total Operating Costs
2027
0
57
-0.39
57
2028
-47
120
-0.82
70
2029
-300
180
-1.3
-120
2030
-430
250
-1.7
-190
2031
-500
330
-2.2
-170
2032
-570
410
-2.7
-160
2033
-610
470
-3.4
-140
2034
-640
530
-4.1
-110
2035
-660
580
-4.8
-82
2036
-660
630
-5.4
-38
2037
-600
680
-6.0
65
2038
-540
720
-6.6
170
2039
-490
760
-7.2
270
2040
-450
800
-7.8
340
2041
-410
840
-8.3
410
2042
-390
870
00
00
1
470
2043
-370
910
-9.3
530
2044
-350
940
-9.7
570
2045
-340
970
-10
620
PV, 3%
-6,200
7,700
-69
1,400
PV, 7%
-4,300
4,900
-43
600
EAV, 3%
-430
540
-4.8
99
EAV, 7%
-420
480
-4.2
58
* Values show 2 significant digits.
367
-------
Table 7-49: Operating Cost Impacts of the Updated Proposed Option 2 Relative to the Baseline Case, Millions
of 2017 dollars *
Calendar Year
Emission Repair Costs
Urea Costs
Pre-tax Fuel Costs
Total Operating Costs
2027
0
55
-0.39
55
2028
-47
110
-0.82
65
2029
-190
170
-1.3
-17
2030
-250
240
-1.7
-17
2031
-300
310
-2.2
-0.079
2032
-220
370
-2.7
160
2033
-140
420
-3.4
280
2034
-71
470
-4.1
400
2035
-8.4
520
-4.8
500
2036
62
560
-5.4
610
2037
150
600
-6.0
740
2038
230
640
-6.6
860
2039
300
670
-7.2
970
2040
350
710
-7.8
1,100
2041
400
740
-8.3
1,100
2042
430
770
00
00
1
1,200
2043
460
800
-9.3
1,300
2044
490
830
-9.7
1,300
2045
510
860
-10
1,400
PV, 3%
1,100
6,800
-69
7,900
PV, 7%
310
4,400
-43
4,700
EAV, 3%
76
480
-4.8
550
EAV, 7%
30
430
-4.2
450
* Values show 2 significant digits.
Note that the ORVR requirements will result in previously evaporated gasoline being used by
in the engines of gasoline vehicles. We have estimated the cost savings that owner/operators will
experience and present those in Section 7.2.2. In this section, we also show the pre-tax fuel
savings that are ultimately part of the benefit-cost analysis presented in Chapter 9. Table 7-50
shows the impacts on fuel tax revenues that will be expected from these changes under the final
program.
368
-------
Table 7-50: Fuel Cost and Transfer Impacts of the Final Program Relative to the Baseline Case, Millions of
2017 dollars
Calendar Year
Retail Fuel Costs
Pre-tax Fuel Costs
Tax Revenues
2027
-0.47
-0.39
-0.076
2028
-0.97
-0.82
-0.16
2029
-1.5
-1.3
-0.26
2030
-2.1
-1.7
-0.35
2031
-2.7
-2.2
-0.44
2032
-3.3
-2.7
-0.53
2033
-4.1
-3.4
-0.65
2034
-4.9
-4.1
-0.77
2035
-5.7
-4.8
-0.88
2036
-6.4
-5.4
-0.99
2037
-7.1
-6.0
-1.1
2038
-7.8
-6.6
-1.2
2039
-8.5
-7.2
-1.3
2040
-9.1
-7.8
-1.3
2041
-9.7
-8.3
-1.4
2042
-10
00
00
1
-1.5
2043
-11
-9.3
-1.5
2044
-11
-9.7
-1.6
2045
-12
-10
-1.7
PV, 3%
-81
-69
-12
PV, 7%
-51
-43
-7.7
EAV, 3%
-5.6
-4.8
-0.85
EAV, 7%
-4.9
-4.2
-0.75
7,3,3 Total Program Costs
The tables shown here present technology costs, operating costs and the sum of the two for
final program and the updated proposed Option 2. Values shown for a given calendar year are
undiscounted values while discounted values are presented at both 3 and 7 percent discount rates.
All values are shown in 2017 dollars.
369
-------
Table 7-51: Total Technology & Operating Cost Impacts of the Final Program Relative to the Baseline Case,
Millions of 2017 dollars
Total
Total
Calendar Year
Technology
Operating
Sum
Costs
Costs
2027
3,800
57
3,900
2028
3,700
70
3,800
2029
3,700
-120
3,600
2030
3,500
-190
3,400
2031
3,600
-170
3,400
2032
3,600
-160
3,400
2033
3,600
-140
3,500
2034
3,600
-110
3,500
2035
3,700
-82
3,600
2036
3,700
-38
3,600
2037
3,700
65
3,800
2038
3,700
170
3,900
2039
3,800
270
4,000
2040
3,800
340
4,200
2041
3,900
410
4,300
2042
3,900
470
4,400
2043
3,900
530
4,500
2044
4,000
570
4,600
2045
4,100
620
4,700
PV, 3%
53,000
1,400
55,000
PV, 7%
38,000
600
39,000
EAV, 3%
3,700
99
3,800
EAV, 7%
3,700
58
3,800
* Values show 2 significant digits.
370
-------
Table 7-52: Total Technology & Operating Cost Impacts of the Updated Proposed Option 2 Relative to the
Baseline Case, Millions of 2017 dollars
Total
Total
Calendar Year
Technology
Operating
Sum
Costs
Costs
2027
2,100
55
2,100
2028
2,000
65
2,100
2029
1,900
-17
1,900
2030
1,800
-17
1,800
2031
1,800
-0.079
1,800
2032
1,800
160
1,900
2033
1,700
280
2,000
2034
1,700
400
2,100
2035
1,700
500
2,200
2036
1,700
610
2,300
2037
1,700
740
2,500
2038
1,700
860
2,600
2039
1,700
970
2,700
2040
1,700
1,100
2,800
2041
1,700
1,100
2,900
2042
1,700
1,200
2,900
2043
1,700
1,300
3,000
2044
1,800
1,300
3,100
2045
1,800
1,400
3,100
PV, 3%
26,000
7,900
34,000
PV, 7%
19,000
4,700
23,000
EAV, 3%
1,800
550
2,300
EAV, 7%
1,800
450
2,300
* Values show 2 significant digits.
Chapter 7 References
1 See HD2027_FRM_CostAnalysis_vl.2.0, Docket ID No. EPA-HQ-OAR-2019-0055.
2 "Cost Reduction through Learning in Manufacturing Industries and in the Manufacture of
Mobile Sources, Final Report and Peer Review Report," EPA-420-R-16-018, November 2016.
3 See the 2010 light-duty greenhouse gas rule (75 FR 25324, May 7, 2010); the 2012 light-duty
greenhouse gas rule (77 FR 62624, October 15, 2012); the 2011 heavy-duty greenhouse gas rule
(76 FR 57106, September 15, 2011); the 2016 heavy-duty greenhouse gas rule (81 FR 73478,
October 25, 2016); the 2014 light-duty Tier 3 rule (79 FR 23414, April 28, 2014).
4 Francisco Posada, Sarah Chambliss, and Kate Blumberg, "Costs of Emission Reduction
Technologies for Heavy-Duty Diesel Vehicles," International Council on Clean Transportation,
February 2016 (ICCT 2016).
5 See ICCT 2016 at page 26.
371
-------
6 Francisco Posada Sanchez, Anup Bandivadekar, John German, "Estimated Cost of Emission
Reduction Technologies for Light-Duty Vehicles," International Council on Clean
Transportation, March 2012 (ICCT 2012).
7 Mamidanna, S. 2021. Heavy-Duty Engine Valvetrain Technology Cost Assessment. U.S. EPA
Contract with FEV North America, Inc., Contract No. 68HERC19D0008, Task Order No.
68HERH20F0041 .Submitted to the Docket.
8 Mamidanna, S. 2021. Heavy-Duty Vehicles Aftertreatment Systems Cost Assessment.
Submitted to the Docket.
9 "Cost Reduction through Learning in Manufacturing Industries and in the Manufacture of
Mobile Sources, Final Report and Peer Review Report," EPA-420-R-16-018, November 2016.
10 See 75 FR 25324, 76 FR 57106, 77 FR 62624, 79 FR 23414, 81 FR 73478, 86 FR 74434.
11 See the 2010 light-duty greenhouse gas rule (75 FR 25324, May 7, 2010); the 2012 light-duty
greenhouse gas rule (77 FR 62624, October 15, 2012); the 2011 heavy-duty greenhouse gas rule
(76 FR 57106, September 15, 2011); the 2016 heavy-duty greenhouse gas rule (81 FR 73478,
October 25, 2016); the 2014 light-duty Tier 3 rule (79 FR 23414, April 28, 2014).
12 Rogozhin, Alex, Michael Gallaher, Gloria Helfand, and Walter McManus. "Using Indirect
Cost Multipliers to Estimate the Total Cost of Adding New Technology in the Automobile
Industry." International Journal of Production Economics 124 (2010): 360-368; Heavy Duty
Truck Retail Price Equivalent and Indirect Cost Multipliers, Draft Report, RTI International, RTI
Project Number 021 1577.003.002, July 2010.
13 Rogozhin, Alex, Michael Gallaher, Gloria Helfand, and Walter McManus. "Using Indirect
Cost Multipliers to Estimate the Total Cost of Adding New Technology in the Automobile
Industry." International Journal of Production Economics 124 (2010): 360-368.
14 See the 2011 heavy-duty greenhouse gas rule (76 FR 57106, September 15, 2011); the 2016
heavy-duty greenhouse gas rule (81 FR 73478, October 25, 2016); the 2014 light-duty Tier 3 rule
(79 FR 23414, April 28, 2014).
15 Heavy Duty Truck Retail Price Equivalent and Indirect Cost Multipliers, Draft Report, RTI
International, RTI Project Number 021 1577.003.002, July 2010.
16 Rogozhin, Alex, Michael Gallaher, Gloria Helfand, and Walter McManus. "Using Indirect
Cost Multipliers to Estimate the Total Cost of Adding New Technology in the Automobile
Industry." International Journal of Production Economics 124 (2010): 360-368.
17 "Nonconformance Penalties for On-highway Heavy-duty Diesel Engines: Technical Support
Document," EPA-420-R-12-014, Figure 3-1 at page 37.
18 "Tier 3 Certification Fuel Impacts Test Program, Appendix A, Table A-l," EPA-420-R-18-
004.
19 See "TheEngineeringToolBox.pdf," generated via
https://www.engineeringtoolbox.com/gross-net-heating-values-d_420.html, accessed May 29,
2020.
372
-------
20 "Tier 3 Certification Fuel Impacts Test Program, Appendix A, Table A-l," EPA-420-R-18-
004.
21 Thomas, M., and S. Rao (1999). "Warranty Economic Decision Models: A Summary and
Some Suggested Directions for Future Research." Operations Research 47(6):807-820.
22 Wu, S (2012). Warranty Data Analysis: A Review. Quality and Reliability Engineering
International 28: 795-805.
23 Guajardo, J., M Cohen, and S. Netessine (2016). "Service Competition and Product Quality in
the U.S. Automobile Industry." Management Science 62(7): 1860-1877.
24 Murthy, D., and N. Jack (2009). "Warranty and Maintenance," Chapter 18 in Handbook of
Maintenance Management and Engineering, Mohamed Ben-Daya et al., editors. London:
Springer.
25 Saidi-Mehrabad, M., R. Noorossana, and M. Shafiee (2010). "Modeling and analysis of
effective ways for improving the reliability of second-hand products sold with warranty."
International Journal of Advanced Manufacturing Technology 46: 253-265.
26 See "Mitigating Rising Maintenance & Repair Costs for Class-8 Truck Fleets, Effective Data
& Strategies to Leverage Newer Trucks to Reduce M&R Costs," Fleet Advantage Whitepaper
Series, 2018.
27 "An Analysis of the Operational Costs of Trucking: 2021 Update," American Transportation
Research Institute, November 2021.
28 "An Analysis of the Operational Costs of Trucking: 2019 Update," American Transportation
Research Institute, November 2019.
29 See "Mitigating Rising Maintenance & Repair Costs for Class-8 Truck Fleets, Effective Data
& Strategies to Leverage Newer Trucks to Reduce M&R Costs," Fleet Advantage Whitepaper
Series, 2018.
30 See "Mitigating Rising Maintenance & Repair Costs for Class-8 Truck Fleets, Effective Data
& Strategies to Leverage Newer Trucks to Reduce M&R Costs," Fleet Advantage Whitepaper
Series, 2018.
373
-------
Chapter 8 Estimated Benefits
8.1 Overview
The highway heavy-duty engines and vehicles subject to the final rule are significant sources
of mobile source air pollution, including directly-emitted PM2.5 as well as NOx and VOC
emissions (both precursors to ozone formation and secondarily-formed PM2.5). The final program
will reduce exhaust emissions of these pollutants from the regulated engines and vehicles, which
will in turn reduce ambient concentrations of ozone and PM2.5 Estimated emission reductions
are presented in Chapter 5, and air quality impacts of the standards are presented in Chapter 6.
Exposures to these pollutants are linked to adverse environmental and human health impacts,
such as premature deaths and non-fatal illnesses (see Chapter 4).
This chapter describes the methods used to estimate health benefits from reducing
concentrations of ozone and PM2.5 As noted in Chapter 6, full-scale photochemical air quality
modeling was performed for the proposal. No further air quality modeling has been conducted to
reflect the emissions impacts of the final program. Because air quality modeling results are
necessary to quantify estimates of avoided mortality and illness attributable to changes in
ambient PM2.5 and ozone, we present the benefits from the proposal as a proxy for the health
benefits associated with the final program. Chapter 5 describes the differences in emissions
between those used to estimate the air quality impacts of the proposal and those that will be
achieved by the final program. Emission reductions associated with the final program are similar
to those used in the air quality modeling conducted for the proposal (Section 5.5.4). We therefore
conclude that the health benefits from the proposal are a fair characterization of those that will be
achieved due to the substantial improvements in air quality attributable to the final program.
Using the air quality modeling from the proposal, we have quantified and monetized health
impacts in 2045, representing projected impacts associated with a year when the program will be
fully implemented and when most of the regulated fleet will have turned over. There are also
benefits associated with the standards that, if quantified and monetized, would increase the total
monetized benefits. These unquantified benefits are discussed in Section 8.8 of this chapter.
Overall, we estimate that the final program will lead to a substantial decrease in adverse PM2.5-
and ozone-related health impacts in 2045.
The approach we used to estimate health benefits is consistent with the approach described in
the technical support document (TSD) that was published for the final Revised Cross-State Air
Pollution Rule (CSAPR) Update RIA. K2-A Estimating the health benefits of reductions in PM2.5
and ozone exposure begins with estimating the change in exposure for each individual and then
estimating the change in each individual's risks for those health outcomes affected by exposure.
The benefit of the reduction in each health risk is based on the exposed individual's willingness
to pay (WTP) for the risk change, assuming that each outcome is independent of one another.
The greater the magnitude of the risk reduction from a given change in concentration, the greater
the individual's WTP, all else equal. The social benefit of the change in health risks equals the
A On March 15, 2021, EPA finalized the Revised Cross-State Air Pollution Rule Update for the 2008 ozone National
Ambient Air Quality Standards (NAAQS). Starting in the 2021 ozone season, the rule will require additional
emissions reductions of nitrogen oxides (NOx) from power plants in 12 states, https://www.epa.gov/csapr/revised-
cross-state-air-pollution-rule-update.
374
-------
sum of the individual WTP estimates across all of the affected individuals. We conduct this
analysis by adapting primary research - specifically, air pollution epidemiology studies and
economic value studies - from similar contexts. This approach is sometimes referred to as
"benefits transfer." Below we describe the procedure for quantifying and monetizing the health
benefits associated with reduced human exposure to PM2.5 and ozone.
8.2 Health Impact Assessment for PM2.5 and Ozone
There are four distinct steps the Agency follows when conducting a health impacts
assessment, each of which are described in this section: (1) prepare air quality modeling data for
health impacts analysis; (2) select air pollution health endpoints to quantify; (3) calculate counts
of air pollution effects using a health impact function; (4) specify the health impact function with
concentration-response parameters drawn from the epidemiological literature.
8.2,1 Preparing Air Quality Modeling Data for Health Impacts Analysis
In RIA Chapter 5, we present the emissions that will be reduced by the final rule, including
NOx, direct PM, and VOCs, all of which contribute to ambient concentrations of PM2.5 and
ozone. These reduced emissions will benefit public health and the environment since exposure to
ozone and PM2.5 is linked to adverse public health and environmental effects.6 In RIA Chapter 6,
we summarize the air quality modeling methods and results. These air quality results, measured
in terms of ambient concentrations of PM2.5 and ozone, are in turn associated with human
populations to estimate changes in health effects. This section describes how the CMAQ
modeling output was converted into a format suitable for the health impacts analysis using
EPA's Environmental Benefits Mapping and Analysis Program - Community Edition
(BenMAP-CE).c
The first step was to extract 2016 base year predicted hourly, surface-layer PM2.5 and ozone
concentrations for each grid cell directly from the standard CMAQ output files (at a 12-km by
12-km resolution). For ozone, we generated predicted ozone concentration surfaces for each of
three different warm seasons defined by the underlying health studies used in the analysis: April-
September, May-September, and June-August. These hourly model predictions were then
combined with monitored observations obtained from the Agency's Air Quality System (AQS)
to interpolate hourly ozone concentrations to 12-km by 12-km grid cells for the contiguous 48
B As noted in Chapter 6, full-scale photochemical air quality modeling was performed for the proposal. No further
air quality modeling has been conducted to reflect the emissions impacts of the final program. Since air quality
modeling results are necessary to quantify estimates of avoided mortality and illness attributable to changes in
ambient PM2 5 and ozone, we present the air quality and associated benefits from the proposal as a proxy for the
health benefits associated with the final program. Chapter 5 describes the differences in emissions between those
used to estimate the air quality impacts of the proposal and those that will be achieved by the final program.
Emission reductions associated with the final program are similar to those used in the air quality modeling
conducted for the proposal. We therefore conclude that the health benefits from the proposal are a fair
characterization of those that will be achieved due to the substantial improvements in air quality attributable to the
final program. We do not expect the magnitude of any differences to materially impact our cost-benefit conclusions.
c BenMAP-CE is an open-source computer program that calculates the number and economic value of air pollution-
related deaths and illnesses. The software incorporates a database that includes many of the concentration-response
relationships, population files, and health and economic data needed to quantify these impacts. More information
about BenMAP-CE, including downloadable versions of the tool and associated user manuals, can be found at
EPA's website www.epa.gov/benmap.
375
-------
states to create gridded 2016 surfaces informed by observational data.D'E We then converted
these warm-season hourly ozone concentrations to an ozone metric of interest, such as the daily
maximum 8-hour average concentration or the daily maximum 1-hour average concentration,
again consistent with the underlying health studies used in the analysis. Gridded fields of relative
response factors (RRFs) were created for each ozone metric and warm season definition of
interest by dividing unadjusted future year (2045) CMAQ concentrations by unadjusted 2016
base year CMAQ concentrations. Separate 12-km gridded RRFs were created for the future year
base case and policy cases for each metric/season combination. Then final future year air quality
surfaces were created by multiplying each of the RRF surfaces by the 2016 eVNA surface. These
surfaces then served as inputs to the health impact functions of the benefits analysis, contained
within BenMAP-CE.
For PM2.5, we also used 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. A gridded field of PM2.5
concentrations was created by interpolating Federal Reference Monitor ambient data and
Interagency Monitoring of Protected Visual Environments (IMPROVE) ambient data. Gridded
fields of PM2.5 species concentrations were created by interpolating EPA Chemical Speciation
Network (CSN) ambient data and IMPROVE data. The ambient data were interpolated to the
CMAQ 12-km grid. Future-year estimates of PM2.5 were calculated using gridded RRFs applied
to gridded 2016 ambient PM2.5 and PM2.5 species concentrations.
The procedures for determining the RRFs are similar to those in EPA's Modeling Guidance
for Demonstrating Air Quality Goals for Ozone, PM2.5, and Regional Haze.3 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 PM2.5 design values
is called the "Speciated Modeled Attainment Test (SMAT)." EPA used this procedure to
estimate the ambient impacts of the proposal.
Table 8-1 provides ozone and PM2.5 metrics for those 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.
D The 12-km grid squares contain the population data used in the health benefits analysis model, BenMAP-CE.
E This approach is a generalization of planar interpolation that is technically referred to as enhanced Voronoi
Neighbor Averaging (eVNA) spatial interpolation. See the BenMAP-CE manual for technical details, available for
download at http://www.epa.gov/benmap.
376
-------
Table 8-1: Summary of CMAQ-Derived Population-Weighted Ozone and PM2.5 Air Quality Metrics for
Health Benefits Endpoints
2045
Statistic3
Baseline
Chang^
Ozone Metric: National Population-Weighted Average (ppb)c
Daily Maximum 8-Hour Average Concentration - May-September
39
0.69
Daily Maximum 8-Hour Average Concentration - April-September
39
0.64
Daily Maximum 8-Hour Average Concentration - June-August
38
0.76
Daily Maximum 1-Hour Average Concentration - April-September
44
0.80
PM2.iMetric: National Population-Weighted Average (/ug/m3)c
Annual Average Concentration
7.3
0.034
a Ozone and PM2.5 metrics were calculated at the CMAQ grid-cell level for use in health effects estimates.
Ozone metrics were calculated over relevant time periods during daylight hours of each "ozone season."
Note that the national, population-weighted PM2.5 and ozone air quality metrics presented in this table
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 RIA Chapter 6, which represent the average
for areas with a current air quality monitor.
b The change is defined as the baseline value minus the control-case value.
0 Calculated by summing the product of the projected CMAQ grid-cell population and the estimated CMAQ
grid concentration and then dividing by the total population.
8.2.2 Selecting Air Pollution Health Endpoints to Quantify
As a first step in quantifying ozone and PIVh.s-related human health impacts, the Agency
consults the Integrated Science Assessment for Ozone and Related Photochemical Oxidants
(Ozone ISA) and the Integrated Science Assessment for Particulate Matter (PM ISA). These two
documents synthesize the toxicological, clinical and epidemiological evidence to determine
whether each pollutant is causally related to an array of adverse human health outcomes
associated with either short-term (i.e., hours to less than one month) or long-term (i.e., one
month to years) exposure; for each outcome, the ISA reports this relationship to be "causal",
"likely to be causal", "suggestive of, but not sufficient to infer, a causal relationship",
"inadequate to infer a causal relationship" or "not likely to be a causal relationship". The Agency
estimates the incidence of air pollution effects for those health endpoints that the ISA classified
as either "causal" or "likely-to-be-causal".
In brief, the ISA for ozone found short-term exposures to ozone to have a "causal"
relationship with respiratory effects, a "likely to be causal" relationship with metabolic effects
and a "suggestive of, but not sufficient to infer, a causal relationship" with central nervous
system effects, cardiovascular effects, and total mortality. The ISA reported that long-term
exposures to ozone are "likely to be causal" for respiratory effects including respiratory
mortality, and "suggestive of, but not sufficient to infer, a causal relationship" for cardiovascular
effects, reproductive effects, central nervous system effects, metabolic effects, and total
mortality. The PM ISA found short-term exposure to PM2.5 to have a "causal" relationship with
cardiovascular effects and mortality (i.e., premature death), "likely to be causal" relationship
with respiratory effects, and "suggestive of, but not sufficient to infer, a causal relationship" with
metabolic effects and nervous system effects. The ISA identified cardiovascular effects and total
mortality as have a "causal" relationship with long-term exposure to PM2.5. A "likely to be
causal" relationship was determined between long-term PM2.5 exposures and respiratory effects,
nervous system effects, and cancer effects, and the evidence was "suggestive of, but not
377
-------
sufficient to infer, a causal relationship" for male and female reproduction and fertility effects,
pregnancy and birth outcomes, and metabolic effects.
Table 8-2 reports the effects we quantified and those we did not quantify in this RIA. The list
of benefit categories not quantified is not exhaustive. And, among the effects quantified, it is not
always possible to completely quantify the full range of human health impacts or economic
values. The table below omits health effects associated with NO2 exposure, and any welfare
effects such as acidification and nutrient enrichment; these effects are described in the Ozone and
PM NAAQS RIAs and summarized later in this chapter.4'5
378
-------
Table 8-2: Health Effects of Ambient Ozone and PMzs
Category
Effect
Effect
Quantified
Effect
Monetized
More
Information
Premature
mortality from
exposure to PM2.5
Adult premature mortality from long-term exposure (age 65-
99 or age 30-99)
¦/
¦/
PMISA
Infant mortality (age <1)
¦/
¦/
PMISA
Nonfatal
morbidity from
exposure to PM2.5
Heart attacks (age >18)
¦/
~a
PMISA
Hospital admissions—cardiovascular (ages 65-99)
¦/
¦/
PMISA
Emergency department visits— cardiovascular (age 0-99)
¦/
¦/
PMISA
Hospital admissions—respiratory (ages 0-18 and 65-99)
¦/
¦/
PMISA
Emergency room visits—respiratory (all ages)
¦/
¦/
PMISA
Cardiac arrest (ages 0-99; excludes initial hospital and/or
emergency department visits)
¦/
~a
PMISA
Stroke (ages 65-99)
¦/
~a
PMISA
Asthma onset (ages 0-17)
¦/
¦/
PMISA
Asthma symptoms/exacerbation (6-17)
¦/
¦/
PMISA
Lung cancer (ages 30-99)
¦/
¦/
PMISA
Allergic rhinitis (hay fever) symptoms (ages 3-17)
¦/
¦/
PMISA
Lost work days (age 18-65)
¦/
¦/
PMISA
Minor restricted-activity days (age 18-65)
¦/
¦/
PMISA
Hospital admissions—Alzheimer's disease (ages 65-99)
¦/
¦/
PMISA
Hospital admissions—Parkinson's disease (ages 65-99)
¦/
¦/
PMISA
Other cardiovascular effects (e.g., other ages)
—
PM ISAb
Other respiratory effects (e.g., pulmonary function, non-
asthma ER visits, non-bronchitis chronic diseases, other ages
and populations)
—
—
PM ISAb
Other nervous system effects (e.g., autism, cognitive decline,
dementia)
—
—
PM ISAb
Metabolic effects (e.g., diabetes)
—
PM ISAb
Reproductive and developmental effects (e.g., low birth
weight, pre-term births, etc.)
—
—
PM ISAb
Cancer, mutagenicity, and genotoxicity effects
—
PM ISAb
Mortality from
exposure to
ozone
Premature respiratory mortality from short-term exposure (0-
99)
¦/
¦/
Ozone ISA
Premature respiratory mortality from long-term exposure
(age 30-99)
¦/
¦/
Ozone ISA
Nonfatal
morbidity from
exposure to
ozone
Hospital admissions—respiratory (ages 65-99)
¦/
¦/
Ozone ISA
Emergency department visits—respiratory (ages 0-99)
¦/
¦/
Ozone ISA
Asthma onset (0-17)
¦/
¦/
Ozone ISA
Asthma symptoms/exacerbation (asthmatics age 5-17)
¦/
¦/
Ozone ISA
Allergic rhinitis (hay fever) symptoms (ages 3-17)
¦/
¦/
Ozone ISA
Minor restricted-activity days (age 18-65)
¦/
¦/
Ozone ISA
School absence days (age 5-17)
¦/
¦/
Ozone ISA
Decreased outdoor worker productivity (age 18-65)
—
Ozone ISAb
Metabolic effects (e.g., diabetes)
—
Ozone ISAb
Other respiratory effects (e.g., premature aging of lungs)
—
Ozone ISAb
Cardiovascular and nervous system effects
—
Ozone ISAb
Reproductive and developmental effects
—
Ozone ISAb
a Valuation estimate excludes initial hospital and/or emergency department visits.
b Not quantified due to data availability limitations and/or because current evidence is only suggestive of causality.
379
-------
8,2.3 Calculating Counts of Air Pollution Effects Using the Health Impact Function
We use BenMAP-CE to quantify individual risk and counts of estimated premature deaths and
illnesses attributable to photochemical modeled changes in warm season average ozone
concentrations and annual mean PM2.5 for the year 2045 using a health impact function.6 A
health impact function combines information regarding the: concentration-response relationship
between air quality changes and the risk of a given adverse outcome; population exposed to the
air quality change; baseline rate of death or disease in that population; and, air pollution
concentration to which the population is exposed.
The following provides an example of a health impact function, in this case for PM2.5
mortality risk. We estimate counts of PIVh.s-related total deaths (yij) during each year i (i=l,.. .,1
where I is the total number of years analyzed) among adults aged 30 and older (a) in each county
in the contiguous U.S. j (j=l,... ,J where J is the total number of counties) as
yij— Ea yija
yija = moija x(ep-ACij-l) x pijaj Eq[l]
where moija is the baseline all-cause mortality rate for adults aged a=30-99 in county j in year
i stratified in 10-year age groups, P is the risk coefficient for all-cause mortality for adults
associated with annual average PM2.5 exposure, Cij is the annual mean PM2.5 concentration in
county j in year i, and Pija is the number of county adult residents aged a=30-99 in county j in
year i stratified into 5-year age groups.F
The BenMAP-CE tool is pre-loaded with projected population from the Woods & Poole
company; cause-specific and age-stratified death rates from the Centers for Disease Control and
Prevention, projected to future years; recent-year baseline rates of hospital admissions,
emergency department visits and other morbidity outcomes from the Healthcare Cost and
Utilization Program and other sources; concentration-response parameters from the published
epidemiologic literature cited in the ISAs for fine particles and ground-level ozone; and, cost of
illness or WTP unit values for each endpoint.
8.2.4 Quantifying Ozone-Attributable Premature Mortality
In 2008, the National Academies of Science (NAS) issued a series of recommendations to
EPA regarding the procedure for quantifying and valuing ozone-related mortality due to short-
term exposures.7 Chief among these was that"... short-term exposure to ambient ozone is likely
to contribute to premature deaths" and the committee recommended that "ozone-related
mortality be included in future estimates of the health benefits of reducing ozone exposures..."
The NAS also recommended that".. .the greatest emphasis be placed on the multicity and
[National Mortality and Morbidity Air Pollution Studies (NMMAPS)] ... studies without
exclusion of the meta-analyses."
F In this illustrative example, the air quality is resolved at the county level. For this RIA, we simulate air quality
concentrations at 12km by 12km grids. The BenMAP-CE tool assigns the rates of baseline death and disease stored
at the county level to the 12km by 12km grid cells using an area-weighted algorithm. This approach is described in
greater detail in the appendices to the BenMAP-CE user manual.
380
-------
Prior to the 2015 Ozone NAAQS RIA, the Agency estimated ozone-attributable premature
deaths using an NMMAPS-based analysis of total mortality, two multi-city studies of
cardiopulmonary and total mortality and effect estimates from three meta-analyses of non-
accidental mortality.8'9'10'11'12'13 Beginning with the 2015 Ozone NAAQS RIA, the Agency
began quantifying ozone-attributable premature deaths using two newer multi-city studies of
non-accidental mortality and one long-term cohort study of respiratory mortality. 14>15>16 The
2020 Ozone ISA included changes to the causality relationship determinations between short-
term exposures and total mortality, as well as including more recent epidemiologic analyses of
long-term exposure effects on respiratory mortality.17 In the final 2021 CSAPRRIA, mortality
from long-term exposures was estimated using the Turner et al. (2016) study extending and
expanding the analysis of the American Cancer Society cohort (ACS). Mortality for short-term
exposures was estimated using the risk estimate parameters from Zanobetti et al. (2008) and
Katsouyanni et al. (2009), which were pooled using a consistent ozone season (May-Sept) and
ozone metric (maximum daily 8-hour average).18'19'20
In this RIA, ozone-attributable respiratory deaths are also estimated using the risk estimate
parameters described in the final 2021 CSAPRRIA. However, instead of pooling results derived
from different ozone air quality surfaces, we have chosen to use only the risk estimates derived
from the Katsouyanni et al. (2009) study because the study includes more cities across the United
States. Furthermore, this analysis uses modeled ozone concentration data that matches the ozone
metric (maximum daily 1-hour average) and season (April-September) used by Katsouyanni et
al. (2009) rather than using a default ozone season (May-Sept) and metric (maximum daily 8-
hour average) used in the CSAPR RIA.
8.2,5 Quantifying PM2.5-Attributable Premature Mortality
When quantifying PM-attributable cases of adult mortality, we use risk estimates from two
epidemiology studies examining two large population cohorts: the American Cancer Society
cohort and the Medicare cohort.21'22 The 2019 PM ISA concluded that the analyses of the ACS
and Medicare cohorts provide strong evidence of an association between long-term PM2.5
exposure and premature mortality with support from additional cohort studies. Both the ACS and
Medicare cohort studies have separate and distinct attributes that make them well-suited to being
used in a PM benefits assessment, so we present PM2.5 related effects derived using relative risk
estimates from both cohorts.
The PM ISA, which was reviewed by the Clean Air Scientific Advisory Committee of EPA's
Science Advisory Board (SAB-CASAC), concluded that there is a causal relationship between
mortality and both long-term and short-term exposure to PM2.5 based on the entire body of
scientific evidence.23 The PM ISA also concluded that the scientific literature supports the use of
a no-threshold log-linear model to portray the PM-mortality concentration-response relationship
while recognizing potential uncertainty about the exact shape of the concentration-response
relationship. The 2019 PM ISA, which informed the setting of the 2020 PM NAAQS, reviewed
available studies that examined the potential for a population-level threshold to exist in the
concentration-response relationship. Based on such studies, the ISA concluded that "evidence
from recent studies reduce uncertainties related to potential co-pollutant confounding and
continues to provide strong support for a linear, no-threshold concentration-response
relationship".24 Consistent with this evidence, the Agency historically has estimated health
impacts above and below the prevailing NAAQS.25' 26>27>28>29> 30,31,32,33,34,35,36
381
-------
Following this approach, we report the estimated PIVk.s-related benefits (in terms of both
health impacts and monetized values) calculated using a log-linear concentration-response
function that quantifies risk from the full range of simulated PM2.5 exposures.37'38 When setting
the 2020 PM NAAQS, the EPA noted that:
. .an important consideration in characterizing the potential for additional public health
improvements associated with changes in PM2.5 exposure is whether concentration- response
relationships are linear across the range of concentrations or if nonlinear relationships exist along
any part of this range. Several recent studies examine this issue, and continue to provide
evidence of linear, no-threshold relationships between long-term PM2.5 exposures and all-cause
and cause-specific mortality.39 However, interpreting the shapes of these relationships,
particularly at PM2.5 concentrations near the lower end of the air quality distribution, can be
complicated by relatively low data density in the lower concentration range, the possible
influence of exposure measurement error, and variability among individuals with respect to air
pollution health effects [85 FR 82696],"40
Hence, we are most confident in the size of the risks estimated from simulated PM2.5
concentrations that coincide with the bulk of the observed PM concentrations in the
epidemiological studies that 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.
To give readers insight to the level of uncertainty in the estimated PM2.5 mortality benefits at
lower ambient concentrations, we report the estimated PM benefits as a distribution, identifying
points along this distribution corresponding to the Lowest Reported Levels (LRLs) of each long-
term exposure mortality study and the PM NAAQS (see Figure 8-1 below). In addition to adult
mortality discussed above, we use risk estimates from a multi-city study to estimate PM-related
infant mortality.41
8.3 Economic Valuation Methodology for Health Benefits
We next quantify the economic value of the ozone and PM2.5-related deaths and illnesses
estimated above. Changes in ambient concentrations of air pollution generally yield small
changes in the risk of future adverse health effects for many people. Therefore, the appropriate
economic measure is "willingness to pay" (WTP) for changes in risk of a health effect. For some
health effects, such as hospital admissions, WTP estimates are not generally available, so we use
the cost of treating or mitigating the effect. These cost-of-illness (COI) estimates are typically a
lower bound estimate of the true value of reducing the risk of a health effect because they reflect
the direct expenditures related to treatment, but not the value of avoided pain and suffering. The
unit values applied in this analysis are provided in Table 21 of the Estimating PM2.5- and Ozone-
Attributable Health Benefits TSD.
The estimated value of avoided premature deaths (PM2.5 plus ozone) account for between 84
percent or 94 percent of total monetized benefits depending on the studies used. The value for
the projected reduction in the risk of premature mortality is the subject of continuing discussion
within the economics and public policy analysis community. Following the advice of the SAB's
Environmental Economics Advisory Committee (SAB-EEAC), EPA currently uses the VSL
approach in calculating estimates of mortality benefits, because we believe this calculation
provides the most reasonable single estimate of an individual's willingness to trade off money
382
-------
for changes in the risk of death.42 The VSL approach is a summary measure for the value of
small changes in the risk of death experienced by a large number of people.
EPA continues work to update its guidance on valuing mortality risk reductions, and the
Agency consulted several times with the SAB-EEAC on this issue. Until updated guidance is
available, the Agency determined that a single, peer-reviewed estimate applied consistently, best
reflects the SAB-EEAC advice it has received. Therefore, EPA applies the VSL that was vetted
and endorsed by the SAB in the Guidelines for Preparing Economic Analyses while the Agency
continues its efforts to update its guidance on this issue.43 This approach calculates a mean value
across VSL estimates derived from 26 labor market and contingent valuation studies published
between 1974 and 1991. The mean VSL across these studies is $4.8 million (1990$). We then
adjust this VSL to account for the currency year and to account for income growth from 1990 to
the analysis year. Specifically, the VSL applied in this analysis in 2017$ after adjusting for
income growth is $11 million for 2045.
The Agency is committed to using scientifically sound, appropriately reviewed evidence in
valuing changes in the risk of premature death and continues to engage with the SAB to update
its mortality risk valuation estimates. In 2016, the Agency proposed new meta-analytic
approaches for updating its estimates, which were subsequently reviewed by the SAB-EEAC.44
EPA is taking the SAB's formal recommendations under advisement.
In valuing PIVh.s-related premature mortality, we discount the value of premature mortality
occurring in future years using rates of 3 percent and 7 percent.45 We assume that there is a
multi-year "cessation" lag between changes in PM exposures and the total realization of changes
in health effects. Although the structure of the lag is uncertain, EPA follows the advice of the
SAB-Health Effects Subcommittee (HES) to use a segmented lag structure that assumes 30
percent of premature deaths are reduced in the first year, 50 percent over years 2 to 5, and 20
percent over the years 6 to 20 after the reduction in PM2.5.46 Changes in the cessation lag
assumptions do not change the total number of estimated deaths but rather the timing of those
deaths.
Because short-term ozone-related premature mortality occurs within the analysis year, the
estimated ozone-related benefits are identical for all discount rates. When valuing changes in
ozone-attributable deaths using the Turner et al. (2016) study, we follow advice provided by the
SAB-HES, which found that".. .there is no evidence in the literature to support a different
cessation lag between ozone and particulate matter. The HES therefore recommends using the
same cessation lag structure and assumptions as for particulate matter when utilizing cohort
mortality evidence for ozone."47
These estimated health benefits do not account for the influence of future changes in the
climate on ambient concentrations of pollutants.48 For example, recent research suggests that
future changes to climate may create conditions more conducive to forming ozone; the influence
of changes in the climate on PM2.5 concentrations are less clear.49 The estimated health benefits
also do not consider the potential for climate-induced changes in temperature to modify the
relationship between ozone and the risk of premature death.50'51'52
383
-------
8.4 Characterizing Uncertainty in the Estimated Benefits
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. The health benefits
TSD that accompanied the final Revised Cross-State Air Pollution Rule (CSAPR) Update RIA
details our approach to characterizing uncertainty in both quantitative and qualitative terms. That
TSD describes the sources of uncertainty associated with key input parameters including
emissions 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 country
(i.e., regulations, technology, and human behavior). Each of these inputs is uncertain and affects
the size and distribution of the estimated benefits. When the uncertainties from each stage of the
analysis are compounded, even small uncertainties can have large effects on the total quantified
benefits.
To characterize uncertainty and variability in this assessment, we incorporate three
quantitative analyses, which are described in greater detail within the TSD (Sections 6.1 and
6.2):
• A Monte Carlo assessment that accounts for random sampling error and between study
variability in the epidemiological and economic valuation studies;
• The quantification of PM- and ozone-related mortality using alternative mortality effect
estimates drawn from different studies; and
• Presentation of 95th percentile confidence interval around each risk estimate.
Quantitative characterization of other sources of uncertainties are also discussed in Sections
6.1 and 6.2 of the TSD:
• For PM2.5-related adult all-cause mortality:
o The distributions of air quality concentrations experienced by the original cohort
population (TSD Section 6.1.2.1);
o Methods of estimating and assigning exposures in epidemiologic studies (TSD
Section 6.1.2.2);
o Confounding by ozone (TSD Section 6.1.2.3); and
o The statistical technique used to generate hazard ratios in the epidemiologic study
(TSD Section 6.1.2.4).
• For ozone-related mortality:
o Confounding by PM2.5 in the long-term ozone-attributable respiratory mortality
risk estimate (TSD Section 6.2.2.1);
o Potential threshold analysis in the short-term ozone-attributable respiratory
mortality risk estimate (TSD Section 6.2.2.2.1); and
o Confounding by PM2.5 in the short-term ozone-attributable respiratory mortality
risk estimate (TSD Section 6.2.2.2.2).
• Plausible alternative risk estimates for asthma onset in children (TSD Sections 6.1.3 and
6.2.4), cardiovascular hospital admissions (TSD Section 6.1.4,), and respiratory hospital
admissions (TSD Section 6.1.5)
• Effect modification of PM2.5- and ozone attributable health effects in at-risk populations
(TSD Sections 6.1.6 and 6.2.5).
384
-------
Quantitative consideration of baseline incidence rates and economic valuation estimates are
provided in Sections 6.3 and 6.4 of the TSD, respectively. Qualitative discussions of various
sources of uncertainty can be found in Section 6.5 of the TSD.
Below are key assumptions underlying the estimates for PIVh.s-related premature mortality,
followed by key uncertainties associated with estimating the number and value of ozone-related
premature deaths.
• 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 PM ISA, which
was reviewed by CASAC, concluded that "across exposure durations and health effects
categories ... the evidence does not indicate that any one source or component is
consistently more strongly related with health effects than PM2.5 mass."53
• We assume that the health impact function for fine particles is log-linear down to the
lowest air quality levels modeled in this analysis. Thus, the estimates include health
benefits from reducing fine particles in areas with varied concentrations of PM2.5,
including both regions that are in attainment with the fine particle standard and those that
do not meet the standard down to the lowest modeled concentrations. The PM ISA
concluded that "the majority of evidence continues to indicate a linear, no-threshold
concentration-response relationship for long-term exposure to PM2.5 and total
(nonaccidental) mortality."54
• 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,
which affects the valuation of mortality benefits at different discount rates. The above
assumptions are subject to uncertainty.55 Similarly, we assume there is a cessation lag
between the change in PM exposures and both the development and diagnosis of lung
cancer.
• We assume that there is no "cessation" lag between the change in ozone exposures and
the total realization of changes in long-term mortality effects. The 20-year segmented lag
for PM2.5 accounts for the onset of cardiovascular-related mortality, an outcome which is
not relevant to the long-term ozone respiratory mortality estimated here. There is no
alternative empirical estimate of the cessation lag for long-term exposure to ozone.
• We use a log-linear impact function without a threshold in modeling short-term ozone-
related mortality. However, we acknowledge reduced confidence in specifying the shape
of the concentration-response relationship in the range of < 40ppb and below.56 Thus, the
benefits estimates include health benefits from reducing ozone in areas with varied
concentrations of ozone down to the lowest modeled concentrations.
In this analysis, we plot estimated PM-related deaths according to where they occur along the
distribution of baseline PM2.5 annual mean concentrations (Figure 8-1). Displaying the data in
such a way allows readers to visualize the portion of population exposed to annual mean PM2.5
levels at or above different concentrations, which provides some insight into the level of
385
-------
uncertainty in the estimated PM2.5 mortality benefits. EPA does not view the level of the PM
NAAQS or the lowest concentration levels reported in the mortality studies as concentration
thresholds below which we would not quantify health benefits of air quality improvements.0
Rather, the PIVh.s-attributable benefits estimates reported in this RIA are the most appropriate
estimates because they reflect the full range of air quality concentrations associated with the
emission reduction program being evaluated. The 2019 PM ISA concluded that the scientific
evidence collectively is sufficient to conclude that there is a causal relationship between long-
term PM2.5 exposures and mortality and that overall, the studies support the use of a no-threshold
log-linear model to estimate mortality attributed to long-term PM2.5 exposure.
Figure 8-1 compares the percentage of the population and PM-related deaths to the annual
mean PM2.5 concentrations in the baseline for the year 2045. The figure identifies the LRL for
each of the major cohort studies and the annual mean PM2.5 NAAQS of 12 |ig/m\ For Turner et
al. (2016), the LRL is 2.8 |ig/m3 and for Di et al. (2017), the LRL is 0.02 |ig/m3.H As PM-related
mortality quantified using risk estimates from the Di et al. (2017) and Turner et al. (2016) are
within 5 percent of one another, in the interest of clarity and simplicity, we present the results
estimated using the risk estimate from Turner et al. (2016) alone in Figure 8-1. Additional
information on low concentration exposures in Turner et al. (2016) and Di et al. (2017) can be
found in section 6.1.2.1 of the Estimating PM2.5- and Ozone-Attributable Health Benefits TSD.
The air quality modeling predicts PM2.5 concentrations to be at or below the level of the annual
PM2.5 NAAQS (12 |ig/m3) in most locations in 2045. As noted in RIA Chapter 6.4.2, we are
more confident in the projected changes in annual mean PM2.5 concentrations than we are in the
projected absolute PM2.5 concentrations in 2045.
G For a summary of the scientific review statements regarding the lack of a threshold in the PM2 5-mortality
relationship, see the TSD entitled Summary of Expert Opinions on the Existence of a Threshold in the
Concentration-Response Function for PM2.5-related Mortality (U.S. EPA, 2010).
H Turner et al. (2016) estimated PM25 exposures using both a hierarchical Bayesian space-time model (HBM) and a
land use regression model with Bayesian Maximum Entropy kriging of residuals (LURBME). As such, two LRLs
are reported in the paper, 2.8 |ig/m3 and 1.4 |ig/m3. As the HBM risk estimate was used in the final 2021 CSAPR
RIA, the HBM LRL is presented here.
386
-------
LRL
Di (2017)
LRL
Turner (2016)
2012
PM;s Annual naaqs
0.2
Attributable Deaths
0 > ¦ ¦
1 2 3 A 5 6 7 S 9 10 11 12 13 14 15
2045 Baseline PM2 5 Concentration (jig/m3)
Figure 8-1: Estimated Percentage of PIVh.s-Related Deaths (Turner et al. 2016) and Number of Individuals
Exposed (30+) by Annual Mean PM2.5 Level in 2045
8.5 Estimated Number and Economic Value of Health Benefits
Below we report the estimated number of reduced premature deaths and illnesses in 2045
attributable to the standards along with the 95 percent confidence interval (Table 8-3 and Table
8-4). The number of reduced estimated deaths and illnesses are calculated from the sum of
individual reduced mortality and illness risk across the population. Table 8-5 reports the
estimated individual economic value of avoided premature deaths and illnesses relative to the
baseline along with the 95 percent confidence interval. Error! Reference source not found.
Table 8-6 reports total benefits associated with the standards in 2045, reflecting alternative
combinations of the economic value of PM2.5- and ozone-related premature deaths summed with
the economic value of illnesses for each discount rate.
387
-------
Table 8-3: Estimated Avoided PM2.5 Mortality and Illnesses in 2045 (95% Confidence Interval) a'b
Avoided Health Incidence
Avoided premature mortality
Turner et al. (2016) - Ages 30+
740
(500 to 980)
Di et al. (2017) - Ages 65+
800
(780 to 830)
Woodruff et al. (2008) - Ages < 1
4.1
(-2.6 to 11)
Non-fatal heart attacks among adults
Short-term exposure
Peters et al. (2001)
790
(180 to 1,400)
Pooled estimate
85
(31 to 230)
Morbidity effects
Long-term exposure
Asthma onset
1,600
(1,500 to 1,600)
Allergic rhinitis symptoms
10,000
(2,500 to 18,000)
Stroke
41
(11 to 70)
Lung cancer
52
(16 to 86)
Hospital Admissions - Alzheimer's disease
400
(300 to 500)
Hospital Admissions - Parkinson's disease
43
(22 to 63)
Short-term exposure
Hospital admissions-cardiovascular
110
(76 to 130)
ED visits- cardiovascular
210
(-82 to 500)
Hospital admissions - respiratory
68
(23 to 110)
ED visits - respiratory
400
(78 to 830)
Asthma symptoms
210,000
(-100,000 to 520,000)
Minor restricted-activity days
460,000
(370,000 to 550,000)
Cardiac arrest
10
(-4.2 to 24)
Lost work days
78,000
(66,000 to 90,000)
a Values rounded to two significant figures.
b PM2.5 exposure metrics are not presented here because all PM health endpoints are based on studies that used
daily 24-hour average concentrations. Annual exposures are estimated using daily 24-hour average concentrations.
388
-------
Table 8-4: Estimated Avoided Ozone Mortality and Illnesses in 2045 (95% Confidence Interval)3
Metric and Seasonb
Avoided Health
Incidence
Avoided premature mortality
Long-term
exposure
Turner et al. (2016)
MDA8
April-September
2,100
(1,400 to 2,700)
Short-term
exposure
Katsouyanni et al (2009)
MDA1
April-September
120
(-69 to 300)
Morbidity effects
Long-term
exposure
Asthma onset0
MDA8
June-August
16,000
(14,000 to 18,000)
Short-term
exposure
Allergic rhinitis symptoms
MDA8
May-September
88,000
(47,000 to 130,000)
Hospital admissions - respiratory
MDA1
April-September
350
(-91 to 770)
ED visits - respiratory
MDA8
May-September
5,100
(1,400 to 11,000)
Asthma symptoms - Coughd
MDA8
May-September
920,000
(-50,000 to 1,800,000)
Asthma symptoms - Chest Tightnessd
MDA8
May-September
770,000
(85,000 to 1,400,000)
Asthma symptoms - Shortness of Breathd
MDA8
May-September
390,000
(-330,000 to 1,100,000)
Asthma symptoms - Wheezed
MDA8
May-September
730,000
(-57,000 to 1,500,000)
Minor restricted-activity daysd
MDA1
May-September
1,600,000
(650,000 to 2,600,000)
School absence days
MDA8
May-September
1,100,000
(-150,000 to 2,200,000)
a Values rounded to two significant figures.
b MDA8 - maximum daily 8-hour average; MDA1 - maximum daily 1-hour average. Studies of ozone vary with
regards to season, limiting analyses to various definitions of summer (e.g., April-September, May-September or
June-August). These differences can reflect state-specific ozone seasons, EPA-defined seasons or another seasonal
definition chosen by the study author. The paucity of ozone monitoring data in winter months complicates the
development of full year projected ozone surfaces and limits our analysis to only warm seasons.
0 The underlying metric associated with this risk estimate is daily 8-hour average from 10am - 6pm (AVG8);
however, we ran the study with a risk estimate converted to MDA8.
d Applied risk estimate derived from full year exposures to estimates of ozone across a May-September ozone
season. When risk estimates based on full-year, long-term ozone exposures are applied to warm season air quality
projections, the resulting benefits assessment may underestimate impacts, due to a shorter timespan for impacts to
accrue.
389
-------
Table 8-5: Estimated Economic Value of PM2.5- and Ozone-Attributable Premature Mortality and Illnesses in
2045 (95% Confidence Interval; millions of 2017$)a
3% Discount Rate
7% Discount Rate
Avoided premature mortality
1/-1
-------
Table 8-6: Total Ozone and PIVb.s-Attributable Benefits in 2045 (95% Confidence Interval; billions of
2017$)a'b
Total Annual Benefits in 2045
3% Discount Rate
$12 $33
($0.72 to $31)° ($3.5 to $87)d
7% Discount Rate
$10 $30
($0.37 to $28)c ($3.0 to $78)d
a The benefits associated with the standards presented here do not include the Ml
complement of health and environmental benefits that, if quantified and monetized,
would increase the total monetized benefits.
b Values rounded to two significant figures. The two benefits estimates separated by
the word "and" signify that they are two separate estimates. The estimates do not
represent lower- and upper-bound estimates though they do reflect a grouping of
estimates that yield more and less conservative benefit totals. They should not be
summed.
0 Sum of benefits using the Katsouyanni et al. (2009) short-term exposure ozone
respiratory mortality risk estimate and the Turner et al. (2016) long-term exposure
PM2 5 all-cause risk estimate.
d Sum of benefits using the Turner et al. (2016) long-term exposure ozone
respiratory mortality risk estimate and the Di et al. (2017) long-term exposure PM2 5
all-cause risk estimate.
8.6 Present Value of Total Benefits
The full-scale benefits analysis reflects spatially and temporally allocated emissions
inventories generated using SMOKE/MOVES (see RIA Chapter 5), photochemical air quality
modeling using CMAQ (see RIA Chapter 6), and PM2.5 and ozone benefits generated using
BenMAP-CE, all for conditions projected to occur in calendar year 2045. As we presented in
RIA Chapter 5 and Chapter 7, national estimates of year-over-year emissions and program costs
were generated from program implementation to a year when the program will be fully phased-in
and the vehicle fleet will be approaching full turnover (2027-2045). The time and resources
required to conduct air quality modeling to support a full-scale benefits analysis for all analysis
years from 2027 to 2044 precluded the Agency from conducting benefits analyses comparable to
the calendar year 2045 benefits analysis. Instead, we have used a reduced-form approach to scale
total benefits in 2045 back to 2027 (including interim years) using projected reductions in year-
over-year NOx emissions so that we can estimate the present value of the stream of estimated
benefits
This approach is similar to the Agency's method for estimating "benefits-per-ton" values over
time.57 For interim analysis years without air quality modeling, we input the program's 2045 air
quality data into BenMAP-CE to generate benefits that occur in earlier analysis years. This
approach allows us to calculate the benefits for interim years by adjusting for changes in
population, baseline mortality incidence, and income growth overtime. Table 8-7 displays the
data used to generate benefits that reflect input data for years 2027, 2030, 2035, 2040, and 2045.1
1 Interim analysis years chosen for computational efficiency at reasonable intervals.
391
-------
Table 8-7: Benefits Inputs that Change Over Time used to Calculate Year-over-Year Estimates
Analysis Year
Air Quality
Modeling &
Emissions Year
Population Year
Baseline
Mortality
Incidence Year
Income Growth
Year
Currency Year
2027
2045
2027
2025
2027
2017
2030
2030
2030
2030
2035
2035
2035
2035
2040
2040
2040
2040
2045
2045
2045
2045
We next calculate the total monetized benefits estimated for each of the analysis years and
divide them by the estimated tons of NOx emissions projected to be controlled by the rule in
2045 (see RIA Chapter 5, Table 5-29) to generate "benefit-per-ton" values that reflect benefits
inputs consistent with the analysis year.J Because NOx is the dominant pollutant controlled by
the program, we make a simplifying assumption that total PM and ozone benefits can be scaled
by NOx emissions, even though emissions of other pollutants are controlled in smaller amounts
by the rule (see RIA Chapter 5, Table 5-21). By using the 2045 air quality modeling surfaces for
the earlier analysis years, we also assume that the spatial distribution of NOx emissions
reductions does not change over time. While there may be localized differences in the rate of
fleet turnover due to state or local incentive programs, we do not currently have sufficient data to
incorporate those differences into our analyses and believe that they would generally even out
over time (as noted in RIA Chapter 5, we use MOVES default vehicle activity data, including
data on age of the fleet or turnover).
To estimate total benefits for the interim years, we multiply the benefit-per-ton values
estimated for each earlier analysis year by the NOx emissions projected to be controlled in that
same year (2027, 2030, 2035, and 2040; see RIA Chapter 5, Table 5-31). For intervening years
between the analysis years, we linearly interpolate total benefits.
Table 8-8 and Table 8-9 present the undiscounted stream of scaled annual total benefits of the
rule between 2027 and 2045. We also estimate the present value and annualized value of the
stream of benefits in these tables. Table 8-8 presents total benefits as the sum of short-term
ozone respiratory mortality benefits for all ages, long-term PM2.5 all-cause mortality benefits for
ages 30 and above, and all monetized avoided illnesses.58'59 Table 8-9 presents total benefits as
the sum of long-term ozone respiratory mortality benefits for ages 30 and above, long-term PM2.5
all-cause mortality benefits for ages 65 and above, and all monetized avoided illnesses.60'61 The
present value of benefits in both tables is discounted back to year 2027 using both a 3 percent
and 7 percent discount rate.
1 Note that these "benefit-per-ton" values are internally consistent with the air quality modeling conducted for the
proposal in 2045. They are appropriate for scaling benefits of the program but should not be used outside of the
context of this rulemaking analysis.
392
-------
Table 8-8: Undiscounted Stream and Present Value of Human Health Benefits from 2027 through 2045:
Monetized Benefits Quantified as Sum of Short-Term Ozone Respiratory Mortality Ages 0-99, and Long-
Term PM2.5 All-Cause Mortality Ages 30+ (Discounted at 3% and 7%; billions of 2017$)a b
Monetized Benefits
3% Discount
7% Discount
2027
$0.66
$0.59
2028
$1.4
$1.2
2029
$2.1
$1.9
2030
$2.8
$2.6
2031
$3.8
$3.4
2032
$4.8
$4.3
2033
$5.5
$5.0
2034
$6.2
$5.6
2035
$6.9
$6.2
2036
$7.5
$6.7
2037
$8.0
$7.2
2038
$8.6
$7.7
2039
$9.1
$8.2
2040
$9.6
$8.7
2041
$10
$9.0
2042
$10
$9.4
2043
$11
$9.7
2044
$11
$10
2045°
$12
$10
Present Value
$91
$53
Annualized Value
$6.3
$5.1
a The benefits associated with the standards presented here do not include the Ml complement of health and
environmental benefits that, if quantified and monetized, would increase the total monetized benefits.
b Benefits calculated as value of avoided: PM25-attributable deaths (quantified using a concentration-response
relationship from the Turner et al. 2016 study); Ozone-attributable deaths (quantified using a concentration-
response relationship from the Katsouyanni et al. 2009 study); and PM2 5 and ozone-related morbidity effects.
0 Year in which PM2 5 and ozone air quality was simulated (2045).
393
-------
Table 8-9: Undiscounted Stream and Present Value of Human Health Benefits from 2027 through 2045:
Monetized Benefits Quantified as Sum of Long-Term Ozone Respiratory Mortality Ages 30+, and Long-Term
PM2.5 All-Cause Mortality Ages 65+ (Discounted at 3% and 7%; billions of 2017$)a'b
Monetized Benefits
3% Discount
7% Discount
2027
$1.8
$1.6
2028
$3.7
$3.3
2029
$5.7
$5.1
2030
$7.9
$7.1
2031
$11
$9.6
2032
$13
$12
2033
$16
$14
2034
$18
$16
2035
$19
$17
2036
$21
$19
2037
$23
$21
2038
$25
$22
2039
$26
$23
2040
$28
$25
2041
$29
$26
2042
$30
$27
2043
$31
$28
2044
$32
$29
2045°
$33
$30
Present Value
$260
$150
Annualized Value
$18
$14
a The benefits associated with the standards presented here do not include the full complement of health and
environmental benefits that, if quantified and monetized, would increase the total monetized benefits.
b Benefits calculated as value of avoided: PM25-attributable deaths (quantified using a concentration-response
relationship from the Turner et al. 2016 study); Ozone-attributable deaths (quantified using a concentration-
response relationship from the Katsouyanni et al. 2009 study); and PM2 5 and ozone-related morbidity effects.
0 Year in which PM2 5 and ozone air quality was simulated (2045).
8.7 Unquantified Benefits
In addition to the PM2.5 and ozone-related health impacts we are unable to quantify or
monetize in Table 8-2, there are additional benefits associated with reductions in exposure to
ambient concentrations ofNC>2,K ecosystem benefits, and visibility improvement that EPA is not
currently able to quantify due to data, resource, or methodological limitations. EPA continues to
pursue data and methods to further improve our assessment of benefits that are currently
unquantified. In particular, we are evaluating the feasibility of assessing impacts on ecosystem
K EPA is considering how to incorporate NO2 health benefits into our rulemakings. The ISA states that a key
uncertainty in understanding the relationship between non-respiratory health effects and short- or long-term
exposure to NO2 is co-pollutant confounding, particularly by other roadway pollutants. The Agency will utilize the
same systematic process for selecting, quantifying, and monetizing NCh-related health impacts as it has for PM25
and ozone. This process includes: applying a criteria for identifying and selecting studies and risk estimates most
appropriate to inform a benefits analysis for a RIA; identifying pollutant-attributable health effects for which the
ISA reports strong evidence and that may be quantified in a benefits assessment; collecting baseline incidence and
prevalence estimates and demographic information; developing appropriate economic unit values, and
characterizing uncertainty with quantified benefits estimates.
394
-------
services from reductions in nitrogen deposition and terrestrial acidification. RIA Chapter 4
provides a qualitative description of both the health and environmental effects of the criteria
pollutants controlled by the program. These additional unquantified health and welfare benefit
categories are listed in Table 8-10.
There will also be benefits associated with reductions in air toxic pollutant emissions that
result from the program (see RIA Chapter 4.1.6 and RIA Chapter 5.3.1), but we did not attempt
to monetize those impacts. This is 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 estimation or benefits assessment. While EPA has worked to improve these tools, there
remain critical limitations for estimating incidence and assessing benefits of reducing mobile
source air toxics.
395
-------
Table 8-10: Unquantified Criteria Pollutant Health and Welfare Benefits Categories
Category
Effect
Effect
Quantified
Effect
Monetized
More
Information
Improved Human Health
Asthma hospital admissions
—
—
NO2 ISA62'3
Chronic lung disease hospital admissions
—
—
NO2 ISA3
Respiratory emergency department visits
—
—
NO2 ISA3
Reduced incidence
Asthma exacerbation
—
—
NO2 ISA3
of morbidity from
Acute respiratory symptoms
—
—
NO2 ISA3
exposure to NO2
Premature mortality
—
—
NO2 ISA3b-c
Other respiratory effects (e.g., airway
hyperresponsiveness and inflammation, lung
function, other ages and populations)
—
—
NO2 ISAbc
Improved Environment
Reduced visibility
Visibility in Class 1 areas
—
—
PM ISA3
impairment
Visibility in residential areas
—
—
PM ISA3
Reduced effects on
Household soiling
—
—
PM ISA3b
materials
Materials damage (e.g., corrosion, increased wear)
—
—
PM ISAb
Reduced effects from
PM deposition
(metals and organics)
Effects on individual organisms and ecosystems
—
—
PM ISAb
Visible foliar injury on vegetation
—
—
Ozone ISA3
Reduced vegetation growth and reproduction
—
—
Ozone ISA3
Yield and quality of commercial forest products and
Ozone ISA3
Reduced vegetation
and ecosystem
effects from
exposure to ozone
Damage to urban ornamental plants
—
Ozone ISAb
Carbon sequestration in terrestrial ecosystems
—
Ozone ISA3
Recreational demand associated with forest aesthetics
—
Ozone ISAb
Other non-use effects
Ozone ISAb
Ecosystem functions (e.g., water cycling,
biogeochemical cycles, net primary productivity,
leaf-gas exchange, community composition)
—
—
Ozone ISAb
Recreational fishing
—
NOx SOx ISA63'3
Tree mortality and decline
—
NOx SOx ISAb
Reduced effects from
acid deposition
Commercial fishing and forestry effects
—
NOx SOx ISAb
Recreational demand in terrestrial and aquatic
ecosystems
—
—
NOx SOx ISAb
Other non-use effects
NOx SOx ISAb
Ecosystem functions (e.g., biogeochemical cycles)
—
NOx SOx ISAb
Species composition and biodiversity in terrestrial
and estuarine ecosystems
—
—
NOx SOx ISAb
Coastal eutrophication
—
NOx SOx ISAb
Reduced effects from
nutrient enrichment
Recreational demand in terrestrial and estuarine
ecosystems
—
—
NOx SOx ISAb
Other non-use effects
NOx SOx ISAb
Ecosystem functions (e.g., biogeochemical cycles,
fire regulation)
—
—
NOx SOx ISAb
Reduced vegetation
Injury to vegetation from SO2 exposure
—
NOx SOx ISAb
effects from ambient
exposure to SO2 and
NOx
Injury to vegetation from NOx exposure
—
NOx SOx ISAb
3 We assess these benefits qualitatively due to data and resource limitations for this RIA.
b We assess these benefits qualitatively because we do not have sufficient confidence in available data or methods.
0 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.
396
-------
Chapter 8 References
1 U.S. Environmental Protection Agency (U.S. EPA). 2021. Regulatory Impact Analysis for the
Final Revised Cross-State Air Pollution Rule (CSAPR) Update for the 2008 Ozone NAAQS.
EPA-452/R-21-002. March.
2 U.S. Environmental Protection Agency (U.S. EPA). 2021. Estimating PM2.5- and Ozone-
Attributable Health Benefits. Technical Support Document (TSD) for the Final Revised Cross-
State Air Pollution Rule Update for the 2008 Ozone Season NAAQS. EPA-HQ-OAR-2020-
0272. March, https://www.epa.gov/sites/default/files/2021-03/documents/estimating_pm2.5-
_and_ozone-attributable_health_benefits_tsd.pdf.
3 U.S. Environmental Protection Agency (U.S. EPA). 2018. Modeling Guidance for
Demonstrating Air Quality Goals for Ozone, PM2.5, and Regional Haze. EPA 454/R-l 8-009.
https://www.epa.gov/sites/production/files/2020-10/documents/o3-pm-rh-modeling_guidance-
2018.pdf.
4 U.S. Environmental Protection Agency (U.S. EPA). 2015. Regulatory Impact Analysis of the
Final Revisions to the National Ambient Air Quality Standards for Ground-Level Ozone, EPA-
452/R-15-07. EPA-452/R-12-003. Office of Air Quality Planning and Standards, Health and
Environmental Impacts Division, Research Triangle Park, NC.
(https://www.epa.gOv/sites/production/files/2016-02/documents/20151001ria.pdf).
5 U.S. Environmental Protection Agency (U.S. EPA). 2012. Regulatory Impact Analysis for the
Final Revisions to the National Ambient Air Quality Standards for Particulate Matter. EPA-
452/R-12-003. Office of Air Quality Planning and Standards, Health and Environmental Impacts
Division, Research Triangle Park, NC. Available at: <
https://www3.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf>.
6 U.S. Environmental Protection Agency (U.S. EPA). 2018. Environmental Benefits Mapping
and Analysis Program - Community Edition. User's Manual. Office of Air Quality Planning and
Standards, Health and Environmental Impacts Division, Research Triangle Park, NC. Available
at: < https://www.epa.gov/sites/production/files/2015-04/documents/benmap-
ce_user_manual_march_2015 ,pdf>.
7 National Research Council (NRC). 2008. Estimating Mortality Risk Reduction and Economic
Benefits from Controlling Ozone Air Pollution. National Academies Press. Washington, DC.
8 Bell, M.L., A. McDermott, S.L. Zeger, J.M. Sarnet, and F. Dominici. 2004. "Ozone and Short-
Term Mortality in 95 U.S. Urban Communities, 1987-2000." Journal of the American Medical
Association. 292(19):2372-8.
9 Huang Y, Dominici F, Bell M. 2004. Bayesian Hierarchical Distributed Lag Models for
Summer Ozone Exposure and Cardio-Respiratory Mortality. Johns Hopkins Univ Dept Biostat
Work Pap Ser.
10 Schwartz, J. 2005. "How Sensitive is the Association between Ozone and Daily Deaths to
Control for Temperature?" American Journal of Respiratory and Critical Care Medicine. 171(6):
627-31.
397
-------
11 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):436-45.
12 Ito, K., S.F. De Leon, and M. Lippmann. 2005. "Associations Between Ozone and Daily
Mortality: Analysis and Meta-Analysis." Epidemiology. 16(4):446-57.
13 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.
14 Smith RL, Xu B, Switzer P. 2009. Reassessing the relationship between ozone and short-
term mortality in U.S. urban communities. Inhal Toxicol 21 Suppl 2:37-61;
doi:10.1080/08958370903161612.
15 Zanobetti A, Schwartz J. 2008. Mortality displacement in the association of ozone with
mortality: an analysis of 48 cities in the United States. Am J Respir Crit Care Med 177:184-9;
doi: 10.1164/rccm.200706-8230C.
16 Jerrett M, Burnett RT, Pope CA, Ito K, Thurston G, Krewski D, et al. 2009. Long-term ozone
exposure and mortality. NEngl J Med 360:1085-95; doi:10.1056/NEJMoa0803894.
17 U.S. Environmental Protection Agency (U.S. EPA). 2020. Integrated Science Assessment
(ISA) for Ozone and Related Photochemical Oxidants (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-20/012, 2020.
18 Turner, MC, Jerrett, M, Pope, A, III, Krewski, D, Gapstur, SM, Diver, WR, Beckerman, BS,
Marshall, JD, Su, J, Crouse, DL and Burnett, RT (2016). Long-term ozone exposure and
mortality in a large prospective study. Am J Respir Crit Care Med 193(10): 1134-1142.
19 Zanobetti A, Schwartz J. 2008. Mortality displacement in the association of ozone with
mortality: an analysis of 48 cities in the United States. Am J Respir Crit Care Med 177:184-9;
doi: 10.1164/rccm.200706-8230C.
20 Katsouyanni, K, Samet, JM, Anderson, HR, Atkinson, R, Le Tertre, A, Medina, S, Samoli, E,
Touloumi, G, Burnett, RT, Krewski, D, Ramsay, T, Dominici, F, Peng, RD, Schwartz, J,
Zanobetti, A and Committee, HEIHR (2009). Air pollution and health: a European and North
American approach (APHENA). Res Rep Health Eff Inst(142): 5-90.
21 Turner, MC, Jerrett, M, Pope, A, III, Krewski, D, Gapstur, SM, Diver, WR, Beckerman, BS,
Marshall, JD, Su, J, Crouse, DL and Burnett, RT (2016). Long-term ozone exposure and
mortality in a large prospective study. Am J Respir Crit Care Med 193(10): 1134-1142.
22 Di, Q, Wang, Y, Zanobetti, A, Wang, Y, Koutrakis, P, Choirat, C, Dominici, F and Schwartz,
JD (2017). Air pollution and mortality in the Medicare population. New Engl J Med 376(26):
2513-2522.
23 U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2019.
Letter from Louis Anthony Cox, Jr., Chair, Clean Air Scientific Advisory Committee, to
Administrator Andrew R. Wheeler. Re: CASAC Review of the EPA's Integrated Science
Assessment for Particulate Matter (External Review Draft - October 2018). April 11, 2019. EPA-
CASAC-19-002. Office of the Administrator, Science Advisory Board U.S. EPA HQ,
398
-------
Washington DC. Available at:
https://casac.epa. gov/ords/sab/f?p=l 05:12:17077188731556::: 12::.
24 U.S. Environmental Protection Agency (U.S. EPA). 2019. Integrated Science Assessment
(ISA) for Particulate Matter (Final Report, 2019). U.S. Environmental Protection Agency,
Washington, DC, EPA/600/R-19/188, 2019.
25 U.S. Environmental Protection Agency (U.S. EPA). 2010. Regulatory Impact Analysis (RIA)
for Existing Stationary Compression Ignition Engines NESHAP Final Draft.
26 U.S. EPA. Greenhouse gas emissions standards and fuel efficiency standards for medium- and
heavy-duty engines and vehicles. Final rules. Fed Reg. 2011 ;76(179):57106—57513.
27 U.S. Environmental Protection Agency (U.S. EPA). 2011. Regulatory Impact Analysis for the
Final Mercury and Air Toxics Standards.
28 U.S. Environmental Protection Agency (U.S. EPA). 2012. Regulatory Impact Analysis for
the Final Revisions to the National Ambient Air Quality Standards for Particulate Matter. EPA-
452/R-12-003. Office of Air Quality Planning and Standards, Health and Environmental Impacts
Division, Research Triangle Park, NC. Available at: <
https://www3.epa.gov/ttnecasl/regdata/RIAs/finalria.pdf>.
29 U.S. Environmental Protection Agency (U.S. EPA). 2013. Regulatory Impact Analysis for the
Final Revisions to the National Ambient Air Quality Standards for Particulate Matter.
30 U.S. Environmental Protection Agency (U.S. EPA). 2014. Regulatory Impact Analysis for
the Proposed Carbon Pollution Guidelines for Existing Power Plants and Emission Standards for
Modified and Reconstructed Power Plants.
31 U.S. Environmental Protection Agency (U.S. EPA). 2014. Regulatory Impact Analysis of the
Proposed Revisions to the National Ambient Air Quality Standards for Ground-Level Ozone.
32 U.S. Environmental Protection Agency (U.S. EPA). 2015. Regulatory Impact Analysis for
Residential Wood Heaters NSPS Revision: Final Report.
33 U.S. Environmental Protection Agency (U.S. EPA). 2015. Regulatory Impact Analysis for
the Clean Power Plan Final Rule.
34 U.S. Environmental Protection Agency (U.S. EPA). 2015. Regulatory Impact Analysis for
the Proposed Federal Plan Requirements for Greenhouse Gas Emissions from Electric Utility
Generating Units Constructed on or Before January 8, 2014; Model Trading Rules; Amendments
to Framework Regulations. 4-52
35 U.S. Environmental Protection Agency (U.S. EPA). 2015. Regulatory Impact Analysis of the
Final Revisions to the National Ambient Air Quality Standards for Ground-Level Ozone, EPA-
452/R-15-07. EPA-452/R-12-003. Office of Air Quality Planning and Standards, Health and
Environmental Impacts Division, Research Triangle Park, NC.
(https://www.epa.gOv/sites/production/files/2016-02/documents/20151001ria.pdf).
399
-------
36 U.S. Environmental Protection Agency (U.S. EPA). 2016. Regulatory Impact Analysis of the
Cross-State Air Pollution Rule (CSAPR) Update for the 2008 National Ambient Air Quality
Standards for Ground-Level Ozone.
37 National Research Council (NRC). 2002. Estimating the Public Health Benefits of Proposed
Air Pollution Regulations. National Academies Press. Washington, DC.
38 U.S. Environmental Protection Agency (U.S. EPA). 2009. Integrated Science Assessment for
Particulate Matter (Final Report). EPA-600-R-08-139F. National Center for Environmental
Assessment - RTP Division, Research Triangle Park, NC. December. Available at:
.
39 U.S. Environmental Protection Agency (U.S. EPA). 2009. Integrated Science Assessment for
Particulate Matter (Final Report). EPA-600-R-08-139F. National Center for Environmental
Assessment - RTP Division, Research Triangle Park, NC. December. Available at:
. Section 11.2.4
40 Review of the National Ambient Air Quality Standards for Particulate Matter, Federal
Register, 85 FR 82696, December 18, 2020.
41 Woodruff, T.J., J. Grillo, and K.C. Schoendorf. 2008. "Air pollution and postneonatal infant
mortality in the United States, 1999-2002." Environmental Health Perspectives. 116(1): 110-
115.
42 U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2000.
An SAB Report on EPA's White Paper Valuing the Benefits of Fatal Cancer Risk Reduction.
EPA-SAB-EEAC-00-013. July.
43 U.S. Environmental Protection Agency (U.S. EPA). 2016. Guidelines for Preparing
Economic Analyses.
44 U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2017.
Letter from Peter S. Thorne, Chair, Science Advisory Board, and Madhu Khanna, Chair, SAB
Environmental Economics Advisory Committee, to Administrator Scott Pruitt. Re: SAB Review
of EPA's Proposed Methodology for Updating Mortality Risk Valuation Estimates for Policy
Analysis. February 23, 2017. EPA-SAB-17-005. Office of the Administrator, Science Advisory
Board U.S. EPA HQ, Washington DC. Available at:
https://cfpub.epa. gov/si/si_public_file_download.cfm?p_download_id=532807.
45 United States Office of Management and Budget (OMB). 2003. Circular A-4: Regulatory
Analysis. Washington, DC. Available at: <
https://obamawhitehouse.archives.gov/omb/circulars_a004_a-4/ >.
46 U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2005.
Advisory Council on Clean Air Compliance Analysis Response to Agency Request on Cessation
Lag. EPA-COUNCIL-LTR-05-001. December. Available at: <
https://council.epa.gov/ords/sab/f?p=l 04:12:968651521971>.
400
-------
47 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
CAA
48 USGCRP. 2016. The Impacts of Climate Change on Human Health in the United States: A
Scientific Assessment.; http://dx.doi.org/10.7930/J0R49NQX.
49 Fann N, Nolte CG, Dolwick P, Spero TL, Brown AC, Phillips S, et al. 2015. The geographic
distribution and economic value of climate change-related ozone health impacts in the United
States in 2030. J Air Waste Manag Assoc 65; doi:10.1080/10962247.2014.996270.
50 Jhun I, Fann N, Zanobetti A, Hubbell B. 2014. Effect modification of ozone-related mortality
risks by temperature in 97 US cities. Environment International. 73:128-34.
51 Ren, C., G.M. Williams, K. Mengersen, L. Morawska, and S. Tong. 2008. "Does
Temperature Modify Short-Term Effects of Ozone on Total Mortality in 60 Large Eastern U.S.
Communities? An Assessment Using the NMMAPS Data." Environment International. 34:451-
458.
52 Ren, C., G.M. William, L. Morawska, K. Mengensen, and S. Tong. 2008. "Ozone Modifies
Associations between Temperature and Cardiovascular Mortality: Analysis of the NMMAPS
Data." Occupational and Environmental Medicine. 65:255-260.
53 U.S. Environmental Protection Agency (U.S. EPA). 2019. Integrated Science Assessment
(ISA) for Particulate Matter (Final Report, 2019). U.S. Environmental Protection Agency,
Washington, DC, EPA/600/R-19/188, 2019.
54 U.S. Environmental Protection Agency (U.S. EPA). 2019. Integrated Science Assessment
(ISA) for Particulate Matter (Final Report, 2019). U.S. Environmental Protection Agency,
Washington, DC, EPA/600/R-19/188, 2019.
55 U.S. Environmental Protection Agency—Science Advisory Board (U.S. EPA-SAB). 2005.
Advisory Council on Clean Air Compliance Analysis Response to Agency Request on Cessation
Lag. EPA-COUNCIL-LTR-05-001. December. Available at: <
https://council.epa.gov/ords/sab/f?p=l 04:12:968651521971>.
56 U.S. Environmental Protection Agency (U.S. EPA). 2020. Integrated Science Assessment
(ISA) for Ozone and Related Photochemical Oxidants (Final Report). U.S. Environmental
Protection Agency, Washington, DC, EPA/600/R-20/012, 2020. Section 6.2.6
57 U.S. Environmental Protection Agency (U.S. EPA). 2018. Technical Support Document:
Estimating the Benefit per Ton of Reducing PM2.5 Precursors from 17 Sectors. February.
58 Turner, MC, Jerrett, M, Pope, A, III, Krewski, D, Gapstur, SM, Diver, WR, Beckerman, BS,
Marshall, JD, Su, J, Crouse, DL and Burnett, RT (2016). Long-term ozone exposure and
mortality in a large prospective study. Am J Respir Crit Care Med 193(10): 1134-1142.
59 Di Q, et al. (2017) Air pollution and mortality in the Medicare population. N Engl J Med
376:2513-2522.
401
-------
60 Katsouyanni, K, Samet, JM, Anderson, HR, Atkinson, R, Le Tertre, A, Medina, S, Samoli,
E, Touloumi, G, Burnett, RT, Krewski, D, Ramsay, T, Dominici, F, Peng, RD, Schwartz, J,
Zanobetti, A and Committee, HEIHR (2009). Air pollution and health: a European and North
American approach (APHENA). Res Rep Health Eff Inst(142): 5-90.
61 Turner, MC, Jerrett, M, Pope, A, III, Krewski, D, Gapstur, SM, Diver, WR, Beckerman, BS,
Marshall, JD, Su, J, Crouse, DL and Burnett, RT (2016). Long-term ozone exposure and
mortality in a large prospective study. Am J Respir Crit Care Med 193(10): 1134-1142.
62 U.S. Environmental Protection Agency (U.S. EPA). 2016. Integrated Science Assessment for
Oxides of Nitrogen - Health Criteria (Final Report). National Center for Environmental
Assessment, Research Triangle Park, NC. July. Available at: <
https://cfpub. epa.gov/ncea/isa/recordisplay. cfm?deid=310879>.
63 U.S. Environmental Protection Agency (U.S. EPA). 2008. Integrated Science Assessment for
Oxides of Nitrogen and Sulfur-Ecological Criteria National (Final Report). National Center for
Environmental Assessment - RTP Division, Research Triangle Park, NC. EPA/600/R-08/139.
December. Available at: .
402
-------
Chapter 9 Comparison of Benefits and Costs
This chapter compares the estimated range of total monetized health benefits to total costs
associated with the final rule. This chapter also presents the range of monetized net benefits
(benefits minus costs) associated with the final rule. Program costs are detailed and presented in
Chapter 7 of this RIA. Those costs include costs for both the new technology and the operating
costs associated with that new technology, as well as costs associated with the final rule's
warranty and useful life provisions. Program benefits are presented in RIA Chapter 8.A Those
benefits are the monetized economic value of the reduction in PM2.5- and ozone-related
premature deaths and illnesses that result from reductions in NOx emissions and directly emitted
PM2.5 attributable to implementation of the final rule.
As noted elsewhere in this RIA, the estimated benefits, costs, and net benefits do not reflect
all the anticipated impacts of the final rule.
9.1 Methods
EPA presents three different benefit-cost comparisons for the final rule:
1. A future-year snapshot comparison of annual benefits and costs in the year 2045, chosen
to approximate the annual health benefits that will occur in a year when the program will
be fully implemented and when most of the regulated fleet will have turned over.
Benefits, costs and net benefits are presented in year 2017 dollars and are not discounted.
However, 3 percent and 7 percent discount rates were applied in the valuation of avoided
premature deaths from long-term pollution exposure to account for a twenty-year
segmented cessation lag.
2. The present value (PV) of the stream of benefits, costs and net benefits calculated for the
years 2027-2045, discounted back to the first year of implementation of the final rule
(2027) using both a 3 percent and 7 percent discount rate, and presented in year 2017
dollars. Note that year-over-year costs are presented in RIA Chapter 7 and year-over-year
benefits can be found in RIA Chapter 8.
3. The equivalent annualized value (EAV) of benefits, costs and net benefits, representing a
flow of constant annual values that, had they occurred in each year from 2027 to 2045,
will yield an equivalent present value to the present value estimated in method 2 (using
either a 3 percent or 7 percent discount rate). Each EAV represents a typical benefit, cost
or net benefit for each year of the analysis and is presented in year 2017 dollars.
The two estimates of benefits (and net benefits) in each of these benefit-cost comparisons
reflect alternative combinations of the economic value of PM2.5- and ozone-related premature
deaths summed with the economic value of illnesses for each discount rate (see Chapter 8 for
more detail).
A As detailed in RIA Chapter 8, estimates of health benefits are based on air quality modeling conducted for the
proposal, and thus differences between the proposal and final rule are not reflected in the benefits analysis. We have
concluded, however, that the health benefits estimated for the proposal are a fair characterization of the benefits that
will be achieved due to the substantial improvements in air quality attributable to the final rule.
403
-------
9.2 Results
Table ES presents the benefits, costs and net benefits of the final rule in annual terms for year
2045, in PV terms, and in EAV terms.
Annual benefits are larger than the annual costs in 2045, with annual net benefits of $5.8 and
$25 billion using a 7 percent discount rate, and $6.9 and $29 billion using a 3 percent discount
rate.B Benefits also outweigh the costs when expressed in PV terms (net benefits of $14 and
$110 billion using a 7 percent discount rate, and $36 and $200 billion using a 3 percent discount
rate) and EAV terms (net benefits of $1.3 and $11 billion using a 7 percent discount rate, and
$2.5 and $14 billion using a 3 percent discount rate).
Given these results, implementation of the final rule will provide society with a substantial net
gain in welfare, notwithstanding the health and other benefits we were unable to quantify (see
RIA Chapter 8.8 for more information about unquantified benefits). EPA does not expect the
omission of unquantified benefits to impact the Agency's evaluation of the costs and benefits of
the final rule, though net benefits would be larger if unquantified benefits were monetized.
Table 9-1: 2045 Annual Value, Present Value and Equivalent Annualized Value of Costs, Benefits and Net
Benefits of the Final Rule (billions, 2017$)a'b
3% Discount
7% Discount
2045
Benefits
$12-$33
$10-$30
Costs
$4.7
$4.7
Net Benefits
$6.9 - $29
$5.8-$25
Present Value
Benefits
$91 -$260
$53 -$150
Costs
$55
$39
Net Benefits
$36 - $200
$14 -$110
Equivalent Annualized
Value
Benefits
$6.3 -$18
$5.1 -$14
Costs
$3.8
$3.8
Net Benefits
$2.5 -$14
$1.3 -$11
a All benefits estimates are rounded to two significant figures; numbers may not sum due to independent
rounding. The range of benefits (and net benefits) in this table are two separate estimates and do not
represent lower- and upper-bound estimates, though they do reflect a grouping of estimates that yield
more and less conservative benefits totals. The costs and benefits in 2045 are presented in annual terms
and are not discounted. However, all benefits in the table reflect a 3 percent and 7 percent discount rate
used to account for cessation lag in the valuation of avoided premature deaths associated with long-term
exposure.
b The benefits associated with the standards presented here do not include the full complement of health
and environmental benefits that, if quantified and monetized, would increase the total monetized benefits
B The range of benefits and net benefits presented in this section reflect a combination of assumed PM25 and ozone
mortality risk estimates and selected discount rate.
404
-------
Chapter 10 Economic Impact Analysis
This rulemaking is considered economically significant, because it is expected to have an
annual impact on the economy of $100 million or more, and thus an economic analysis has been
completed as part of this RIA. This rule is not expected to have measurable inflationary or
recessionary effects.
The benefits to human health and the environment are discussed in Chapter 8, and the costs of
the final standards are discussed in Chapter 7. The benefit-cost analysis for this rule is presented
in Chapter 9. This chapter provides an analysis of the impacts of the standards on vehicle sales
and employment.
10.1 Impact on Sales, Fleet Turnover and Mode Shift
As explained in Chapter 7, this rule is expected to increase the cost of heavy-duty (HD)
vehicles by requiring emissions control technologies capable of controlling NOx at lower levels
than are currently permitted, as well as longer emissions warranty periods for emissions control
technology components. In addition, there is an expected increase in operating costs due to an
increase in the use of diesel exhaust fluid (DEF) and emission-related repair costs beyond the
new useful life periods.
Three sectors expected to be most immediately affected by this action are: 1) HD vehicle and
engine manufacturers, 2) HD vehicle and engine buyers, and 3) HD engine equipment suppliers
(e.g., suppliers of emissions control components). Effects on industries downstream of these
sectors, such as HD vehicle dealerships or delivery industries, will be relatively smaller due to
the limited role of the cost of a HD vehicle in pricing in those sectors. The three sectors will also
be responding to the 2016 rulemaking, "Greenhouse Gas Emissions and Fuel Efficiency
Standards for Medium- and Heavy-Duty Engines and Vehicles - Phase 2" (the Phase 2 rule). The
final standards will be implemented during the same time frame as the final Phase 2 rule
standards. Both this rulemaking and the Phase 2 rule will require HD engine manufacturers to
develop and implement improvements in engine emissions controls.
As discussed in the Phase 2 rule RIA,1 increases in costs of HD vehicles from improved
emissions controls will be likely to lead to increases in final prices for HD vehicles; the
magnitude of that effect will depend on how much of the cost is passed along to potential buyers.
These price increases may affect HD vehicle sales in several ways. First, as basic economic
supply and demand theory suggests, higher prices are expected to reduce HD vehicle sales.
Second, HD vehicle buyers may strategically seek to avoid increased prices by "pre-buying,"
increasing the purchases of new vehicles before the compliance deadline for the new
requirements. This might lead to an associated period immediately afterward of "low-buying,"
during which purchases decrease, and thereby impact the rate of fleet turnover. A third potential
effect is transportation mode shift, changing from on-highway trucking to other modes of
transportation (e.g., shipping via barge or rail instead of by truck). The magnitude of each of
405
-------
these three categories of effects (sales, fleet turnover, mode shift) will depend on the costs. This
section discusses these impacts.0
10.1.1 Sales
The effects of the final standards on HD vehicle sales depend on the magnitude of the cost
increase associated with implementing improved emissions controls to comply with the
requirements, and on the degree to which the costs get passed through to vehicle buyers.
As discussed in Chapter 7, an increase in cost of HD vehicles will result from the standards
requiring the use of emissions control technologies capable of controlling NOx at lower levels, as
well as imposing longer useful life and emissions warranty periods for emissions control
technology components. While the final requirement for longer emissions warranty periods will
likely increase the purchase price of new HD vehicles, the corresponding lengthened useful life
periods are expected to make emissions control technology components more durable. More
durable components coupled with manufacturers paying for repairs during a longer warranty
period will in turn reduce repair costs, which, especially in the long term, may increase (or
reduce the decrease in) sales of new HD vehicles due to fleets and independent owner-operators
being inclined to purchase vehicles with lower repair costs.0 The exact purchase behavior of
fleet owners and independent owner-operators is challenging to predict, particularly in the time
period immediately after new standards go into effect, when buyers may be waiting to see how
the new vehicles perform relative to manufacturer claims.
If cost increases are small, either purchasers or sellers may absorb the cost increase without
measurable changes in behavior. Significant cost increases passed through to buyers may lead
potential buyers to purchase fewer vehicles than without the higher costs, or to buy vehicles
sooner than they would have otherwise, in advance of the requirements. In a report on the HD
Phase 2 rule, The National Academies of Sciences, Engineering, and Medicine stated that both
pre-buy and low-buy are likely to be short-lived phenomena, and potentially unavoidable.2
Allowing manufacturers to generate NOx emissions credits before the final standards are
required to be met may mitigate pre-buy; instead of early compliance imposing only net costs,
early compliance now provides a benefit in terms of reduced compliance costs in the future, as
well as early emissions reductions (see preamble Section III for details on the timing of the final
standards).3
c We recognize that additional external factors, including the current global COVID-19 pandemic, might impact the
heavy-duty vehicle market, however due to data limitations we are unable to include possible effects of such
external factors in our analyses. However, though sales have not rebounded to levels seen directly before the
COVID-19 pandemic, trends indicate they are heading that way. Seasonally adjusted data from The Bureau of
Economic Analysis (published at https://fred.stlouisfed.org/series/HTRUCKSSA) show heavy weight truck retail
sales (trucks weighing over 14,000 pounds) started increasing in mid-2009, through about May of 2019, and then
fell dramatically until May of 2020, though sales never fell as low as they were in the mid-2009 time-frame. Sales
increased again through March of 2021, before falling again through Sept 2021, and are currently increasing again.
This indicates that possible shocks stemming from the global COVID-19 pandemic are likely short term, and the
industry may return to historical levels by 2027.
D The reduced repair costs may counteract some of the sales effect of increased vehicle purchase cost. As a result,
they may reduce incentives for pre- and low-buy, and mitigate adverse sales impacts.
406
-------
Measuring the existence and magnitude of pre- and low-buy depends on separating those
effects from other factors that affect HD vehicle sales. If, for example, the timing of the
standards coincides with a decrease in HD vehicle sales due to an economic downturn, as likely
happened with initial implementation of the 2007/2010 HD rule that went into effect during the
eve of the Great Recession in 2007, then the estimated effects of the standard would somehow
have to be disentangled from the effects of the economic slowdown. Researchers estimating pre-
or low-buy may seek to control for underlying sales patterns like this by including other factors,
such as diesel price and gross domestic product (GDP), that also influence new HD vehicle sales.
They then look for deviations from these trends at the time that the standards go into effect.
Using this approach to control for other influential factors, Lam and Bausell found a pre-buy
of around 18,000 to 21,000 HD trucks, totaling about 20 to 25 percent of total production, in the
6-month period before the October 2002 compliance deadline for HD engine manufacturers to
reduce NOX emissions.4 Similarly, Rittenhouse and Zaragoza-Watkins (RZW) looked for pre-
buy in the seven months preceding compliance deadlines for EPA HD criteria pollutant standards
in 1998, 2002, 2007 and 2010, as well as for low-buy in the seven months after those compliance
deadlines.5 For the 2007 standards, they found a sales increase of about 31,000 vehicles over the
preceding seven months compared to the baseline, matched by an "approximately symmetric"
drop in sales in the following months. For 2002, they found a similarly symmetric, though
smaller, result of 14,000 - 18,000 Class 8 vehicles. These results suggest that the standards led to
vehicles purchases being pulled forward that would have otherwise been purchased after the
standards were promulgated in the absence of the standards. The resulting effect was slower
adoption of lower-emissions vehicles, compared to the assumed rate of sales in the absence of
new emissions standards, although there was essentially no net change in sales (the sum of pre-
buy and low-buy was not statistically different from zero in 2007 or 2002). RZW did not find
evidence of pre- or low-buy for the standards that went into effect in 1998 and 2010. They
speculated that the standards in those years were less costly and involved use of less risky,
already available technologies.
A limitation of the method used by the researchers discussed above is that they do not suggest
a way to predict how future cost changes may influence sales. This is because they do not
include price impacts in the approach to estimate pre- and low-buy impacts, yet vehicle price is
expected to change when the standard goes into effect due to an increase in cost to the
manufacturers. Both the change in price and the timing of the standard influence pre- and low-
buy because they occur at the same time, and it is statistically difficult to separate the two
effects. In addition, manufacturers of HD vehicles may affect either the magnitude or timing of
price increases in response to cost increases, confusing the effect of price on sales. Thus, while
these studies suggest that the current rule may lead to increases in sales through pre-buying
behavior, and decreases in sales through low-buying, the estimation approaches used by the
studies do not allow EPA to predict existence or magnitude of potential pre- and low-buy
impacts from future standards. In an effort to improve our analyses, EPA has been working on a
method to estimate these impacts. The approach and an example are explained in RIA Chapter
10.1.2, below.
An unpublished report attempted to develop a predictive model based on the impact of the
2007 standards. The authors assumed that a change in cost translates directly into a change in
price, which was then converted into a change in sales (Harrison and LeBel).6 The price change
407
-------
was based on asking manufacturers to estimate the costs of meeting the standards. The study then
applied a price elasticity of demand of -1.9 (that is, a 1 percent increase in price will lead to a 1.9
percent decrease in sales) to estimate sales increases of 104,000 trucks during 2005-2006, and
sales reductions of 149,000 trucks over 2007-2008. (The study did not provide details on the
source of this elasticity estimate.) The study then reported "actual"E results of sales increases of
120,000 vehicles in 2005-6 and decreases of 183,000 vehicles in 2007-8, based on comparing
estimated sales to an EPA estimate of baseline sales increased by a constant amount each year.
Unlike the published studies reviewed above, Harrison and LeBel did not control for GDP, diesel
prices, or other factors that might independently affect vehicle sales.6 As a result, the EPA
baseline used for the "actual" results is not likely to reflect actual sales in the absence of the
standards, and the "actual" pre- and low-buy values likely do not reflect changes due only to the
standards. For comparison, RZW's finding of pre-buy of about 31,000, based on controlling for
other factors, is about one-third of Harrison and LeBel's prediction and one-fourth of their
"actual" estimate.5
In sum, existing literature does not provide sufficient insight on the relationship between a
change in vehicle cost due to a new standard and sales impacts. Neither Lam and Bausell nor
RZW links a change in vehicle cost to sales impacts (a major interest of the EPA); instead, both
papers focus on the magnitude of sales impacts in the periods surrounding compliance deadlines
of HD emission regulations45 The method proposed by Harrison and LeBel links costs to sales
via a demand elasticity, but omits controls for shifts in baseline conditions as well as omitting
details on the source of the demand elasticity they used.6
For this rule, EPA acknowledges that these standards may lead to some pre-buy before the
standards go into effect, and some low-buy after the standards are effective. The estimated
increase in costs is not expected to have much effect on pre- or low-buy behavior because the
increase is small relative to the cost of the vehicle. Based on the literature previously described,
EPA is not able to quantify these effects for this rule. In the following subsection we outline an
approach that could be used to quantify sales effects, and we illustrate how this method could be
used to estimate pre- and low-buy as a function of the estimated costs outlined in Chapter 7.
10J .2 EPA's Research to Estimate Sales Effects
In 2020 EPA contractors conducted a review of available peer reviewed literature on the
effects of EPA's HD standards on HD sales (see RIA Chapter 10.1.1 for literature review
results). The contractors then conducted an original analysis of the effects of previous EPA
standards on pre- and low-buy for HD vehicles.7
The analysis uses monthly vehicle sales data from the twelve-month period before and after
previous EPA HD standards went into effect (2002,F 2007, 2010, and 2014) to estimate pre- and
E Quotation marks around "actual" are included in Harrison and LeBel (2008).
F Due to a consent decree in 1998 requiring six major HD engine manufacturers in the U.S. to meet a 2.5 g/bhphr
limit on NMHC+NOX by October 1, 2002, much of the regulatory implementation of the 2004 HD rule was pulled
forward. Therefore, we will refer to the implementation of that regulation as the 2002 standards, instead of the 2004
standards, in order to keep the focus on compliance dates. More information on these consent decrees can be found
on EPA's Civil Cases and Settlements by Statute webpage:
https://cfpub.epa.gov/compliance/cases/index.cfm?templatePage=12&ID=l&sortby=RELEASE_D ATE, RELEASE
_DATE&stat=Clean%20Aii%20Act
408
-------
low-buy due to each standard. The analysis examined controls for the effects of month of year,
GDP, Brent Oil price, total imports and exports, and consumer sentiment, and then used binary
indicator variables from 1 through 12 months pre- and post-regulation to identify deviations from
trends in sales specifically around those regulations' implementation dates. All other variables
(except for the binary variables of interest and the month of year) were transformed into log-
differences to address statistical issues associated with time series data. Independent regressions
were estimated for vehicle Classes 6 through 8, and for each of the four previous HD
regulations' implementation dates. Additional details of this analysis are available in the
contractors' report.
Results show no statistically significant sales effects for Class 6 vehicles. There were a few
statistically significant results for Class 7, but the majority were of the opposite sign than
expected (that is, reduced sales before the standards and increased sales afterwards). For Class 8
vehicles, there were statistically significant results in the expected directions, with evidence of
short-lived pre-buy before the 2010 and 2014 standards, and evidence of short-lived low-buy
after the 2002, 2007, and 2010 standards. The rest of this section focuses only on Class 8
vehicles. For more discussion on Classes 6 and 7, see Appendix 7.2 and Chapter 4.4.3 of the
report.
The results provide estimates of the percent deviation in sales from trend for the combined
months leading up to and following the start of new emissions standards. For pre-buy,
statistically significant results range from no change persisting for the eleven months before the
2002 standards, to a 13.2 percent increase in the percent change in sales persisting for one month
before the 2014 standards. Statistically significant effects persist for up to eleven months before
the 2002 standards. For low-buy, statistically significant effects range from no change to a 14.9
percent decrease in the percent change in sales persisting for six months after the 2007 standards.
Statistically significant effects persist for up to twelve months following the 2007 standards.
Importantly, in addition to capturing the effects due to price changes associated with the
regulations, the coefficients also capture unobserved factors, such as concerns over vehicle
reliability and control technology uncertainty. Table 10-1 provides the results for the coefficients
on the pre- and low-buy indicators, along with their length of persistence and the regulation to
which the result is attributed. The significant coefficients are shown in bold face type.
409
-------
Table 10-1: Pre and Low-Buy Sales Effects Coefficients
2002
2007
2010
2014
Combined Months Pre-Regulation
12
0.024
0.004
0.009
0.000
11
-0.0**
-0.006
0.021
0.000
10
0.0**
-0.005
0.041
0.010
9
0.032
-0.008
0.032
0.032
8
0.041
-0.004
0.057**
0.013
7
0.044
-0.006
0.059**
0.019
6
0.037
-0.004
0.043
0.021
5
0.029
0.003
0.054*
0.019
4
0.004
-0.011
0.079***
0.030
3
0.047
-0.013
0.071**
0.014
2
-0.017
-0.012
0.105***
0.003
1
-0.032
-0.01
0.078***
0.132***
Combined Months Post-Regulation
1
0.065***
_0
-0.144***
-0.009
2
-0.051
-0.099***
-0.083*
-0.012
3
-0.115
-0.133***
-0.051
-0.015
4
-0.065
-0.143***
-0.052
0.003
5
-0.066
-0.144***
-0.075**
-0.009
6
-0.076*
-0.149***
-0.052
-0.006
7
-0.017
-0.121***
-0.022
0.001
8
-0.018
-0.114***
-0.034
0.000
9
-0.018
-0.099***
-0.020
0.003
10
-0.007
-0.073**
-0.030
0.006
11
-0.027
-0.07**
-0.010
-0.013
12
-0.014
-0.065**
-0.005
0.000
*** p <0.01; ** p <0.05; *p<0.1
As can be seen in Table 10-1, results vary by regulation. For the purposes of this discussion,
we focus on results for the 2007 and 2010 standards.0 For the 2007 standards, there is no
statistically significant pre-buy. There is statistically significant low-buy for all the periods from
the period of one month after the standard through the combined period of 12 months after the
standard, with magnitude increasing up to, and falling after, the combined period of 6 months
post-standard. For the 2010 standards, there is some evidence of both pre- and low-buy.
Statistically significant pre-buy can be seen for the period of 1 month up to the combined period
of 5 months, and again at the combined periods of 7 and 8 months pre-standard. There is
significant low-buy for the periods of 1, 2 and 5 months post-regulation. Results indicate that the
observed effects are short-lived, on the order of months rather than years.
" We do not consider the results of the 2002 compliance date to be generalizable for several reasons. Litigation may
have affected purchase plans for many firms resulting from the pulling forward of compliance dates from 2004 to
2002. In addition, there may have been greater concerns over the reliability of new engines compared to other
regulatory actions, which may have led to more low-buy. Also, the cost of compliance in 2002 was estimated to be
lower than that of other regulations. We do not consider the 2014 standards to be generalizable either. This rule
reduced greenhouse gas (GHG) emissions, which had lower technology costs and fuel savings relative to other rules.
In addition, numerous pathways for compliance leads to difficulty estimating the price change in HD vehicles due to
the regulation. More details and discussion on the 2002 and 2004 standards are available in the contractors' report.
410
-------
10.1.2.1 Estimating Elasticities
To estimate a change in Class 8 vehicle sales due to future EPA emission standards, we
transform the coefficients on the indicator variables, explained above, into demand elasticities.
These elasticities (s) measure the percent change in vehicle sales due to a percent change in
vehicle prices:
Equation 10-1
% AS ales
e =
%A Price
The percent change in sales for Class 8 vehicles comes from the coefficients on the indicator
variables from Table 10-1. In estimating elasticities, we only use the significant coefficients,
while noting that no response (an elasticity of 0) is also represented in the results. The percent
change in price is estimated by dividing the estimated cost of compliance published in the EPA
RIAs associated with the relevant standard (2007/2010 HD rule) by the estimated purchase price
of a Class 8 vehicle in that year (adjusted to 2010 dollars).11 Table 10-2 shows the regulatory
cost, the HD vehicle prices and the resulting percent change in price we are using to estimate the
elasticities from the 2007 and 2010 standards.
Table 10-2: Regulatory Costs and HD Vehicle Prices Used to Estimate Elasticities
Statutory
Deadline
Regulatory
Cost (2010$)
HD Vehicle
Price
% Change in
Price
2007
$9,741
$98,900
9.8%
2010
$7,662
$108,250
7.1%
From the statistically significant pre- and low-buy sales effects for Class 8 vehicles, we
estimate a set of pre- and low-buy elasticities by dividing the percent change in sales, from Table
10-1, by the percent change in price, from Table 10-2.
Table 10-3 shows the estimated statistically significant coefficients (percent change in sales)
from Table 10-1, their period of effect, and the associated estimated elasticity. We expect pre-
buy elasticities to be positive (more sales before new emission standards) and low-buy
elasticities to be negative (fewer sales after new emission standards). Because the smallest
statistically significant sales effect is zero, and a number of other effects are not statistically
different from zero, the smallest pre- and low-buy elasticities are zero - no effect due to the
standards. It should also be noted that not only the magnitude of the elasticity matters, but also
the time period over which the elasticity applies. A large elasticity for a short period of months
may measure less effect than a small elasticity over a longer period.
H The estimated cost of compliance was based on EPA's cost of compliance in the RIA for the relevant standard.
The price of a Class 8 HD vehicle for each year was calculated as an average of a high and low list price from an
online source for HD vehicle sales (Commercial Truck Trader, a site that advertises new and used trucks for sale).
411
-------
Table 10-3: Elasticity Estimates
Statutory
Deadline
Period of
Effect
(Months)
% Change
in Sales (P)
Estimated
Elasticity
All
Any
0
0
2010
8
0.057
0.805
7
0.059
0.834
5
m
6
Ph
5
0.054
0.763
4
0.079
1.116
3
0.071
1.003
2
0.105
1.483
1
0.078
1.102
All
Any
0
0
2007
12
-0.065
-0.660
11
-0.070
-0.711
10
-0.073
-0.741
9
-0.099
-1.005
8
-0.114
-1.157
7
-0.121
-1.229
b
m
6
-0.149
-1.513
£
o
5
-0.144
-1.462
4
-0.143
-1.452
3
-0.133
-1.350
2
-0.099
-1.005
1
-0.070
-0.711
2010
5
-0.075
-1.060
2
-0.083
-1.173
1
-0.144
-2.034
There are several limitations to the results presented in Table 10-3. As noted in Chapter
10.1.2, the sales coefficients used to estimate the elasticities likely capture aspects of the final
regulation not solely limited to changes in price (e.g., adverse fuel consumption effects, or
concerns about the reliability of untested control technology). Similarly, the base vehicle prices
and estimated regulatory costs discussed above are estimates and may not correspond with
observed base prices or increased regulatory costs. A commenter on the proposed rule also noted
that a limitation of this method is that it assumes 100% cost pass-through to consumers.1 In
addition, though we estimate a range of possible effects, including zero, this method assumes
buyers will continue to respond to regulation similarly as they have in the past. This may change
over time as market offerings change, for example if vehicles become more durable, or as the
HD vehicle market includes more electrified or fuel cell HD vehicles.
1 See Section 25 of the Response to Comments for our response to this comment.
412
-------
Since these elasticities are based on monthly data, it is appropriate to apply the estimated
elasticities to monthly series. Analysis of the coefficients over time indicates that the observed
effects are short-lived, on the order of months rather than years. As noted above and described
below, the time period is a critical factor for estimating the impacts.
10.1.2.2 Illustrative Example
This subsection outlines how we could apply the pre- and low-buy elasticities presented in
Table 10-3 to this rulemaking. Though the methodology to develop the elasticities has been peer
reviewed,8 the application in a rulemaking is new, and thus, in this subsection, we are illustrating
how we could use this method to estimate pre- and low-buy as a function of the estimated costs
outlined in Chapter 7.
Expanding Equation 10-1, elasticity measures can be approximated as
Equation 10-2
%A Sales ASales Price
e = = *
%A Price Sales APrice
In this application, we want to estimate how a change in price leads to a change in sales.
Therefore, rearranging Equation 10-2, we get
Equation 10-3
A Price
ASales = s * * Sales
Price
The elasticity measures come from the estimates explained above.
For this example, A Price is the estimated cost of compliance for a Class 8 HD vehicle to
meet the final MY 2027 standards, $4,827 (see RIA, Chapter 7, Table 7-24). We assume
implementation starts January 1 for the vast majority of the heavy-duty engine industry/
Price is the estimated price of a Class 8 truck, which we set to $130,000.K
Sales are the estimated monthly Class 8 vehicle sales in 2026 and 2027.L Monthly sales are
derived from Class 8 vehicle population data from projected sales volumes using EPA's MOVES
1 This is an illustrative example, and thus may not folly represent the final program; see preamble Section III. A for
additional discussion on implementation dates in the final rule.
K The price of HD vehicles varies greatly, and in part due to the features or options of the vehicle. The price we use
here comes from the estimated price of a low-end, new, semi-truck from Truckers Bookkeeping Service from June,
2021; a pdf of the page "How much does a semi truck cost?" can be found in the docket for this rule, Docket ID
EPA-HQ-OAR-2019-0055.
L We do not have the sales data for 2026, therefore we approximate 2026 vehicles sales in this illustrative example
with 2027 vehicle sales.
413
-------
modelM and month-specific effects from the contractors' report.N To estimate pre-buy, we use
the estimated monthly HD vehicle sales in the months before January 1, 2027. That is, pre-buy
for the 2027 compliance date is estimated with the calculated monthly sales in 2026. To estimate
low-buy, we use the estimated monthly HD vehicle sales in 2027.
To get the sales effects, the elasticity estimates from Table 10-3 are multiplied by the change
in price divided by the base price. This value is then multiplied by the estimated Class 8 HD
vehicle sales for each month over the period of effect for that elasticity measure. This results in a
change in sales for each month over the period of effect. The monthly results are then summed to
get a total affect for each elasticity estimate.
For pre-buy, for example, the elasticity measurement of 1.10 has a period of effect of one
month, so we use the Class 8 sales from December, 2026 and make the pertinent multiplication
for just December, 2026. For the elasticity measurement of 0.81, the period of effect is 8 months,
so we use the Class 8 sales estimates from May, 2026 (8 months before January, 2027) through
December, 2026 and the pertinent multiplication estimation is made for each affected month.
Then, the changes in sales for each affected month are added together to get the total effect. The
results for pre-buy are in Table 10-4 below. As discussed in Chapter 10.1.2.1, there are sales
effects results that are statistically indistinguishable from zero, which means that zero impact on
sales is the lower bound on effects. Total sales of Class 8 vehicles under this final rule are
estimated to increase by between 0 and approximately 2 percent on an annual basis before the
2027 compliance deadline. In addition, the duration of the effects is a critical component in the
calculation of sales impacts. For example, the elasticity of 0.83 for a duration of 7 months has a
larger aggregate impact than the larger elasticity of 1.12 for a duration of 4 months. The result
that produces the largest estimate for an aggregate increase in sales is the elasticity of 0.81 for 8
months.
Table 10-4: Illustrative Pre-Buy Results
Period of Effect
(Months)
Elasticity
Aggregate Sales
Change
Cumulative %
Change in Sales
Any
0
0
0
8
0.81
4,701
2.01%
7
0.83
4,264
1.83%
5
0.76
2,815
1.21%
4
1.12
3,317
1.42%
3
1.00
2,255
0.97%
2
1.48
2,196
0.94%
1
1.10
936
0.40%
M See Population and Activity of Onroad Vehicles in MOVES3 (EPA-420-R-21-012) available online at
https://nepis.epa.gov/Exe/ZyPDF.cgi?Dockey=P 101 lTF8.pdf for more information on how vehicle population data
is estimated in MOVES.
N Because these populations are annual, and the elasticities are monthly, we have to distribute the annual sales
throughout the year. To do so, we estimate the average monthly sales and then use the monthly sales effect (the
percent change in sales by month) estimated in the contractors' report to better approximate the sales by month.
414
-------
Low-buy is estimated the same way, though we use monthly sales estimates for the requisite
number of months following the January 1, 2027 compliance date. Low-buy results for the final
rule are show in Table 10-5. For an elasticity of-1.51 over the course of 6 months (an estimate
from the 2007 standards), we use the monthly sales estimates for January, 2027 through June,
2027, make the pertinent multiplication estimations for each affected month, and add the
monthly results to get the aggregate sales change. As discussed in Chapter 10.1.2.1, for all rules
there are sales effects results that are statistically indistinguishable from zero, which means that
zero impact on sales is the lower bound on effects. This example estimates sales of Class 8
vehicles in the months following the 2027 compliance date to fall by between 0 and just under 3
percent on an annual basis. As with pre-buy, both the magnitude of the elasticity and the duration
of effect are important in estimating the total effect. For example, the elasticity of -1.01 for the
duration of 9 months (from the 2007 standards) results in a larger aggregate effect than the larger
elasticity of-1.23 for the duration of 7 months. The result that produces the largest estimate for
an aggregate decrease in sales is the elasticity of -1.16 for a duration of 8 months.
Table 10-5: Illustrative Low-Buy Results0
Statutory
Deadline
Period of Effect
(Months)
Elasticity
Aggregate Sales
Change
Cumulative %
Change in Sales
All
Any
0
0
0
12
-0.66
(5,717)
-2.45%
11
-0.71
(5,552)
-2.38%
10
-0.74
(5,323)
-2.28%
9
-1.01
(6,447)
-2.76%
8
-1.16
(6,586)
-2.82%
r-
7
-1.23
(6,108)
-2.62%
o
-------
initially, vehicles meeting the standards will only be a small portion of the total fleet. For
instance, in 2019 about 369,000 new medium-/heavy-duty trucks and buses were sold, compared
to over 14 million total medium-/heavy-duty vehicle registrations.9 Overtime, as more vehicles
subject to the standards enter the market and older vehicles leave the market, the emissions
reductions due to the standards will increase. This relationship holds true even if new vehicle
sales are unaffected by the standards.
If pre-buy and low-buy behaviors occur, they can shift emissions impacts in several ways.
First, under low-buy, there is slower adoption of new vehicles, which implies that emissions
reductions will be slower than under the assumption of no change in vehicle sales (RZW).5 On
the other hand, the pre-bought HD vehicles are likely to displace older, more polluting vehicles,
which may provide an earlier reduction in emissions than would have occurred without the
standards. However, although the pre-bought new HD vehicles are likely to have lower
emissions than the older, displaced vehicles, the emissions reductions are likely to be smaller
than the reductions that will be realized from the purchase and use of new vehicles subject to
these standards, so that the net effect of pre-buy is to slow reductions in emissions.5
Another potential effect of the standards is a net reduction in new vehicle sales. This could
result from either a smaller pre-buy than the post-standards low-buy,p or some potential buyers
deciding not to purchase at all. In this case, the vehicle miles traveled (VMT) of older vehicles
may increase to make up for the VMT that would otherwise have been expected of the newer
("missing") vehicles. To the extent that the older vehicles emit more than the missing vehicles,
emissions may increase.*2 However, because the VMT is likely to be shifted to the newer HD
vehicles among the existing fleet, and most of those vehicles are expected to be in compliance
with the existing HD vehicle standards, this effect is expected to be small.
Quantifying these effects requires a robust method to estimate the effects of the standards on
pre-buy and low-buy, as well as a method to estimate shifts in VMT among vehicle vintages in
the case of an expected change in the net sales of newer vehicles. In the absence of robust
methods to estimate these effects, EPA is not quantifying the fleet turnover or emissions impacts
of sales effects in this rule, though, as with pre-buy and low-buy, we acknowledge these potential
impacts.
The estimated increase in operating costs due to an increase in the use of DEF may lead to a
slower fleet turnover if VMT does not shift from older (lower DEF requirements) vehicles to the
newer (higher DEF requirements) vehicles. This leads to slower emission reductions than if those
newer vehicles were used. However, as this increase in operating costs is small, and may be
offset in part by reduced repair costs, we expect minimal effects on fleet turnover due to the
change in operating costs.
Another factor that may impact fleet turnover is the increase in HD vehicles' operational life.
With an increase in operational life, vehicles compliant with this regulation may stay on the road
p Though this is a possibility, it should be noted that the RZW (2018) study found that pre-buy approximately
equaled low-buy. Harrison and LeBel (2008) found that low-buy exceeded pre-buy.
Q This effect is sometimes called the "Gruenspecht effect," based on the theory presented in Gruenspecht, Howard
(1982), "Differentiated Regulation: The Case of Auto Emissions Standards.". hnerican Economic Review 72: 328-
331.
416
-------
longer, leading to reduced fleet turnover. As the vehicles that might be in operation longer due to
this regulation are in compliance with this regulation, we do not expect emission impacts due to
reduced fleet turnover as a function of increased operational life.
We do not have data to estimate the effect the final regulation might have on HD scrappage
rates. If the regulation leads to HD vehicle owners holding onto used vehicles longer (a reduction
in scrappage), this could result in slower fleet turnover, even if new vehicle purchases are
unaffected by the regulation. If the regulation leads to vehicle owners preferring to shift
ownership to newer vehicles, this could result in increased fleet ownership and associated
increased scrappage. Modeling dynamic scrappage (scrappage that changes due to the regulation)
relies on consumer choice modeling, which is especially difficult to estimate in the HD context,
as there are many different kinds of consumers and products spanning Classes 4 through 8.
Though EPA has included estimates of dynamic scrappage in previous light-duty rules, analyses
based on dynamic scrappage were only included in the recent 2023-2026 model year light-duty
rule, where we relied on the CCEMS model.R Though we do not have data to estimate dynamic
scrappage, HD vehicle scrappage is accounted for in the MOVES model in the same way light-
duty vehicle scrappage is estimated.8 The MOVES scrappage algorithm uses historical vehicle
survival rates to predict future year scrappage. Scrappage in the MOVES model is static and does
not have a consumer choice component.
10,1,4 Potential for Mode Shift
Another possible response to the new emissions standards is shifting freight shipments to
other transportation modes, such as rail or barge. This may happen, for example, if the new
standards were to raise operating costs enough to make truck transportation more expensive than
rail or marine alternatives.
EPA does not expect this rule to result in a transportation mode shift. Generally, shipping
cargo via truck is more expensive per ton-mile than barge or rail, and less expensive than air. 10'u
This is due to many factors, not the least of which is labor costs (each truck has at least one
driver). Even though trucking is more expensive than rail or marine on a ton-mile basis, it is a
very attractive transportation alternative for several reasons: shipping via truck is generally faster
and more convenient than rail or marine, trucks can reach more places, and trucks may be less
constrained by available infrastructure than barge or rail. In addition, shipping via truck does not
require trans-shipments (transferring from one mode to another, for example to deliver cargo to
or from the port or rail yard), and it allows partial deliveries at many locations. This speed,
infrastructure availability, and delivery flexibility make trucking the transportation solution of
choice for many kinds of cargo across most distances. As a result, smaller shipments of higher-
valued goods (e.g., consumer goods) tend to be transported by air or truck, while larger
shipments of lower-valued goods (e.g., raw materials) tend to go via rail or barge.10'12
R Final Rule to Revise Existing National GHG Emissions Standards for Passenger Cars and Light Trucks Through
Model Year 2026. Found online at https://www.epa.gov/regulations-emissions-vehicles-and-engines/final-rule-
revise-existing-national-ghg-emissions. See Chapter 4 of the final Regulatory Impact Analysis for details regarding
EPA's use of CCEMS for that rule.
s This is documented in Section 6.1 and Appendix C of Population and Activity of Onroad Vehicles in MOVES3
(PDF) (232 pp, 4.3 MB, April 2021, EPA-420-R-21-012).
417
-------
Studies of intermodal freight shifts, such as Comer et al. or Bushnell and Hughes, focus on
changes in cost per ton-mile as a potential source of transportation mode shift.10'12 Comer et al.
note, for instance, that fuel consumption "depend[s] on the type of freight being moved, route
characteristics, transport speed, and locomotive/truck characteristics."10 Bushnell and Hughes
estimate that increased fuel prices for truck transportation lead to small substitutions between
truck and rail for small or large shipments, and higher shifts for intermediate-sized shipments.12
The findings from this study suggest that the variation (in kinds and values) of goods shipped by
different means likely results in only a small amount of mode shift in response to a change in
operating cost (e.g., fuel prices). However, due to data availability, this study approximates
freight rates with fuel costs, assumes shipping distances using different modes are the same, and
mostly does not consider transportation availability constraints affecting some modes in some
regions. These limitations may distort the effects they estimate.
A mode shift study EPA carried out in 2012 in the context of new sulfur limits for fuel used in
large ships operating on the Great Lakes may help address some of these limitations.11 The
methodology used a combination of geospatial modeling and freight rate analysis to examine the
impact of an increase in ship operating costs. While the focus of the study was transportation
mode shift away from marine and toward land, it noted that truck transportation is far more
expensive than both rail and marine on a ton-mile basis.1 It also shows that even a large
percentage increase in marine operating costs did not raise freight rates by a similar percentage,
because fuel costs are only part of total operating costs. In the case of truck transportation,
operating costs are a much smaller portion of total costs. The results of this study combined with
the others cited in this section indicate that changing the cost of truck transportation is unlikely to
create mode shift.
The primary effect of the standards on operating costs is an increase in the use of DEF, which
is expected to be partially offset by reductions in emission repair costs. The increase in total
operating cost is a very small part of the total increase in cost impacts estimated in this final rule
(see Chapter 7.3). Because the cost effect is expected to be small relative to the price of a HD
vehicle, and substitution between trucks and other modes is limited by the nature of the goods
and their routes, we do not expect significant effects on mode shift. Finally, given the higher
costs of truck transportation, a relatively small increase in truck freight rates due to the small
increase in operating costs are unlikely to affect the competitive dynamics of the transportation
sector.
10.1.5 Effects on Domestic and International Shares of Production
The final standards are not expected to provide incentives for manufacturers to shift between
domestic and foreign production. This is because the standards apply to any 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 these final rules might lead to application and use
T Figure 1-5 in U.S. EPA Office of Transportation and Air Quality. "Economic Impacts of the Category 3 Marine
Rule on Great Lakes Shipping." EPA-420-R-12-005. 2012.
418
-------
of technologies that other countries may seek now or in the future, developing this capacity for
domestic producers now may provide some additional ability to serve those markets.
10.1.6 Summary of Sales, Turnover, and Mode Shift Impacts
As discussed in Chapter 7, EPA expects the cost of new HD vehicles to increase (see Chapter
7 for details). This increase may have some impact on new vehicle sales; in particular, it is
possible that truck buyers increase purchases before the standards become effective (pre-buy),
and reduce purchases afterwards (low-buy). Studies of pre-buy and low-buy suggest that these
phenomena have occurred in the past, but those studies do not provide methodologies for
estimating the impact of new rules on future vehicle sales. For that reason, EPA conducted an
analysis to develop a relationship between estimated changes in vehicle price due to a new
regulation and corresponding changes in vehicle sales (i.e., pre- and low-buy elasticities). We
present the details of this new analysis and provide an illustrative example of applying pre- and
low-buy elasticities to estimate potential sales impacts on Class 8 vehicles. For pre-buy and low-
buy, the illustrative example shows that sales impacts on Class 8 vehicles are of limited duration
and range from zero impact to approximately 2 percent for pre-buy and from zero to just under
three percent for low-buy.
Whether shippers might choose a different mode for freight depends not only on the cost per
ton-mile of the shipment, but also the value of the shipment, the time needed for shipment, and
the availability of infrastructure. This rule is expected to affect the cost per ton-mile by only a
small amount. For that reason, EPA expects little transportation mode shift to occur due to the
final standards. EPA also does not expect changes in where production happens in response to
these standards.
10.2 Employment Impacts
This section explains the methods and estimates of employment impacts due to this
regulation. Though the rule primarily affects HD vehicles, the employment effects may be felt
more broadly in the motor vehicle and parts sectors due to the effects of the standards on sales.
Thus, we focus our assessment on the motor vehicle manufacturing and the motor vehicle parts
manufacturing sectors, with some assessment of impacts on additional sectors likely to be most
affected by the standards. Chapter 10.2.1 offers a brief, high-level explanation of employment
impacts due to environmental regulation and discusses a selection of the peer-reviewed literature
on this topic. Chapter 10.2.2 discusses EPA's qualitative and quantitative estimates of the partial
employment impacts of this rule on regulated industries. Chapter 10.2.3 examines employment
impacts in some closely related sectors, and Chapter 10.2.4 summarizes expected employment
impacts.
10.2,1 Economic Framework for Employment Impact Assessment
Economic theory of labor demand indicates that employers affected by environmental
regulation may increase their demand for some types of labor, decrease demand for other types
of labor, or for still other types, not change it at all. A variety of conditions can affect
employment impacts of environmental regulation, including baseline labor market conditions
and employer and worker characteristics such as industry, and region. Isolating employment
impacts of regulation is difficult as they are a challenge to disentangle from employment impacts
caused by a wide variety of ongoing concurrent economic changes.
419
-------
A growing literature has investigated employment effects of environmental regulation.
Morgenstern et al. decompose the labor consequences in a regulated industry facing increased
abatement costs. They identify three separate components.13 First, there is a demand effect
caused by higher production costs raising market prices. Higher prices reduce consumption (and
production), reducing demand for labor within the regulated industry. Second, there is a cost
effect where, as production costs increase (for example, due to pollution control activities that
require additional labor to produce the same output quantity), plants use more of all inputs
including labor to produce the same level of output. Third, there is a factor-shift effect where
post-regulation production technologies may have different labor intensities. These three effects
outlined by Morgenstern et al. provide the foundation for EPA's analysis of the impacts of the
current regulation on labor.13
Additional papers approach employment effects through similar frameworks. Berman and Bui
model two components that drive changes in firm-level labor demand: output effects and
substitution effects.14>u The output effect happens when prices increase, leading to a decrease in
quantity demanded, and results in a decrease in production. The substitution effect happens when
regulation affects labor intensity of production (holding output constant). Deschenes describes
environmental regulations as requiring additional capital equipment for pollution abatement that
does not increase labor productivity.15 These higher production costs induce regulated firms to
reduce output and decrease labor demand (an output effect) while simultaneously shifting away
from the use of more expensive capital towards increased labor demand (a substitution effect).v
Ehrenberg and Smith describe how at the industry level, labor demand is more likely to be
responsive to regulatory costs if: (1) the elasticity of labor demand is high relative to the
elasticity of labor supply, and (2) labor costs are a large share of total production costs.16 Labor
demand might also respond to regulation if compliance activities change labor intensity in
production.
Arrow, Cropper, at al. state that, in the long run, environmental regulation is expected to
cause a shift of employment among employers rather than affect the general employment level.17
Even if they are mitigated by long-run market adjustments to full employment, many regulatory
actions have transitional effects in the short run.18'19 These movements of workers in and out of
jobs in response to environmental regulation are potentially important distributional impacts of
interest to policy makers. Of particular concern are transitional job losses experienced by
workers operating in declining industries, exhibiting low migration rates, or living in
communities or regions where unemployment rates are high.
Workers affected by changes in labor demand due to regulation may experience a variety of
impacts including job gains or involuntary job loss and unemployment. Compliance with
environmental regulation can result in increased demand for the inputs or factors (including
labor) used in the production of environmental protection. However, the regulated sector
generally relies on revenues generated by their other market outputs to cover the costs of
supplying increased environmental quality, which can lead to reduced demand for labor and
u Berman and Bui 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.
v For an overview of the neoclassical theory of production and factor demand, see Chapter 9 of Layard and Walters
(1978).
420
-------
other factors of production used to produce the market output. Workforce adjustments in
response to decreases in labor demand can be costly to firms as well as workers, so employers
may choose to adjust their workforce over time through natural attrition or reduced hiring, rather
than incur costs associated with job separations (see, for instance, Curtis and Hafstead and
Williams).Error! Bookmark not defined.'20
Employment impacts, both positive and negative, in sectors upstream and downstream from
the regulated sector, or in sectors producing substitute or complimentary products, may also
occur.
10.2.2 Employment Impacts in the Motor Vehicle and Parts Manufacturing Sectors
In this section, EPA presents partial estimates of industry-level employment effects of the
final rule. We use the labor intensity of production for motor vehicle manufacturing and motor
vehicle parts manufacturing to provide a range of potential employment impacts.w
Our analysis follows the structure of Morgenstern et al., as described above, to estimate the
impacts of this rule on the regulated sector.13 We qualitatively describe the employment impacts
due to the factor-shift and demand effects, provide an illustrative example of demand effects, and
quantitatively estimate the employment impacts due to the cost effect. Due to a variety of
reasons, including that our quantitative estimates of the demand effect are merely illustrative,
and we do not estimate factor-shift effects, our estimates do not reflect the total effects on
employment in the regulated industries.
10.2.2.1 The Factor-Shift Effect
The factor-shift effect reflects employment changes due to changes in labor intensity of
production resulting from compliance activities. The labor intensity of manufacturing HD
vehicle engines or HD vehicles might increase or decrease because of the rule. Due to a lack of
information on expected changes in labor intensity, the estimated employment impacts in this
chapter do not include the factor-shift effect. In addition, EPA is aware of HD market shifts
toward battery electrification, and that there may be employment effects due to that shift at least
in part due to different labor intensity needs. Our results do not reflect market transitions toward
battery electrification, in large part because we do not have data on employment differences in
traditional manufacturing sectors and battery electric manufacturing sectors, especially for future
expected effects. In addition, as discussed in preamble Section III, battery-electric and fuel cell
electric vehicles are subject to the final standards, but we did not evaluate these technologies in
setting the level of the final standards. Further, battery-electric and fuel cell electric vehicles
cannot participate in the final ABT program (see Section IV.G). The combination of not
including battery-electric technology when setting the final standards, and not allowing
manufacturers to generate NOx emissions credits from these technologies, leads us to not expect
this regulation to induce a significant shift toward battery-electric heavy-duty vehicle production.
w We do not identify impacts separately for these sectors because we do not have information on the division of
costs between them.
421
-------
10.2.2.2 The Cost Effect
The cost effect reflects the impact on employment due to increased costs from adopting
technologies needed for vehicles to meet the standards. The analysis holds output constant,
meaning that it does not include sales impacts. We estimate the cost effect using the historic
share of labor in the cost of production to extrapolate future estimates of impacts on labor due to
new compliance activities in response to this regulation. Specifically, we multiply the share of
labor in production costs by the production cost increase estimated as an impact of this rule. This
provides a sense of the magnitude of potential impacts on employment.
The use of the ratio of the share of labor in production costs to estimate "cost effect"
employment has both advantages and limitations. It is often possible to estimate these ratios for
specific sectors, for example, the average number of workers in the HD vehicle manufacturing
sector per $1 million spent in that sector, rather than using ratios from more aggregated sectors,
such as the motor vehicle manufacturing sector. This means that it is not necessary to extrapolate
employment ratios from possibly unrelated sectors. On the other hand, these estimates are
averages, covering all the activities in these sectors and may not be representative of the labor
effects when expenditures are required for specific activities, or when manufacturing processes
change due to compliance activities in such a way that labor intensity changes. For instance, the
ratio of workers to production cost for the HD motor vehicles manufacturing sector represents
this ratio for all HD vehicle manufacturing and not just for production processes related to
emission reductions compliance activities. In addition, these estimates do not include changes in
sectors that supply these sectors, such as steel or electronics producers. The effects estimated
here can be viewed as effects on employment in the HD motor vehicle sector due to the changes
in expenditure in that sector, rather than as an assessment of all employment changes due to the
final standards. In addition, labor intensity is held constant in the face of increased expenditures;
this approach does not take in account changes in labor intensity due to changes in the nature of
production (the factor-shift effect), which could either increase or decrease the employment
impacts estimated here.
Some vehicle parts are made in-house by HD vehicle manufacturers. Other parts are made by
independent suppliers who are not directly regulated but will be affected by this rulemaking as
well. Because EPA does not know whether abatement equipment to comply with the final
standards will be produced by the original equipment manufacturers (OEMs) or by suppliers, we
use labor ratios for both sectors (and their subsectors) to provide a range of estimates for the cost
effect impacts.
We include estimates from two sectors that are broadly defined and two that are more
narrowly defined. Specifically, we estimate labor impacts for the aggregated sectors 'motor
vehicle manufacturing' and 'motor vehicle parts manufacturing', and for the more specific sectors
'light truck and utility vehicle manufacturing' and 'heavy-duty truck manufacturing.'x
We rely on three different public sources to get a range of estimates of employment per $1
million expenditures: the Annual Survey of Manufactures (ASM) and the Economic Census
x The 'light truck and utility vehicle manufacturing' sector is included because these estimates include results for
trucks from class 2b/3 through 6. See Preamble Section I of this rule for more discussion on the HD engine classes
included in this rulemaking.
422
-------
(EC), both provided by the U.S. Census Bureau, and the Employment Requirements Matrix
(ERM) provided by the U.S. Bureau of Labor Statistics (BLS). The EC is conducted every 5
years, most recently in 2017.Y The ASM is an annual subset of the EC and is based on a sample
of establishments. The latest set of data from the ASM is from 2020. The EC and ASM have
more sectoral detail than the ERM, providing estimates out to the 6-digit North American
Industry Classification System (NAICS) code level. They provide separate estimates of the
number of employees and the value of shipments, which we convert to a ratio in this
employment analysis.2 The ERM provides direct estimates of employees per $1 million in
expenditures for a total of 202 aggregated sectors that roughly correspond to the 4-digit NAICS
code level, and provides data through 2020. Table 10-6 below shows the sector definition, the
NAICS code, and the ERM sector number where appropriate that EPA uses to estimate
employment effects in this analysis.
Table 10-6: Sectors Used in this Analysis
Sector Definition
NAICS
ERM Sector Number
Motor vehicle manufacturing
3361
80
Light truck and utility vehicle manufacturing
336112
Heavy-duty truck manufacturing
33612
Motor vehicle parts manufacturing
3363
82
Table 10-7 provides the estimates of employment per $1 million of expenditure for each
sector for each data source, adjusted to 2017 dollars using the U.S. Bureau of Economic Analysis
Gross Domestic Product Implicit Price Deflator retrieved from the Federal Reserve Bank of St.
Louis. The values are adjusted to remove effects of imports through the use of a ratio of domestic
production to domestic sales of 0.81.AA While the estimated labor ratios differ across data
sources, they each exhibit a similar pattern across sectors. Within the 4-digit NAICS code level,
motor vehicle parts manufacturing seems to be the most labor-intensive sector, followed by the
motor vehicle manufacturing sector. Within motor vehicle manufacturing, heavy-duty truck
manufacturing appears to be more labor-intensive than light truck and utility vehicle
manufacturing.
Y Though the Economic Census was conducted in 2022, data from 2022 will not begin to be released until March
2024.
z The total employment across the two 4-digit NAICS code sectors used in this analysis (see Table 10-6) as reported
in the ASM and the EC ranges from 775,016 to 787,640 depending on which data source is used; as noted above the
most recent ASM and EC were conducted in 2020 and 2017, respectively.
A A To estimate the proportion of domestic production affected by the change in sales, we use data from WardsAuto
for total car and truck production in the U.S. compared to total car and truck sales in the U.S. Over the period 2009-
2021, the proportion averages 83 percent. From 2016-2021, the proportion average is slightly lower, at 81 percent.
423
-------
Table 10-7: Employment per $1 Million Expenditures (2017$) in the Motor Vehicle Manufacturing Sectora
Source
Sector (NAICS)
Ratio of Workers
per $1 Million
Expenditures
Ratio of Workers
per $1 Million
Expenditures,
Adjusted for Domestic
vs. Foreign Production
BLS ERM 2017
Motor vehicle manufacturing (3361)
0.504
0.408
ASM 2016
Motor vehicle manufacturing (3361)
0.754
0.611
EC 2017
Motor vehicle manufacturing (3361)
0.615
0.498
BLS ERM 2017
Motor vehicle parts manufacturing (3363)
1.846
1.496
ASM 2016
Motor vehicle parts manufacturing (3363)
2.642
2.141
EC 2017
Motor vehicle parts manufacturing (3363)
2.231
1.808
ASM 2016
Heavy-duty truck manufacturing (33612)
1.451
1.175
EC 2017
Heavy-duty truck manufacturing (33612)
0.988
0.800
ASM 2016
Light truck and utility vehicle
manufacturing (336112)
0.640
0.519
EC 2017
Light truck and utility vehicle
manufacturing (336112)
0.478
0.388
a BLS ERM refers to the U.S. Bureau of Labor Statistics' Employment Requirement Matrix, 2020 values. ASM
refers to the U.S. Census Bureau's Annual Survey of Manufactures, 2020 values. EC refers to the U.S. Census
Bureau's Economic Census, 2017 values. These are the most recent data available.
Over time, the amount of labor needed in the motor vehicle industry has changed: automation
and improved methods have led to significant productivity increases. The BLS ERM, for
instance, provides estimates that, in 1997, about 1.2 workers in the Motor Vehicle Manufacturing
sector were needed per $1 million, but only 0.5 workers by 2020 (in 2017$).BB Because the ERM
is available annually for 1997-2020, we use these data to estimate productivity improvements
over time. We regress logged ERM values on a year trend for the Motor Vehicle Manufacturing
and Motor Vehicle Parts Manufacturing sectors. We use this approach because the coefficient
describing the relationship between time and productivity is a direct measure of the average
percent change in productivity per year. The results of the regressions suggest a 4.4 percent per
year productivity improvement in the Motor Vehicle Manufacturing Sector and a 4.0 percent per
year improvement in the Motor Vehicle Parts Manufacturing Sector.
We then use those estimated percent improvements in productivity to project the number of
workers per $1 million of production expenditures through 2031. Although the costs and benefits
analyses in the preceding chapters go out to 2045, we chose to model the employment effect due
to cost increases through 2031, for a total of five years. This is because our method is an
approximate, partial employment analysis, as well as being dependent on future, uncertain,
macro-economic conditions. The results provided below represent an order of magnitude effect,
rather than definitive impacts. We calculate separate sets of projections (adjusted to 2017$) for
each set of data, ERM, EC, and ASM, for all four sectors described above. The ERM projections
BB http://www.bls.gov/emp/ep_data_emp_requirements.htm; this analysis used data for sectors 80 (Motor Vehicle
Manufacturing) and 82 (Motor Vehicle Parts Manufacturing) from "Chain-weighted (2009 dollars) real domestic
employment requirements tables;" see "Cost Effect Employment Impacts calculation" in the docket.
424
-------
are calculated directly from the fitted regression equations used to estimate the projected
productivity growth, since the regressions themselves used ERM data. For the ASM and EC
projections, we use the ERM's ratio of the projected productivity growth value in each future
year to the projected production expenditure value in 2020 for the ASM and 2017 for the EC (the
base years in our data) to determine how many workers are needed per $1 million of
expenditures, in 2017$. In other words, we apply the projected productivity growth estimated
using the ERM data to the ASM and EC numbers.
To simplify the results, we compare the projected employment across data sources and report
only the maximum and minimum effects in each year across all sectors.cc We provide a range
rather than a point estimate because of the inherent difficulties in estimating employment
impacts as well as the uncertainty over how the costs are expended. The reported ranges provide
an estimate of the expected magnitude of the cost effect. In Table 10-7, the Motor Vehicle Parts
Manufacturing Sector value from the ASM provides the maximum employment estimates per $1
million; the Light Truck and Utility Vehicle Manufacturing Sector value from the EC provides
the minimum estimates.
Cost estimates developed for this rule are provided in Chapter 7. We use the technology cost
estimates from that chapter to estimate the employment impacts of a change in cost of
manufacturing HD vehicles due to this rule. The technology cost estimates (in $ million) are
multiplied by the estimates of workers per $1 million in costs. The projected estimates of
technology costs and corresponding minimum and maximum estimated employment impacts for
each year are shown in Table 10-8, below. The effects are shown in job-years, where a job-year
is, for example, one year of full-time work for one person or two years of half-time work for two
workers. Increased technology costs of vehicles and parts is, by itself and holding labor intensity
and output constant, expected to increase employment by between 800 and 5,300 per year
between 2027 and 2031 under this final rule. While we estimate employment impacts, measured
in job-years, beginning with program implementation, some of these employment gains may
occur earlier as vehicle manufacturers and parts suppliers hire staff in anticipation of compliance
with the standards.
cc To see details, as well as results for all sources, see "Final Cost Effect Employment Impacts Calculation" in the
docket.
425
-------
Table 10-8: Estimated Employment Effects Due to Increased Costs of Vehicles and Parts (Cost Effect), in
Job-Years
Year
Minimum
Employment due
to Cost Effect
Maximum
Employment due
to Cost Effect
2027
1,000
5,300
2028
900
5,000
2029
800
4,700
2030
800
4,400
2031
800
4,200
10.2.2.3 The Demand Effect
The demand effect reflects employment changes due to changes in new vehicle sales. If HD
vehicle sales decrease, fewer people would be needed to assemble trucks and the components
used to manufacturer them. On the other hand, if pre-buy occurs, HD vehicle sales may increase
temporarily, leading to temporary increases in employment. RIA Chapter 10.1.1 and Chapter
10.1.2 discuss the factors influencing the effect of final requirements on demand for new HD
vehicles, explains why that effect is difficult to quantify, outlines a new method to quantify the
impacts and explains how we might use it to estimate pre- and low-buy sales effects in future
rulemakings.
EPA received many comments on the proposed rule requesting that we expand our current
employment analysis to include demand effects. We have responded to those requests with the
example method laid out below to estimate illustrative demand effects due to a change in sales.
Pre- and low-buy are short-term sales effects. As such, it should be noted that employment
effects due to pre- and low-buy may be short term as well. Some of these effects may also be
transitional as workers shift from one sector to another.
Using the illustrative results on pre- and low-buy sales effects as outlined in Chapter 10.1,
combined with employment information from the ASM and EC as described above and domestic
HD truck production from Wards Automotive Group, we estimate the increase in job-years due
to pre-buy in the months before rule implementation, and the decrease in job-years due to low-
buy in the months after rule implementation.
We sum the annual employment values from the Motor Vehicle Manufacturing sector
(NAICS 3361) and the Motor Vehicle Parts Manufacturing sector (NAICS 3363) from the ASM
and EC data sets.™ For the ASM, we use data from 2018, 2019 and 2020. For the EC, we only
have data from 2017. We then divide the annual employment for each year by domestic truck
DD We only sum value from these four digit NAICS code sectors because NAICS codes are nested and summing the
more detailed (longer) NAICS codes with the more general (shorter) NAICS codes will result in double counting.
We do not use data from the ERM because it provides employment per million dollars in sales as opposed to
employment in job-years.
426
-------
production in that year to get a value for job-years per truck for each year. Employment,
production and job-years per truck can be seen in Table 10-9.
Table 10-9: Annual HD Employment and Production
Data
Source
Year
Annual
Employment
Domestic Truck
Production™
Job-Years/Truck
Produced
ASM
2020
775,000
6,890,000
0.112
ASM
2019
821,000
8,380,000
0.098
ASM
2018
810,000
8,510,000
0.095
EC
2017
788,000
8,150,000
0.097
It should be noted that year-over-year percentage change in domestic truck production is
greater than the year-over-year percentage change in annual employment. In addition, production
fell between 2018 and 2019, though employment increased in the directly affected sectors in our
analysis. This indicates that employment changes in the sectors measured does not always follow
changes in domestic HD truck production. This could be due to many factors, including workers
transitioning between manufacturing jobs or sectors. There may also be lag factors, production
changes happening ahead of employment changes, or future planning by manufacturers. For
example, if manufacturers believe changes in production will be temporary, they may not want to
change employment by much, with the understanding that they will need to revert back to
previous levels of employment after production returns to previous levels. Additionally,
employers may not want to face costs associated with layoffs, preferring to reduce employment
through attrition.
Using the data in the table above, we estimate the average job-years per truck in the directly
affected segments to be 0.101 job-years per HD truck. We apply this ratio to the estimated
maximum total annual change in sales in 2026 due to pre-buy and in 2027 due to low-buy as
shown in the illustrative sales effects example in Chapter 10.1.2. This results in an estimate of a
change in job-years due to a change in demand. Table 10-10 shows the results.
Table 10-10: Estimated Maximum Total Change in Sales and Illustrative Change in Job-Years due to
Demand Effect
Year
Max Total
Annual Sales
Effect
Change in Job-
Years
2026 (pre-buy)
4,700
450
2027 (low-buy)
(6,600)
(640)
We assume these demand effects would be short-term, as they would be due to a short-term
change in demand, with our illustrative pre-buy estimates ranging up to 8 months, and low-buy
up to 12 months. As mentioned above, employment changes may lag production changes.
EE This data comes from Wards Automotive Group. U.S. Vehicle Production by Manufacturer, UsaPr05.xls
427
-------
10.2.2.4 Summary of Employment Effects in the Motor Vehicle Sector
As explained above, the overall effect of the final standards on motor vehicle sector
employment depends on the relative magnitude of the demand, cost, and factor-shift effects, as
described by Morgenstern, et al.13 We quantitatively estimate employment impacts of the
standards due to the cost effect. To provide a sense of its magnitude, EPA provides a range of
estimates of the cost effect. Due to a lack of data, we do not estimate factor-shift effects. And
though we are not estimating demand employment effects for this rule, we do outline a method
and illustrate how that method could be used to estimate these effects in Chapter 10.2.2.3. This
proposed method relies on the illustrative results for pre- and low-buy outlined in Chapter
10.1.2.2.
For the regulated sector, the partial employment impact due to the effect of increased
manufacturing costs due to compliance activities is estimated to range between 800 and 5,300
job-years between 2027 and 2031. We expect the demand effect to reduce these employment
increases, as qualitatively discussed above. Finally, we are unable to predict the direction of the
factor-shift effect.
10.2,3 Employment Impacts on Related Sectors
The rule may affect employment in several related sectors including downstream on
purchasers and dealers.
10.2.3.1 Effects on Purchasers of Heavy-Duty Vehicles
Because of the diversity of HD vehicles, entities from a very wide range of transportation
sectors will be purchasing vehicles subject to these standards. As discussed in Chapter 10.1,
vehicles subject to these standards are likely to be more expensive to purchase compared to
vehicles not subject to the standards. HD vehicles are typically commercial, and typically
provide an "intermediate good:" that is, they are used to provide a commercial service, rather
than being a final consumer good. As a result, the higher costs of the vehicles may result in
higher prices for the services provided by these vehicles, and potentially reduced demand for
those services. In turn, there might be less employment in the sectors providing those services.
One commenter on the ANPR provided data indicating that, between 2010 and 2018, the up-
front costs of HD vehicles were approximately 10 to 16 percent of per-mile costs.21 Given that
we expect the potential impacts of the final standards on up-front costs of vehicles to be only a
few percent of total vehicle cost, any increase in per-mile costs are likely to be less than 1
percent (even a 5 percent increase in vehicle costs would result in only a 0.8 percent increase in
per-mile costs, at the high end of the range). Therefore, we expect only negligible to small
impacts on transportation services demand, and related employment in transportation services
sectors. Per-mile cost increases for some sectors will be higher than this average, while they will
be lower in other sectors due to factors such as differences in how the vehicles are used,
including average mileage accumulation of the vehicles in the sector. The actual effects on
demand for the services and related employment will depend on cost pass-through, and
responsiveness of demand to transportation cost increases.
Lengthening the warranty period may provide some positive impacts on employment for
vehicle purchasers. Some commenters submitted comments on the proposed rule conveying
428
-------
concerns over the effects on productivity of downtime due to problems with emission control
systems.22 As discussed in Chapter 7.2.3, the extended warranty provisions finalized in this rule
are expected to not only reduce repair costs for vehicle purchasers, but may also provide
incentives to manufacturers to improve quality and thus reduce the need for repairs. These
effects are expected to reduce costs, and thus mitigate adverse impacts on employment for
vehicle purchasers.
10.2.3.2 Effects on Heavy-Duty Vehicle Dealers and Service Providers
If sales of HD vehicles decrease, then HD vehicle dealers will have fewer sales, and may
employ fewer people as a result. At the same time, dealers, and other independent service repair
shops, often provide repair and maintenance services. The extended warranty provisions are
expected to facilitate repair and maintenance of emission control system components, which
could result in increased demand for workers servicing vehicles. On the other hand, as discussed
in Chapter 7.2.3, the extended useful life provisions are also be expected to provide incentives to
OEMs to improve quality and may reduce the need for warranty claims. Thus, effects on
employment for service providers, including dealers, may be positive or negative.
Similarly, as discussed in Preamble Section IV.B.3, this rule aims to improve access to
serviceability information to improve owner experiences operating and maintaining HD engines
and provide greater assurance of long-term in-use emission reductions by reducing likelihood of
occurrences of tampering. One commenter on the ANPR noted that it is currently difficult for
anyone other than dealers to service vehicles, and commented on the proposed rule that
finalizing the proposed serviceability provisions will help drivers maintain the emissions
equipment themselves. It is possible that improving serviceability will improve maintenance due
to lower costs of conducting service.23 It is also possible that improved serviceability may shift
some of the work, and thus employment, from dealers to independent service repair shops or to
the owners themselves, with an unclear impact on the overall level of employment.
10,2,4 Summary of Employment Impacts
Employment in the HD manufacturing sector depends on three effects: how the effects of the
standards on vehicle prices affect demand for new vehicles (demand effect); the labor demand
needed to meet the standards (cost effect); and any change in labor intensity of production due to
complying with the standards (factor-shift effect). In Chapter 10.2.2, we outline a method to
estimate demand effects on employment using the illustrative sales effects results discussed in
Chapter 10.1.2. We are unable to quantify the factor-shift effect, and therefore we are unable to
estimate the impact on net employment in the HD manufacturing sector. To give an estimate of
the range of cost-effect-related employment changes due to the final regulation, the analysis
estimates a range between 1,000 and 5,300 job-years in 2027, with impacts falling each year. By
2031, estimated impacts range from 800 to 4,200 job years. For comparison, in May 2021, the
Bureau of Labor Statistics reports about 244,000 employees in Motor Vehicle Manufacturing.24
Other sectors that sell, purchase, or service HD vehicles might also experience employment
impacts due to the standards. The effects on these sectors will depend on the degree of cost pass-
through to prices for HD vehicles and the effects of useful life and warranty requirements on
demand for vehicle repair and maintenance.
429
-------
Chapter 10 References
1 U.S. EPA and National Highway Traffic Safety Administration. "Greenhouse Gas Emissions
and Fuel Efficiency Standards for Medium- and Heavy-Duty Engines and Vehicles - Phase 2
Regulatory Impact Analysis." EPA-420-R-16-900. 2016.
2 National Academies of Sciences, Engineering, and Medicine. "Reducing Fuel Consumption
and Greenhouse Gas Emissions of Medium- and Heavy-Duty Vehicles, Phase Two: Final
Report," p. 329. Washington, DC: The National Academies Press. 2020.
https://doi.org/10.17226/25542.
3 Rubin, J. "A Model of Intertemporal Emission Trading, Banking, and Borrowing." Journal of
Environmental Economics and Management 31: 269-286. 1996.
4 Lam, T., and Bausell, C. "Strategic Behaviors Toward Environmental Regulation: A Case of
Trucking Industry." Contemporary Economic Policy 25(1): 3-13. 2007.
5 Rittenhouse, K., and Zaragoza-Watkins,M. "Anticipation and Environmental Regulation."
Journal of Environmental Economics and Management 89: 255-277. 2018.
6 Harrison, D., Jr., and LeBel, M. "Customer Behavior in Response to the 2007 Heavy-Duty
Engine Emission Standards: Implications for the 2010 NOX Standard." NERA Economic
Consulting. 2008. Available in the docket, Docket ID EPA-HQ-OAR-2019-0055-0576.
7 U.S. Environmental Protection Agency. "Analysis of Heavy-Duty Vehicle Sales Impacts Due
to New Regulation." EPA-420-R-21-013. 2021.
https://cfpub.epa.gov/si/si_public_pra_view.cfm?dirEntryID=349838&Lab=OTAQ.
8 U.S. Environmental Protection Agency. "Analysis of Heavy-Duty Vehicle Sales Impacts Due
to New Regulation: Response to Reviewers Comments." EPA-420-R-21-014. 2021.
https://cfpub.epa.gov/si/si_public_pra_view.cfm?dirEntryID=349838&Lab=OTAQ.
9 New vehicle sales from Wards Intelligence. "U.S. Factory Sales of Medium/Heavy Trucks and
Buses by GVW" UsaFsOlO.xlsx. Registration data from Wards Intelligence. "U.S. Total Truck
Registrations by State and Type-2019," sum of "Truck Tractors and "Other Med./Hvy."
UsaRe 10_2019 .xl sx. 2022
10 Comer, B., Corbett, J. J., Hawker, J.S., Korfmacher, K., Lee , E.E., Prokop, C., and
Winebrake. J. "Marine Vessels as Substitutes for Heavy-Duty Trucks in Great Lakes Freight
Transportation." Journal of the Air & Waste Management Association 60: 884-890. 2010.
11 U.S. EPA Office of Transportation and Air Quality. "Economic Impacts of the Category 3
Marine Rule on Great Lakes Shipping." EPA-420-R-12-005. 2012.
12 Bushnell, J., and Hughes, J. "Mode Choice, Energy Consumption and Emissions in U.S.
Freight Transportation." Working paper. 2019. Available online:
https://spot.colorado.edu/~jonathug/Jonathan_E._Hughes/Main_files/Freight_Modes.pdf,
accessed 10/21/2019.
430
-------
13 Morgenstern, R.D.; Pizer, W.A.; and Shih, J.-S. "Jobs Versus the Environment: An Industry-
Level Perspective." Journal of Environmental Economics and Management 43: 412-436. 2002.
14 Berman, E. and Bui, L. T. M. "Environmental Regulation and Labor Demand: Evidence from
the South Coast Air Basin." Journal of Public Economics 79(2): 265-295. 2001.
15 Deschenes, O. "Balancing the Benefits of Environmental Regulations for Everyone and the
Costs to Workers and Firms." IZA World of Labor 22v2. 2018. Available online:
https://wol.iza.org/uploads/articles/458/pdfs/environmental-regulations-and-labor-markets.pdf
16 Ehrenberg, R. G., and Smith, R.S. Modern Labor Economics: Theory and Public Policy
(Addison Wesley Longman, Inc.), p. 108. 2000.
17 Arrow, K.J., Cropper, M.L., Eads, G.C., Hahn, R.W., Lave, L.B., Noll, R.G., Portney, P.R.,
Russel, M., Sch mal en see, R., Smith, V.K., and Stavins, R.N. "Is there a role for benefit-cost
analysis in environmental, health, and safety regulation9" Science 272(5259): 22 1-222. 1996.
18 Smith, V. K. "Should Benefit-Cost Methods Take Account of High Unemployment?
Symposium Introduction," Review of Environmental Economics and Policy 9(2):165-178. 2015
19 U.S. OMB. 2015. "2015 Report to Congress on the Benefits and Costs of Federal Regulations
and Agency Compliance with the Unfunded Mandates Reform Act", available
at: https://obamawhitehouse.archives.gov/sites/default/files/omb/inforeg/2015_cb/2015-cost-
benefit-report.pdf/
20 Hafstead ,M.A.C. and Williams III, R.C. "Unemployment and Environmental Regulation in
General Equilibrium." Journal of Public Economics 160: 50-65. 2018.
21 "Comments of the American Trucking Associations on the Control of Air Pollution from New
Motor Vehicles: Heavy-Duty Engine Standards Advance Notice of Proposed Rulemaking."
Docket EPA-HQ-OAR-2019-0055-03 57.
22 Comments of the National Association of Small Trucking Companies, "Draining the Swamp"
Docket EPA-HQ-OAR-2019-0055-0456; Comments of the Owner-Operator Independent Drivers
Association, "Re: Docket # EPA-HQ-OAR-2019-005 5 'Control of Air Pollution from New
Motor Vehicles: Heavy-Duty Engine Standards,'" Docket EPA-HQ-OAR-2019-0055-0379.
23 Comments of the Owner-Operator Independent Drivers Association, "Re: Docket # EPA-HQ-
OAR-2019-0055 'Control of Air Pollution from New Motor Vehicles: Heavy-Duty Engine
Standards,"' Docket EPA-HQ-OAR-2019-0055-0379.
24 U.S. Bureau of Labor Statistics. National Industry-Specific Occupational Employment and
Wage Estimates: NAICS 336100 - Motor Vehicle Manufacturing. May 2021. Available online:
https://www.bls.gov/oes/2021/may/naics4_336100.htm, accessed 9/2/2022.
431
-------
Chapter 11 Small Business Analysis
This chapter presents our analysis of the economic impacts of this action on small entities that
are subject to the highway heavy-duty engine and vehicle provisions of this rule. These are:
heavy-duty vehicle manufacturers, heavy-duty secondary vehicle manufacturers, and heavy-duty
alternative fuel engine converters. Other entities that are subject to the rule are either not small
(e.g., engine and incomplete vehicle manufacturers) or are not expected to incur any additional
burden from the rule (e.g., manufacturers in sectors other than highway heavy-duty engines and
vehicles and that are subject only to the regulatory amendments contained in Section XII of the
Preamble for this rule).
11.1 Definition and Description of Small Businesses
Under the Regulatory Flexibility Act (5 USC 601 et seq.), a small entity is defined as: (1) a
business that meets the definition for small business based on the Small Business
Administration's (SBA) size standards;1>A (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; or (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 analysis considers only small business entities subject to the rule. Small governmental
jurisdictions and small not-for-profit organizations are not subject to the rule as they have no
certification or compliance requirements.
11.2 Overview of the Heavy-Duty Program and Type of Entities Covered
As described in the preamble and elsewhere in this RIA, this rulemaking sets out a
comprehensive approach to reduce air pollution from highway heavy-duty engines. The key
provisions can be grouped into three broad categories: 1) reducing emissions under a wider range
of engine operating conditions than those covered in existing requirements, including refueling
events (e.g., revised exhaust and refueling emission standards and updated test procedures); 2)
maintaining emission control over a greater portion of an engine's operational life (e.g.,
lengthened useful life and regulatory emission warranty periods), and 3) providing manufacturers
with flexibilities to meet the new emission standards while clarifying our regulations.
While the rule also includes regulatory amendments for sectors other than highway heavy-
duty engines and vehicles, these amendments for other sectors correct, clarify, and streamline
existing regulatory provisions, and they will impose no additional burden on small entities in
these other sectors.
There are three categories of highway heavy-duty engine and vehicle entities that are subject
to the rule:
A EPA relied on the 2019 SBA size standards that were current at the time of the analysis. We acknowledge that new
size standards went into effect after we conducted the analysis, but those new standards would not change the
analysis. See https://www.sba.gov/article/2022/oct/03/sba-issues-final-rule-adopt-naics-2022-small-business-size-
standards. The October 2022 version of the size standards are available online:
https://www.sba.gov/document/support-table-size-standards.
432
-------
• Heavy-duty engine manufacturers
• Heavy-duty conventional vehicle manufacturers, including incomplete and secondary
vehicle manufacturers
• Alternative fuel engine converters
Heavy-duty engine manufacturers have been developing, testing, and certifying engines for
many years in compliance with EPA rulemakings adopted under the CAA. These companies
will be required to produce engines that meet new emission standards and certify them using
revised test procedures. The heavy-duty engine manufacturers that certify engines to EPA's
program include no small entities.
Heavy-duty vehicle manufacturers are one of two types. The first type of company
manufactures and certifies a complete or incomplete vehicle and its associated engine.6 These
companies are not small entities. The second type of company manufactures a vehicle of its own
design using a certified engine or incomplete vehicle produced by a different company.
Manufacturers that finish an incomplete vehicle produced and certified by a different company
(i.e., "secondary manufacturers") complete the vehicle by adding the truck body and other
equipment. While these secondary manufacturers are not required to certify with EPA (because
they use an incomplete vehicle certified by another company), they may incur costs to
accommodate any changes made to the certified incomplete vehicle to meet the new emission
requirements. Several secondary manufacturers are small entities under the SBA definition, and
the economic impacts of the rule on them are described in Section 11.3.
Alternative fuel engine converters are also subject to the rule. Two of these companies are
small entities under the SBA definitions, and the impacts on them are described in Section 11.4.
Finally, Section 11.5 contains a summary table of the expected economic impacts on small
entities subject to the rule.
11.3 Impacts on Small Entities: Heavy-Duty Secondary Vehicle Manufacturers
A secondary vehicle manufacturer is defined as anyone that produces a vehicle by modifying
a complete vehicle or completing the assembly of a partially complete vehicle, although a
manufacturer controlled by the manufacturer of the base vehicle (or by an entity that also
controls the manufacturer of the base vehicle) is not a secondary vehicle manufacturer; instead,
both entities are considered to be one manufacturer (40 CFR 1037.801 definition of secondary
vehicle manufacturer). EPA's heavy-duty vehicle program allows an engine manufacturer to
introduce partially complete vehicles into U.S. commerce to be completed by a secondary
vehicle manufacturer (see 40 CFR 1037.622). These incomplete vehicles will typically be
certified by the engine manufacturer. The program also allows a manufacturer to introduce
complete vehicles into U.S. commerce for modification by a small manufacturer (e.g., a
recreational vehicle manufacturer); these also will typically be certified by the engine
manufacturer. The provisions specify that a secondary vehicle manufacturer may finish assembly
of partially complete vehicles if it obtains a vehicle that is not fully assembled with the intent to
manufacture a complete vehicle in a certified configuration.
B See the definition of "vehicle" in 40 CFR 1037.801.
433
-------
The impacts of this rule on secondary vehicle manufacturers are different depending on
whether the vehicle is produced using a compression-ignition incomplete vehicle or a spark-
ignition incomplete vehicle.
Secondary vehicle manufacturers that produce a heavy-duty vehicle using a compression-
ignition incomplete vehicle are not expected to need to modify their manufacturing or other
processes to comply with the rule. As discussed in Chapter 3 of this RIA, compression-ignition
engine manufacturers are expected to achieve the new criteria pollutant engine emission
standards by modifying the engine and aftertreatment system technologies already applied to
these engines to meet the existing standards (e.g., selective catalytic reduction). As a result, the
engine and aftertreatment systems needed to meet the new criteria pollutant emission standards
are expected to be similar to the systems vehicle manufacturers install today and secondary
vehicle manufacturers are not expected to need to redesign or modify their products or
production processes to accommodate these compression-ignition engine-based certified
systems. Therefore, we do not expect secondary vehicle manufacturers that use compression-
ignition incomplete vehicles to experience adverse economic impacts as a result of this rule.
The analysis of impacts of the final rule on secondary vehicle manufacturers that produce a
heavy-duty vehicle using a spark-ignition incomplete vehicle includes multiple steps. Similar to
manufacturers of compression-ignition engines, Chapter 3 indicates that spark-ignition engine
manufacturers are expected to achieve the new criteria pollutant engine exhaust emission
standards by modifying the engine and aftertreatment system technologies already applied to
these engines to meet the existing standards (e.g., three-way catalysts). As a result, the engine
and aftertreatment systems are expected to be similar to the systems vehicle manufacturers install
today and secondary vehicle manufacturers are not expected to need to redesign or modify their
products or production processes to accommodate spark-ignition engine-based certified systems.
Therefore, we do not expect secondary vehicle manufacturers that use spark-ignition incomplete
vehicles to experience adverse economic impacts as a result of the new spark-ignition criteria
pollutant exhaust emission standards.
However, these spark-ignition incomplete vehicles will also be required to comply with new
refueling emission standards for vehicles fueled by gasoline, other volatile liquid fuels, and
gaseous fuels. Compliance with these standards may require some secondary vehicle
manufacturers to change their manufacturing or other processes to accommodate compliant
refueling emission control systems. Historically, an incomplete vehicle that is sold to a
secondary vehicle manufacturer includes the fuel system and its evaporative emission controls as
part of the incomplete vehicle's certified configuration. When manufacturers of chassis-certified
complete heavy-duty vehicles (i.e., Classes 2b and 3) adopted ORVR technology to meet
refueling emission standards (59 FR 16262, April 6, 1994), the design changes were contained in
the fuel system and did not require changes to the vehicle body to accommodate the new
technology (i.e., filler door location and designs remained the same). To comply with the final
refueling emission standards being adopted in this rule, manufacturers of incomplete heavy-duty
vehicles fueled by volatile fuels may add to or replace existing components of the evaporative
emission control systems currently being installed on incomplete vehicles. We expect incomplete
vehicle manufacturers will strive to design compliant ORVR systems that maintain continuity
from previous fuel system designs, minimizing the need for vehicle body redesign to
accommodate any changes to emission control systems. Our expectation is reinforced by the
434
-------
comments we received on the proposal to this rule, in which ORVR suppliers expressed
confidence in the relationship between engine OEMs and delegated assemblers (i.e., secondary
vehicle manufacturers) to effectively implement refueling requirements for incomplete heavy-
duty vehicles.0 Secondary manufacturers that finish the vehicle bodies have many years of
experience installing evaporative emission control systems as delegated assemblers, and the
ORVR instructions are expected to add very few, if any, steps to the evaporative system
instructions currently provided by the chassis manufacturers.
Secondary vehicle manufacturers are not required to certify with EPA, and so we do not have
a list of secondary manufacturers that will be subject to the rule. Instead, we used the Hoovers
D&B database to identify small companies engaged in the Motor Vehicle Body Manufacturing
(NAICS Code 336211 with 1,000 employees or fewer) or Motor Home Manufacturing (336213
with 1,250 employees or fewer) sectors.2 We limited our search to companies located in the
United States. This approach is reasonable because it is unlikely that a foreign entity would
purchase a certified incomplete vehicle from a manufacturer located in the United States,
transport that vehicle to a location outside the United States, complete the vehicle, and then
export that completed vehicle back into the United States with the associated transportation
costs. If there were such a company, the cost of the additional transportation to and from the
assembling country would likely exceed the expected costs of compliance with the rule ($2,528
per company per year; see below). Also, the additional transportation costs would likely make
the completed vehicle uncompetitive in the U.S. market (with the exception of luxury trucks or
recreational vehicles, in which case the company would likely have revenue that can
accommodate the costs of complying with the requirements).
We adjusted the initial list of 1,190 companies to remove those that are subsidiaries of another
company (they have a parent or ultimate parent company). For the Motor Vehicle Body
Manufacturing sector, we further adjusted the list to reflect only companies engaged in truck and
bus body manufacturing (as identified by their Standard Industrial Classification (SIC) code) and
removed companies that do not make truck bodies or that make light-duty trucks (these
companies are not subject to the rule).D For the Motor Home Manufacturing sector, we selected
only companies engaged in motor home manufacturing (using their SIC code) and removed van
conversions (those are light-duty vehicles not subject to this rule). Finally, we removed
companies with four or fewer employees because it is not likely that a company with fewer than
five total employees manufactures completed trucks. It should be noted that we also removed
one company from the list that does not appear to have become operational (the $100 annual
revenue reported in Hoovers for this company was likely a placeholder). This procedure yielded
a list of 249 small entities engaged in the manufacture of secondary vehicles and thereby
potentially subject to the refueling standards. It should be noted that the final list of companies
does not distinguish between those that produce spark-ignition and compression-ignition
vehicles.
c See comments from the Manufacturers of Emission Controls Association (EPA-HQ-OAR-2019-0055-1320) and
Ingevity Corporation (EPA-HQ-OAR-2019-0055-1213).
D Removed: cranes, overhead traveling; dump truck lifting mechanism; fifth wheel, motor vehicle; truck beds; truck
bodies (motor vehicles); truck bodies and parts; truck cabs, for motor vehicles; truck tops; van bodies.
435
-------
We estimated the impacts of the rule on these small entities using the following information.
We assume each company will have one-time costs associated with reviewing new instructions
for ORVR (10 hours/family), possible vehicle design R&D if the change to the evaporative
control cannister requires different mounting assemblies or the area where it is mounted must be
adjusted (8 hours/family), and training associated with installing the new ORVR system (1
hour/family). We also assume a recurring production cost for installing the new ORVR system
(1 hour/unit). Assuming 2 families with 20 unit per family for each company (40 vehicles
produced per year) and $43.58/hour, the total cost of the program in the first year is expected to
be $2,528 per company. We then compared this to the annual revenue reported in Hoovers for
each of the small entities.E
For these secondary vehicle manufacturers, the expected costs of $2,528 is less than 1 percent
of revenue for 201 of the companies and between 1 and 3 percent of revenue for 48 companies.
Figure 11-1 contains a graphical representation of the revenue distribution of these companies.
The impacts are summarized in Table 11-1 presented in Section 11.5.
E While EPA's guidance for Regulatory Flexibility Act analysis specifies that annual sales should be used in the
analysis for small companies, it also indicates that "revenue or receipts (though technically different than sales) can
usually serve as a reasonable proxy for sales." Footnote 19, page 21. EPA's Action Development Process, Final
Guidance for EPA Rulewriters: Regulatory Flexibility Act as amended by the Small Business Regulatory
Enforcement Fairness Act. November 2006.
436
-------
Small Entities - Compliance Costs as Percent of Annual Revenue
(Source: Hoovers D&B, accessed August 2021)
100
90
80
70
60
50
40
30
20
10
0
69
89
Jl
14
d
JUL
~ MAICS 336211 Motor Vehicle Body
Manufacturing, n=217
~ NAICS 336213 Motor Home
Manufacturing, n=32
10
*1
5
l—| 0
r i
0 1
'1
3 1
H1
——'i
o\o
o\o o\o o\o o\o o\o o\o o\o o\o o\o o\o o\o
<§> & ^ ^ A* c? ^ t? A* <£' <£'
q- q- q- q- v <$>' AV 0Q' & <$>' AV 0°' # A*5' '
o- o- O- Cr V V V V V V V V
Estimated Compliance Costs as Percent of Annual Revenue
Figure 11-1: Secondary Vehicle Manufacturers, Estimated Impacts as a Percent of Annual Revenues
11.4 Impacts on Small Entities: Heavy-Duty Alternative Fuel Engine Converters
Companies that convert compression-ignition or spark-ignition heavy-duty engines to use
alternative fuels are also subject to the final rule. Alternative fuel converters do not always need
to certify the conversions to the emission standards. However, they may need to perform testing
to show that modified engines continue to meet the applicable new standards as part of the
process for meeting 40 CFR part 85, subpart F, to be exempt from the tampering prohibition. For
this analysis, we conservatively assumed the process for these fuel converters will include
emission testing using a new test procedure (SET if converting an SI engine or LLC if converting
a CI engine) and preparing the data for submission to EPA to obtain an EPA Certificate of
Compliance.
EPA identified 5 companies that convert heavy-duty engines to run on alternative fuel. These
are companies that certified engines with EPA as of 2020. We obtained their employment and
annual revenue numbers from Hoovers. Two of these companies are small entities under the
SBA definitions based on annual receipts.
To estimate the impacts of the rule on these small entities we assume that each company will
have one-time costs associated with developing and performing the emission test of 20
hours/family. There are no recurring costs per vehicle expected. Assuming 4 families and
$43.58/hour, the total cost of the program in the first year is expected to be $3,486 per company.
437
-------
Our examination of the annual revenues for the two small alternative fuel engine converters
reveals that these costs, $3,486 per company per year, is not expected to impose a significant
impact on either of them. Even a low significant impact threshold of 1 percent of revenue would
correlate to an annual revenue of $348,640 or less. Each of the two small alternative fuel engine
converters has annual revenues in excess of that amount and therefore will not experience a
significant impact from the rule. These results are summarized in Table 11-1 presented in
Section 11.5.
11.5 Summary Table of Impacts on Small Businesses Subject to the Rule
Table 11-1: Summary of Impacts on Small Businesses Subject to the Rule
Number of
Impact as
percent of annual
NAICS
Category
Sector description
SBA
Threshold
small companies
subject to the
revenue,
number of small
rule
companies
>3%
1-3%
<1%
336211
Secondary manufacturer:
Motor vehicle body manufacturing
1,250
employees
217
0
41
176
336213
Secondary manufacturer:
Motor home manufacturing
1,250
employees
32
0
7
25
Total secondary manufacturer
249
0
48
201
$8.0 million
811198
Alternative fuel engine converters
annual
receipts
2
0
0
2
TOTAL
251
0
48
203
Chapter 11 References
1 U.S. Small Business Administration. Table of Small Business Size Standards Matched to North
American Industry Classification System Codes. 2019 Version. Available online:
https://www.sba.gov/sites/default/files/2019-
08/SBA%20Table%20of%20Size%20Standards_Effective%20Aug%2019%2C%202019_Rev.p
df
2 NAICS Association. NAICS & SIC Identification Tools. Available online:
https://www.naics.com/search
438
-------
439
-------
440
------- |