EPA420-D-99-001
Draft Regulatory Impact Analysis
Control of Air Pollution from
New Motor Vehicles:
Tier 2 Motor Vehicle Emissions Standards
and Gasoline Sulfur Control Requirements
U.S. Environmental Protection Agency
Office of Air and Radiation
Office of Mobile Sources
April 1999
-------
-------
Executive Summary
Executive Summary
Proposed Vehicle Standards
Today's notice proposes new federal emission standards ("Tier 2 standards") for passenger
cars and light trucks. The program is designed to reduce vehicle emissions of nitrogen oxides
(NOx) and non-methane organic gases (NMOG) (which consist primarily of hydrocarbons (HC)
and volatile organic compounds (VOCs)); NOx and NMOG contribute to the formation of ozone
and particulate matter (PM) which are harmful air pollutants. The program would also, for the
first time, apply the same federal standards to passenger cars and all light trucks ("light LDTs"
and "heavy LDTs").
The proposed Tier 2 standards would reduce new vehicle NOx levels to an average of 0.07
grams per mile (g/mi). For new passenger cars and light LDTs, these standards would phase in
beginning in 2004, with the standards to be fully phased in by 2007. For heavy LDTs, the
proposed Tier 2 standards would be phased in beginning in 2008, with full compliance in 2009.
During the phase-in period from 2004-2007, all passenger cars and light LDTs not certified to
Tier 2 standards would have to meet an interim average standard of 0.30 g/mi NOx, equivalent to
the current NLEV standards for LDVs. During the period 2004-2008, heavy LDTs not certified
to Tier 2 standards would phase in an average standard of 0.20 g/mi NOx, with an emissions cap
of 0.60 g/mi NOx
Manufacturers would be allowed to comply with the very stringent proposed new standards
in a flexible way, assuring that the average emissions of a company's production met the target
emission levels while allowing the manufacturer to choose from several more- and less-stringent
emission categories for certification. The proposed requirements also include more stringent PM
standards, which primarily affect diesel vehicles, and more stringent hydrocarbon controls
(exhaust NMOG and evaporative emissions standards).
Proposed Gasoline Sulfur Requirements
The other major part of today's proposal would significantly reduce average gasoline sulfur
levels nationwide beginning in 2004, and likely earlier due to the proposed incentive program to
encourage early sulfur reductions. Refiners would generally install advanced refining equipment
to remove sulfur in their refining processes. Importers of gasoline would be required to import
and market only gasoline meeting the proposed sulfur limits. Temporary, less stringent
standards would apply to certain small refiners.
EPA is proposing that gasoline produced by refiners and sold by gasoline importers
generally meet an average sulfur standard of 30 ppm and a cap of 80 ppm. The proposed
program builds upon the existing regulations covering gasoline content as it relates to emissions
performance. It includes provisions for trading of sulfur credits, increasing the flexibility
in
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
available to refiners for complying with the new requirements. The proposed credit program is
intended to ease compliance uncertainties by providing refiners the flexibility to phase in early
controls in 2000-03 and use credits generated in these years to delay some control to 2006. As
proposed, the program would achieve expected environmental benefits while providing
substantial flexibility to refiners. The effect of the credit program is that those refiners that
participate would have the opportunity for more overall lead-time to attain the final sulfur levels.
Cost-Effectiveness of the Proposed Tier2/Sulfur Program
A comparison of the costs of our proposed program with the emission reductions it is
estimated to achieve leads us to conclude that it is a cost-effective means of reducing pollution.
The cost-effectiveness of the Tier 2/gasoline sulfur proposal, considering only the NOx and
hydrocarbon reductions which it will yield, ranges from $1,800 to $2,180 per ton. This range
compares favorably with other mobile and stationary source controls. For example, both the Tier
1 and NLEV vehicle standards had similar cost-effectiveness to the standards we are proposing
today. For stationary sources, similar levels of reductions in NOx and hydrocarbon emissions
could cost up to $10,000 per ton. We believe that the program we are proposing today will be an
efficient and significant step towards reaching attainment and maintenance of the NAAQS.
Highlights of the Benefit-Cost Analysis
We also made an assessment of the monetary value of the health and general welfare
benefits that would result from our proposed standards. This assessment made use of many of
the same data sets, models, and assumptions already used in previous EPA rulemakings. As a
result, our benefits assessment included methods which have already received review by the
public, other Federal agencies, and/or the independent Science Advisory Board.
In our benefits assessment, we estimated that our proposed standards would, in the long
term, result in the yearly avoidance approximately 800 to 2400 premature deaths, approximately
4700 to 8000 cases of bronchitis, and significant numbers of hospital visits, lost work days, and
an assortment of respiratory ailments. Our proposed standards will also produce welfare benefits
relating to agricultural crop damage, visibility, and nitrogen deposition in rivers and lakes. The
results indicate that, based on the particular assumptions, models, and data used in this
preliminary benefit-cost analysis, the range of monetary benefits realized after full turnover of
the fleet to Tier 2 vehicles would be approximately 3.3 billion to 19.5 billion dollars per year.
Comparing this estimate of the economic benefits with the adjusted cost estimate indicates that
the net economic benefit of the proposed standards to society could be from a net cost of 0.2
billion to a net benefit of 16.0 billion dollars per year. Our benefit-cost analysis should be
considered preliminary due to limitations in the data and models available for analysis in
advance of today's proposal. Additional, more refined analysis will be conducted prior to
issuance of final standards.
IV
-------
Table of Contents
Table of Contents
Executive Summary iii
List of Acronyms xi
List of Tables xv
List of Figures xx
Chapter I: Introduction 1-1
A. Background 1-1
B. Overview of the Proposal 1-2
1. Vehicle Emission Standards 1-3
2. Gasoline Sulfur Standards 1-3
Chapter II: Health and Welfare Concerns II-l
A. Ozone II-l
1. Health and Welfare Effects of Ozone II-2
B. Paniculate Matter II-4
1. Health and Welfare Effects of Particulate Matter II-5
C. Carbon Monoxide II-6
D. Visibility and Regional Haze II-7
Chapter II. References II-9
Chapter III: Environmental Impact III-l
A. Inventory Impacts of Tier 2/Sulfur III-l
1. NOx III-5
a. Light-Duty NOx Trends Without Tier 2/Sulfur III-5
b. NOx Reductions Due To Tier 2/Sulfur 111-10
c. NOx Emission Reductions From Other Options Ill-15
2. VOC 111-17
a. Light-Duty VOC Trends Without Tier 2/Sulfur Ill-17
b. VOC Reductions Due To Tier 2/Sulfur 111-21
c. VOC Emission Reductions From Other Options 111-26
3. SOx 111-27
a. Light-Duty SOx Trends Without Sulfur Control 111-27
b. SOx Reductions Due To Sulfur Control 111-29
4. Particulate Matter 111-30
a. "No Growth" Diesel Sales Scenario III-31
b. "Increased Growth" Sales Scenario 111-35
B. Air Quality Measures 111-39
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
1. Ozone 111-40
a. Measures of Current Attainment and Non-attainment 111-40
b. General Description of Ozone Modeling in the OTAG Domain to Estimate the
Effect of Tier 2/Sulfur Controls 111-42
c. The "Rollback Method" for Estimating Design Values Resulting from Control
Measures 111-43
d. Specific Simulations Used to Evaluate Tier 2/Sulfur 111-46
e. Results of the NOx-only and VOC-only Runs 111-46
f. Details of the Tier 2/Sulfur Ozone Modeling Runs 111-47
2. Visibility/Regional Haze 111-49
C. Air Toxics 111-50
1. Health Effects 111-51
a. Benzene 111-51
b. 1,3-Butadiene 111-53
c. Formaldehyde 111-53
d. Acetaldehyde 111-54
e. Diesel Paniculate Matter 111-55
2. Assessment of Emissions and Exposure 111-57
a. Emissions Modeling 111-57
b. Nationwide Toxic Emissions Estimates - Baseline Scenario 111-62
c. Exposure - Baseline Scenario 111-63
d. Impact of Potential Vehicle and Fuel Controls 111-65
e. Limitations 111-69
Chapter III. References 111-70
Chapter IV: Technological Feasibility IV-1
A. Feasibility of Tier 2 Exhaust Emission Standards for LDVs and LDTs IV-1
1. NMOG and NOx Emissions from Gasoline-Fueled Vehicles IV-1
a. Technology Description IV-2
b. Data Supporting Tier 2 Technical Feasibility IV-14
c. Lean-Burn Technology IV-25
2. CO Emissions from Gasoline Fueled Vehicles IV-25
3. Formaldehyde Emissions from Gasoline Fueled Vehicles IV-26
4. Evaporative Emissions IV-27
5. Diesel Vehicles IV-28
B. Feasibility of Removing Sulfur from Gasoline IV-29
1. Source of Gasoline Sulfur IV-29
2. Current Levels of Sulfur in Gasoline IV-30
3. Feasibility of Meeting the Proposed Gasoline Sulfur Standards IV-32
4. Meeting a Low Sulfur Gasoline Standard IV-33
5. Improved Gasoline Desulfurization Technology IV-37
6. Expected Desulfurization Technology to be Used by Refiners IV-40
7. Feasibility for a Low Gasoline Sulfur Standard in 2004 IV-40
8. Phase In of Compliance with the Proposed Sulfur Standards IV-45
vi
-------
Table of Contents
Chapter IV. References IV-51
Chapter V: Economic Impact V-l
A. Impact of Tier 2 Standards on Vehicle Costs V-l
1. Manufacturer Costs for Tier 2 Vehicles V-l
a. Methodology V-l
b. Hardware Costs for Exhaust Emissions Control V-3
c. Hardware Costs for Evaporative Emissions Control V-l5
d. Assembly Costs V-17
e. Development and Capital Costs V-17
f. Total Near-term and Long-term Manufacturer Costs V-l9
2. Tier 2 Vehicle Consumer Costs V-22
3. Annual Total Nationwide Costs for Tier 2 Vehicles V-23
a. Overview of Nationwide Vehicle Costs V-23
b. Methodology V-24
c. Estimates of Total Nationwide Vehicle Costs by Vehicle Class V-26
B. Gasoline Desulfurization Costs V-30
1. Methodology V-30
a. Cost Inputs V-33
b. Determination of Blendstock Sulfur Levels V-41
2. The Cost of Desulfurizing Gasoline V-46
a. The Cost of the Averaging Standard V-46
b. Verification of the Desulfurization Cost Based on the Improved
Technologies V-50
c. Future Cost of Desulfurization V-52
d. Comparison with Previous Cost Estimates V-55
3. Other Effects of This Program V-58
a. Effect of the Cap Standard V-58
b. Other Effects on the Refining Industry V-60
c. Refinery Energy and Global Warming Impacts V-62
4. Per Vehicle Life-Cycle Fuel Costs V-70
5. Aggregate Annual Fuel Costs V-73
a. Methodology V-74
b. Explanation of Results V-75
C. Combined Vehicle and Fuel Costs V-80
1. Combined Costs Per Vehicle V-80
2. Combined Total Annual Nationwide Costs V-80
Chapter V. References V-83
Chapter VI: Cost-Effectiveness VI-1
A. Overview of the Analysis VI-1
1. Temporal and Geographic Applicability VI-1
2. Baselines VI-2
vii
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
B. Costs VI-4
1. Near and Long-Term Cost Accounting VT-4
2. Vehicle and Fuel Costs VI-5
3. Cost Crediting for PM and SO2 VI-5
C. Emission Reductions VI-7
1. NOx and NMHC VI-7
2. Irreversibility VI-11
3. Primary Particulate Matter VI-13
4. Sulfur Dioxide VI-14
D. Results VI-15
APPENDIX VI-A : Discounted Lifetime Tonnage Values for Exhaust Emissions VT-20
APPENDIX VI-B : Discounted Lifetime Tonnage Values for Evaporative Emissions . VI-24
Chapter VI. References VI-25
Chapter VII: Benefit-Cost Analysis VII-1
A. Emissions VII-2
B. Air Quality Impacts VII-6
1. Ozone Air Quality Estimates VII-6
a. Modeling Domain VII-7
b. Simulation Periods VII-8
c. UAM-V Model Output VII-8
d. Converting Episode Estimates to Full-Season Profiles VII-8
e. Extrapolating from Monitored to Unmonitored Locations VII-9
f. Ozone Air Quality Results VII-10
2. PM Air Quality Estimates VII-11
a. Climatological Regional Dispersion Model VII-12
b. Development of the S-R Matrix VII-12
c. Fugitive Dust Adjustment Factor VII-13
d. Normalizing S-R Matrix Results to Measured Data VII-14
e. Development of Annual Median PM25 Concentrations VII-15
f. PM Air Quality Results VII-15
3. Nitrogen Deposition Estimates VII-16
4. Visibility Degradation Estimates Using the S-R Matrix VII-18
C. Benefits Assessment VII-19
1. Overview of Benefits Estimation VII-20
2. Issues in Estimating Changes in Health Effects VII-27
a. Baseline Incidences VII-31
b. Thresholds VII-31
3. PM- and Ozone-related Health Effects VII-31
a. Premature Mortality VII-32
b. Chronic Bronchitis VII-37
c. Hospital Admissions VII-39
viii
-------
Table of Contents
d. Acute Bronchitis VII-43
e. PM-related Upper Respiratory Symptoms VII-43
f. PM-related Lower Respiratory Symptoms VII-44
g. Ozone-related Any of 19 Respiratory Symptoms VII-45
h. Work Loss Days VII-46
i. Minor Restricted Activity Days VII-46
j. Worker Productivity VII-47
4. Ozone- and PM-Related Welfare Effects VII-48
a. Commodity Agricultural Crops VII-48
b. Nitrogen Deposition VII-49
c. Household Soiling Damage VII-52
d. Visibility VII-53
e. Ozone- and PM-related Welfare Effect Benefits Estimation VII-57
5. Total Aggregated Benefits VII-57
6. Sensitivity Analyses VII-58
7. Limitations of the Analysis VII-60
a. PM Mortality Risk and Health Effects VII-62
b. Unquantifiable Benefits VII-63
c. Potential Disbenefits VII-66
d. Projected Income Growth VII-68
D. Cost VII-68
Chapter VII References VII-70
Chapter VIII: Regulatory Flexibility VIII-1
A. Requirements of the Regulatory Flexibility Act VIII-1
B. Description of Affected Entities VIII-2
1. Small Refiners VIII-3
2. Small Petroleum Marketers VIII-4
3. Small Certifiers of Covered Vehicles VIII-4
C. Projected Costs of the Proposed Gasoline Sulfur Standards VIII-5
D. The Types and Number of Small Entities to Which the Proposed Rule
Would Apply VIII-5
E. Projected Reporting, Recordkeeping, and Other Compliance Requirements of the
Proposed Rule VIII-6
F. Other Relevant Federal Rules Which May Duplicate, Overlap, or Conflict with the
Proposed Rule VIII-7
G. Regulatory Alternatives VIII-7
1. Small Refiners VIII-7
a. Interim Sulfur Standards VIII-8
b. Hardship Relief VIII-10
2. Small Certifiers of Covered Vehicles VIII-11
Chapter VIII. References VIII-13
IX
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Appendix A: 47-State and Four-Cities Analyses A-l
Appendix B: Evidence Supporting the Irreversibility of Sulfur's Emission Impact B-l
A. Exhaust Emission Sensitivity to Sulfur Content B-l
B. Theory Supporting the Reversibility and Irreversibility of Sulfur's Emission Impact B-3
C. Results of Sulfur Reversibility Test Programs B-6
1. Coordinating Research Council (CRC) Sulfur Reversibility Program B-6
2. American Petroleum Institute Sulfur Reversibility Program B-8
3. Johnson Matthey Sulfur Reversibility Program B-l3
4. Other Testing B-15
D. Criteria for Evaluating Sulfur Reversibility Data B-15
E. Projected Levels of Sulfur Irreversibility In-Use B-l8
Appendix C: One-Hour and Eight-Hour County Design Values C-l
Appendix D: EPA's Legal Authority for Proposing Gasoline Sulfur Controls D-l
A. EPA's Current Regulatory Requirements for Gasoline D-l
B. How the Proposed Gasoline Sulfur Control Program Meets the CAA Section 21 l(c)
Criteria D-2
1. Health and Welfare Concerns of Air Pollution Caused by Sulfur in Gasoline . D-2
2. Impact of Gasoline Sulfur Emission Products on Emission Control Systems . D-3
3. Sulfur Levels that Tier 2 Vehicles Can Tolerate D-4
4. Sulfur Sensitivity of Other Catalysts D-6
5. Effect of Gasoline Sulfur Control on the Use of Other Fuels or Fuel Additives D-7
-------
Acronyms, Tables, and Figures
List of Acronyms
A/F
AML
ANPRM
API
ASTM
bbl
BCA
BTU
C-R
CAA or the Act
CAP
CARS
CASAC
CG
CML
CO
COI
COPD
cpsi
CRC
CRDM
DOE
EGR
EHC
EIA
EPA or the Agency
ERIC
FCC
FTP
GDI
GVWR
HAPEM
air/fuel ratio
acute myeloid leukemia
Advanced Notice of Proposed Rulemaking
American Petroleum Institute
American Society for Testing and Materialss
barrel
benefit-cost analysis
British Thermal Unit
concentration response
Clean Air Act
Compliance Assurance Program (2000)
California Air Resources Board
Clean Air Science Advisory Committee
conventional gasoline
chronic myeloid leukemia
carbon monoxide
cost of illness
chronic obstructive pulmonary disease
cells per square inch
Coordinating Research Council
Climatological Regional Dispersion Model
U.S. Department of Energy
exhaust gas recirculation
electrically heated catalyst
Energy Information Administration
U.S. Environmental Protection Agency
Emissions Reduction and Intercept Control (system)
fluidized catalytic cracker
Federal Test Procedure
gasoline direct injection
gross vehicle weight rating
Hazardous Air Pollutant Exposure Model
XI
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
HC
HDV
HEGO
I/M
ICI
IRFA
LOT
LDV
LEV
LPG
MDV
MECA
MLE
MRAD
MSCF
MTBE
NAAQS
NAPAP
NFRAQS
NLEV
NMHC
NMOG
NO2
NOx
NPC
NPRA
NPRM
OAQPS
OBD
OMB
QMS
QMS
ORNL
OSTP
OTAG
hydrocarbon
heavy-duty vehicle
heated exhaust gas oxygen (sensor)
inspection/maintenance
independent commercial importer
initial regulatory flexibility analysis
light-duty truck
light-duty vehicle
low emission vehicle
liquid petroleum gas
medium-duty vehicle
Manufacturers of Emission Controls Association
maximum likelihood estimate
minor restricted activity days
thousand standard cubic feet
methyl tertiary-butyl ether
National Ambient Air Quality Standards
National Acid Precipitation Assessment Program
Northern Front Range Air Quality Study
national low emission vehicle
non-methane hydrocarbons
non-methane organic gases
nitrogen dioxide
oxides of nitrogen
National Petroleum Council
National Petrochemical & Refiners Association
Notice of Proposed Rulemaking
Office of Air Quality Planning and Standards
on-board diagnostics
Office of Management and Budget
Office of Mobile Sources
Office of Mobile Sources
Oak Ridge National Laboratory
(White House) Office of Science and Technology Policy
Ozone Transport Assessment Group
xn
-------
Acronyms, Tables, and Figures
PADD
PCM
Pd
PM
PNGV
ppm
Pt
R&D
RFA
RfC
RFG
Rh
ROI
ROTR
RPE
RVP
S-R Matrix
S&P DRI
SAB
SBA
SBARP or the Panel
SBREFA
SCR
SER
SFTP
SIC
SIGMA
SIP
SO2
SOx
SULEV
SVM
SVM
SwRI
TOG
Petroleum Administrative Districts for Defense
powertrain control module
palladium
particulate matter
Partnership for a New Generation of Vehicles
part per million
platinum
research and development
Regulatory Flexibility Act
reference concentration
reformulated gasoline
rhodium
return on investment
Regional Ozone Transport Rule
retail price equivalent
Reid Vapor Pressure
Source-Receptor Matrix
Standard & Poor's Data Research International
Science Advisory Board
U.S. Small Business Administration
Small Business Advocacy Review Panel
Small Business Regulatory Enforcement Fairness Act
selective catalytic reduction
Small Entity Representative
Supplemental Federal Test Procedure
Standard Industrial Classification
Society of Independent Gasoline Marketers of America
State Implementation Plan
sulfur dioxide
oxides of sulfur
super ultra low emission vehicle
small volume manufacturer
small volume manufacturer (of vehicles)
Southwest Research Institute
total organic gases
Xlll
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
TW
UAM
UCL
UEGO
ULEV
UV
VMT
VNA
voc
WLD
WTP
test weight
Urban Airshed Model
upper confidence limit
universal exhaust gas oxygen (sensor)
ultra low emission vehicle
ultra violet
vehicle miles traveled
Voronoi Neighbor Averaging
volatile organic compound
work loss days
willingness to pay
xiv
-------
Acronyms, Tables, and Figures
List of Tables
Table III-l. 47-State Light Duty NOx Emissions Without Tier 2/Sulfur III-6
Table III-2. Four-City Light-Duty NOx Emissions Without Tier 2/Sulfur III-7
Table III-3. Light-Duty Contribution to Total NOx Inventory Without Tier 2/Sulfur Ill-10
Table III-4. 47-State Light-Duty NOx Reductions Due To Tier 2/Sulfur Ill-10
Table III-5. Four-City Light-Duty NOx Emissions With Tier 2/Sulfur 111-12
Table III-6. Percent Reduction From Light-Duty and
Total Baseline NOx Emissions in Four Cities 111-13
Table III-7. Light-Duty Contribution to Total NOx Inventory With Tier 2/Sulfur Ill-15
Table III-8. 47-State Light-Duty NOx Reductions From Alternative Control Options .... Ill-16
Table III-9. NOx Reduction Shortfall From Alternative Control
Options Relative to Today's Proposal 111-17
Table III-10. 47-State Light-Duty VOC Emissions Without Tier 2/Sulfur Ill-18
Table III-l 1. Four-City Light-Duty VOC Emissions Without Tier 2/Sulfur Ill-19
Table 111-12. Light-Duty Contribution to Total VOC Inventory Without Tier 2/Sulfur . . . 111-21
Table 111-13. 47-State Light-Duty VOC Reductions Due to Tier 2/Sulfur 111-21
Table 111-14. Four-City Light-Duty VOC Reductions Due To Tier 2/Sulfur 111-23
Table III-l5. Percent Reduction From Light-Duty and
Total Baseline VOC Emissions in Four Cities 111-24
Table 111-16. Light-Duty Contribution to Total VOC Inventory With Tier 2/Sulfur 111-26
Table 111-17. 47-State Light-Duty VOC Reductions From Alternative Control Options . . . 111-26
Table III-l8. VOC Reduction Shortfall From Alternative
Control Options Relative to Today's Action 111-27
Table 111-19. 47-State SOx Emissions Without Sulfur Control 111-28
Table 111-20. 47-State Light-Duty SOx Reductions Due To Sulfur Control 111-29
Table 111-21. 47 State Light-Duty Direct Exhaust PM2 5 Emissions Without Tier 2/Sulfur
No Growth in Diesel Sales 111-32
Table 111-22. 47-State Light-Duty Direct Exhaust PM2 5 Reductions Due To Tier 2/Sulfur
No Growth in Diesel Sales 111-33
Table 111-23. 47-State Light-Duty PM10 Emissions With and Without Tier 2/Sulfur Control
No Growth in Diesel Sales 111-35
Table 111-24. Diesel LDT Sales Penetration Under Increased Growth Scenario 111-36
Table 111-25. 47 State Light-Duty Direct Exhaust PM25 Emissions Without Tier 2/Sulfur
Increased Diesel Growth Scenario 111-37
Table 111-26. 47-State Light-Duty Direct Exhaust PM2 5 Reductions Due To Tier 2/Sulfur
Increased Diesel Growth Scenario 111-38
Table 111-27. Percent Reductions for Tier 2/Sulfur Ozone Modeling Runs 111-46
Table 111-28. Percentage Reductions from the 2007 Post-ROTR Inventory of NOX and NMHC
111-48
Table 111-29. NFRAQS Compositional Analysis of PM25 Samples 111-49
Table 111-30. Percentage of PM25 Coming from Gasoline Vehicles 111-49
Table 111-31. Percentage of Total PM25 From Gasoline Vehicles 111-50
Table 111-32. Example of Data File Format for Toxic Adjustment Factors 111-59
xv
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table 111-33. Average Nationwide Highway Vehicle Toxic Emission Rates (mg/mi)
In 1990, 1996, 2007, and 2020, for Baseline Scenarios 111-62
Table 111-34. Average Nationwide Highway Vehicle Toxic Exposure (|ig/m3)
In 1990, 1996, 2007, and 2020, for Baseline Scenarios 111-64
Table 111-35. Average Nationwide Highway Vehicle
Toxic Emission Rates (mg/mi) in 2007, for Various Scenarios 111-67
Table 111-36. Average Nationwide Highway Vehicle
Toxic Emission Rates (mg/mi) in 2020, for Various Scenarios 111-67
Table 111-37. Average Nationwide Highway Vehicle Toxic Exposures for the Entire Population
(|ig/m3) in 2007, for Various Scenarios 111-68
Table 111-38. Average Nationwide Highway Vehicle Toxic Exposures for the Entire Population
(|ig/m3) in 2020, for Various Scenarios 111-68
Table IV-1. Emission Control Hardware and Techniques Projected to Meet Proposed Tier 2
Vehicle Standards IV-2
Table IV-2. Number of Engine Families with One or More Engine/Vehicle Configurations with
Low Full-life NOx Levels IV-15
Table IV-3. 1999 Engine Families with One or More Vehicle Configurations
with Full-life NOx Certification Levels at or below 0.07 g/mile NOx IV-15
Table IV-4. MECA Test Program: Emissions with Catalysts Aged to 100,000 Miles .... IV-19
Table IV-5. ARE Production LEV LDV Passenger Car Emission Data IV-19
Table IV-6. ARE Modified Passenger Car Emission Data IV-20
Table IV-7. ARB Ford Expedition Baseline Emission Test Results IV-20
Table IV-8. ARB Expedition Emission Results with Advanced Catalyst Systems IV-21
Table IV-9. Emission Technology Gaps Between LDVs and LDTs IV-23
Table IV-10. CO Emissions from California LEVs IV-26
Table IV-11. Formaldehyde Emissions from California LEVs IV-27
Table IV-12. Average Sulfur Levels by PADD and for the Nation IV-32
Table IV-13. Leadtime Required Between Promulgation of the Final Rule and Implementation
of the Gasoline Sulfur Standard (years) IV-41
Table IV-14. Distribution of Refineries by Current Gasoline Sulfur Level IV-46
Table IV-15. Effect of Phase-In of Sulfur Control to Meet Sulfur Caps in 2004 and 2005 . IV-49
Table V-l. Main or Underfloor Catalyst Cost Breakdown V-6
Table V-2. Total Estimated Per Vehicle Manufacturer
Incremental Hardware Costs for the Tier 2 Standards V-9
Table V-3. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDV Compared to
NLEV LDV V-10
Table V-4. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT1 Compared to
NLEV LDT1 V-l 1
Table V-5. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT2 Compared to
NLEV LDT2 V-12
Table V-6. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT3 Compared to
Tier 1 LDT3 V-13
Table V-7. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT4 Compared to
Tier 1 LDT4 V-14
Table V-8. Potential Evaporative Improvements and Their Costs to Manufacturers V-l6
xvi
-------
Acronyms, Tables, and Figures
Table V-9. Per Vehicle Fixed Costs V-19
Table V-10. Total Per Vehicle Manufacturer Costs - Near Term V-20
Table V-l 1. Long-term Total Incremental Per Vehicle Manufacturer Costs V-22
Table V-12. Incremental Per Vehicle Costs to Consumers for Tier 2 Vehicles V-23
Table V-13. Estimated Annual Nationwide Costs V-24
Table V-14. Estimated Annual 49-State Vehicle Sales V-25
Table V-l 5. Projected Overall Industry Phase-in of Tier 2 Vehicles and Improved Evaporative
Emissions Controls For Purposes of the Aggregate Cost Analysis V-26
Table V-16. Annual Nationwide Costs For Tier 2 LDVs V-27
Table V-17. Annual Nationwide Costs For Tier 2 LDTls V-27
Table V-l8. Annual Nationwide Costs For Tier 2 LDT2s V-28
Table V-19. Annual Nationwide Costs For Tier 2 LDT3s V-28
Table V-20. Annual Nationwide Costs For Tier 2 LDT4s V-29
Table V-21. Annual Nationwide Costs For Tier 2 LDVs and LDTs V-29
Table V-22. Raw Material and Utility Needs and Desulfurization Capabilities for Several
Desulfurization Technologies V-34
Table V-23. Offsite Factors for Different Sized Refineries V-36
Table V-24. Labor Costs in Selected Cities V-37
Table V-25. Capital Cost Factors Which Vary by PADD V-37
Table V-26. Economic Cost Factors Used and the Resulting Capital Amortization Factor . V-39
Table V-27. Summary of Costs Taken From EIA and NPC Data Tables V-40
Table V-28. PADD 1 Blendstock Sulfur Levels and Gasoline Pool Fraction V-43
Table V-29. PADD 2 Blendstock Sulfur Levels and Gasoline Pool Fraction V-44
Table V-30. PADD 3 Blendstock Sulfur Levels and Gasoline Pool Fraction V-44
Table V-31. PADD 4 Blendstock Sulfur Levels and Gasoline Pool Fraction V-45
Table V-32. PADD 5 Outside of California Blendstock
Sulfur Levels and Gasoline Pool Fraction V-45
Table V-33. Projected Volume of Gasoline Produced by an Average Refinery in each PADD, by
Each PADD of Refineries and for the U.S.* in 2004 V-46
Table V-34. Per-Gallon Cost of Desulfurizing Gasoline V-47
Table V-35. Estimated Average Per-Refinery Capital and Operating Cost of Desulfurizing
Gasoline to 150 ppm and 30 ppm V-48
Table V-36. Aggregate Cost of Desulfurizing Gasoline to 150 ppm and 30 ppm V-49
Table V-37. Summary of the Cost of Desulfurization by the CDTECH Process Based on 90
percent Desulfurization Severity V-51
Table V-38. Projected Future Average Per-Gallon National Cost of Desulfurizing
Gasoline to 30 ppm V-55
Table V-39. Cost of Desulfurizing Gasoline by Refiners in PADDs 1 & 3 Reported in the "EPA
Staff Paper on Gasoline Sulfur Issues," and our Current Costs V-56
Table V-40. A Comparison of the Per-Gallon Gasoline Desulfurization Cost of Improved
Desulfurization Technologies to that of the Earlier Mobil Octgain Process for PADD 3
V-57
Table V-41. Energy Consumed by PADD 3 Refineries in 1994, Projected to 1997 V-63
Table V-42. Estimated Yearly Energy and Hydrogen Demand of CDTECH
Desulfurization Units in PADD 3 V-64
xvii
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-43. Estimated Yearly Energy and Hydrogen Demand of OCTGAIN
Desulfurization Units in PADD 3 V-65
Table V-44. Undiscounted Per-vehicle Costs of
Low Sulfur Gasoline V-71
Table V-45. Discounted Per-vehicle Costs of
Low Sulfur Gasoline V-72
Table V-46. Fleet Average Per-vehicle Costs V-73
Table V-47. Increased Annualized Fuel Cost as a Result of Today's Proposed Tier 2 Gasoline
Sulfur Controls V-74
Table V-48. Stratification of Heavy-Duty Gasoline Fleet using Vehicle Count Projections V-77
Table V-49. Calculation of Gasoline Consumption V-78
Table V-50. Aggregate Annualized Fuel Costs per Year from 2004 to 2020 V-79
Table V-51. Total Incremental Per Vehicle Costs to Consumers
Over the Life of a Tier 2 Vehicle V-80
Table V-52. Total Annualized Costs to the Nation for Tier 2 Vehicles
and Low Sulfur Gasoline V-82
Table VI-1. Fleet-average, Per-vehicle Costs Used in Cost-effectiveness VI-5
Table VI-2. Fleet Average Per-vehicle Costs Used in Cost-effectiveness VI-7
Table VI-3. Weighting Factors for NOx andNMHC Lifetime Tonnage Values VT-9
Table VI-4. Vehicle Class Sales Weighting Factors VI-10
Table VI-5. Fleet-average, Per-vehicle Discounted
Lifetime Tons for the NLEV Baseline VI-10
Table VI-6. Fleet-average, Per-vehicle Discounted Lifetime Tons
for Proposed Tier 2 Standards VT-10
Table VI-7. Fleet-average, Per-vehicle Discounted Lifetime Tons
Used in Cost-effectiveness Analysis VI-13
Table VI-8. Cost-effectiveness of the Proposed Standards VI-15
Table VI-9. Alternative program options evaluated by EPA VI-16
Table VI-10. Cost-effectiveness of Previously Implemented
Mobile Source Programs (Costs Adjusted to 1997 Dollars) VI-17
Table VII-1. Emission Estimates by Vehicle Type and Reductions Associated with Adoption of
the Tier 2 Rule VII-5
Table VII-2. Summary of UAM-V Derived Hourly Ozone Air Quality VII-11
Table VII-3. Summary of S-R Matrix Derived PM Air Quality VII-16
Table VII-4. Summary of 2010 Nitrogen Deposition in Selected Estuaries VII-17
Table VII-5. Summary of 2010 Visibility Degradation Estimates VII-19
Table VII-6. Unquantified Benefit Categories VII-25
Table VII-7. Quantified and Monetized Primary Health and Welfare Effects VII-26
Table VII-8. Key Differences Between Low and High Assumption Sets VII-27
Table VII-9. PM and Ozone Health Concentration-Response Function Summary Data . . VII-29
Table VII-10. Summary of Mortality Valuation Estimates VII-36
Table VII-11. Derivation of Cost of Illness (COI) and Total WTP Estimates for Hospital
Admissions Endpoints VII-42
Table VII-12. Quantified Welfare Effects Included in the Benefits Analysis VII-48
Table VII-13. Class I Areas Included in Visibility Study By Region VII-55
xviii
-------
Acronyms, Tables, and Figures
Table YD-14. Estimated Coefficients Used in the Valuation of
WTP for Improved Visibility VII-56
Table VII-15. Welfare Endpoint Monetary Benefits VII-57
Table VII-16. Avoided Incidence and Monetized Benefits Associated with the
Tier 2 Rule for a Range of Assumption Sets VII-59
Table VII-17. Sources of Uncertainty in the Benefit Analysis VII-61
Table VII-18. PM Health Effects and Benefits (No Lag and Lag of up to Five Years) . . . VII-63
Table VII-19. Adjusted Cost for Comparison to Benefits VII-69
Table VIII-1. Industries Containing Small Businesses
Potentially Affected by Today's Proposed Rule VIII-3
Table VIII-2. Costs for a 19,000 bbls gasoline/day
Refinery to Produce 30 ppm Gasoline VIII-5
Table VIII-3. Types and Number of Small Entities to
Which the Proposed Tier 2/Gasoline Sulfur Rule Would Apply VIII-6
Table VIII-4. Federal Gasoline Sulfur Program with Sulfur
Standards of 30 ppm on Average and an 80 ppm Per-Gallon Cap VIII-9
Table VIII-5. Federal Gasoline Sulfur Program with Sulfur
Standards Above 30 ppm on Average and an 80 ppm Per-Gallon Cap VIII-9
xix
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
List of Figures
Figure III-l. 47-State Light-Duty NOx Emissions Without Tier 2/Sulfur (Annual Tons) . . . III-6
Figure III-2: Breakdown of Total 2020 47 State NOx Inventory Without Tier 2 III-9
Figure III-3. Breakdown of Total 2020 Atlanta NOx Inventory Without Tier 2 III-9
Figure III-4. 47-State Light-Duty NOx Emissions With Tier 2/Sulfur III-l 1
Figure III-5. Breakdown of Total 2020 47-State NOx Inventory With Tier 2/Sulfur 111-14
Figure III-6. Breakdown of Total 2020 Atlanta NOx Inventory With Tier 2/Sulfur 111-14
Figure III-7. 47-State Light-Duty VOC Emissions Without Tier 2/Sulfur Ill-18
Figure III-8. Breakdown of Total 2020 47-State VOC Inventory Without Tier 2/Sulfur . . 111-20
Figure III-9. Breakdown of Total 2020 Atlanta VOC Inventory Without Tier 2/Sulfur . . . 111-20
Figure III-10. 47-State Light-Duty VOC Emissions With Tier 2/Sulfur 111-22
Figure III-l 1. Breakdown of Total 2020 47 State VOC Inventory With Tier 2/Sulfur 111-25
Figure 111-12. Breakdown of Total 2020 Atlanta VOC Inventory With Tier 2/Sulfur 111-25
Figure 111-13. 47-State Light-Duty SOx Emissions Without Sulfur Control 111-28
Figure 111-14. 47-State Light-Duty SOx Emissions With Tier 2/Sulfur 111-30
Figure III-l5. 47-State Light-Duty Direct Exhaust PM25 Emissions Without Tier 2/Sulfur -
No Diesel Growth 111-32
Figure III-16. 47-State Light-Duty Direct Exhaust PM25 Emissions With Tier 2/Sulfur -
No Diesel Growth 111-33
Figure III-17. 47-State Light-Duty Direct Exhaust PM25 Without Tier 2/Sulfur -
Increased Diesel Sales 111-37
Figure 111-18. 47-State Light-Duty Direct Exhaust PM25 Emissions With Tier 2/Sulfur Increased
Diesel Growth 111-38
Figure III-19. Example Plot of Target Fuel Benzene Versus
Baseline Fuel TOG under FTP Conditions 111-60
Figure IV-1. Impact of Coating Architecture on HC and NOx Emissions IV-10
Figure IV-2. Map of U.S. Petroleum Administrative Districts for Defense IV-31
Figure IV-3. Diagram of a Typical Complex Refinery IV-34
Figure V-l. Distribution of Progress Ratios V-21
Figure V-2. Cost of Reducing Gasoline Sulfur in PADD 1 V-67
Figure V-3. Cost of Reducing Gasoline Sulfur in PADD 2 V-67
Figure V-4. Cost of Reducing Gasoline Sulfur in PADD 3 V-68
Figure V-5. Cost of Reducing Gasoline Sulfur in PADD 4 V-68
Figure V-6. Cost of Reducing Gasoline Sulfur in PADD 5 Outside of California V-69
Figure V-7. National Cost of Reducing Gasoline Sulfur Outside of California V-69
Figure V-8. Total Annualized Costs of Tier 2 Vehicles and Low Sulfur Gasoline V-81
Figure VII-1. UAM-V Modeling Domain for Eastern U.S VII-7
Figure VII-2. VNA Spatial Interpolation VII-9
Figure VII-3. Example Benefits Analysis Method VII-22
xx
-------
Chapter I: Introduction
Chapter I: Introduction
The U.S. Environmental Protection Agency (EPA, the Agency) prepared this Regulatory
Impact Analysis (RIA) for its proposed rule on Tier 2 Motor Vehicle Emissions Standards and
Gasoline Sulfur Control Requirements. The purpose of this RIA is to present EPA's estimates of
the likely costs, benefits, and industry impacts associated with the implementation of both the
Tier 2 vehicle standards and the gasoline sulfur requirements.
The remainder of this chapter summarizes the background information and provisions of the
proposed rulemaking. Subsequent chapters in this RIA present the following information:
• Chapter II presents the health and welfare concerns of motor vehicle emissions including,
ozone (and ozone precursors), particulate matter, and carbon monoxide.
• Chapter III summarizes our analysis of the impact of the Tier 2/Sulfur proposal on
emission inventories, ozone and visibility levels, and air toxics emissions and exposures.
Chapter IV examines the technological feasibility of the proposed Tier 2 exhaust emissions
standards for light-duty vehicles (LDVs) and light-duty trucks (LDTs), as well as the
feasibility of removing sulfur from gasoline.
Chapter V talks about the economic impact of the proposed rule, including the impact of
the proposed Tier 2 standards on vehicle costs, the impact of the gasoline sulfur
requirements on gasoline desulfurization costs, and the combined vehicle and fuel costs per
vehicle and nationwide.
Chapter VI discusses the cost-effectiveness of the proposed vehicle and fuel standards.
The analysis in this chapter focuses on the costs and emission reductions associated with a
single vehicle meeting the Tier 2 emissions standards while operating on low sulfur fuel.
Chapter VII analyzes and estimates the economic impact of the proposed standards by
defining and quantifying the various expected consequences and representing those
consequences in terms of dollars. This analysis provides a means for comparing the
expected benefits of the proposed standards to the expected costs.
Chapter VIII concludes this RIA with a presentation of the Initial Regulatory Flexibility
Analysis for the proposed rule. This analysis evaluates the impacts of the proposed Tier 2
motor vehicle and gasoline sulfur standards on small businesses.
A. Background
On July 31, 1998, EPA submitted its Tier 2 Report to Congress., a formal report which
1-1
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
contained the results of its draft Tier 2 Study.a The purpose of the Tier 2 Study was to examine
the appropriateness of requiring more stringent emission standards for new passenger cars and
light-duty trucks. More specifically, EPA was directed by Congress to assess the air quality
need, technical feasibility, and cost-effectiveness of more stringent motor vehicle emission
standards-emission standards more stringent than federal "Tier 1" standards.
The results of the study indicated that, beginning in 2004, emission reductions will be
necessary to meet and maintain the National Ambient Air Quality Standards (NAAQS) for both
ozone and particulate matter (PM). Air quality modeling showed that during 2007-10, when Tier
2 standards would be fully implemented, several areas in the U.S. would still be in nonattainment
for ozone and PM, even after the implementation of existing emission controls. EPA also found
ample evidence that technologies would be available to meet more stringent Tier 2 standards. In
addition, the Tier 2 Study provided evidence that such standards could be implemented at a cost
(per ton of reduced pollutants) comparable to the costs of other programs designed for similar air
quality problems. Finally, the study identified several additional issues in need of further
examination, including the relative stringency of car and light truck emission standards, the
appropriateness of identical versus different standards for gasoline and diesel vehicles, and the
effects of sulfur in gasoline on vehicle catalyst efficiency.
On May 1, 1998, EPA released its Staff Paper on Gasoline Sulfur Issues which presented its
understanding of the impact of gasoline sulfur on emissions from motor vehicles and explored
what gasoline producers and automobile manufacturers could do to reduce sulfur's impact on
emissions. The staff paper noted that gasoline sulfur is a catalyst poison and that high sulfur
levels in commercial gasoline could affect the ability of future automobiles to meet more
stringent standards in use. It also pointed out that sulfur control would provide additional
benefits by lowering emissions from the current fleet of vehicles.
Based on the statutory requirements described above and the evidence provided in the Tier 2
Study, as updated in the Preamble, EPA is proposing its determination that new, more stringent
emission standards are indeed needed, technologically feasible, and cost effective.
B. Overview of the Proposal
Although the nation's air quality is improving, tens of millions of Americans will continue
to be exposed to unhealthy air pollution levels in the future if new emission controls are not
imposed on motor vehicles. EPA is therefore proposing a major, comprehensive program
designed to significantly reduce emissions from passenger cars and light trucks (including sport-
utility vehicles, minivans, and pickup trucks). Under the proposed program, automakers would
produce vehicles designed to have very low emissions when operated on low-sulfur gasoline,
a On April 28, 1998, EPA published a notice of availability announcing the release of a draft of the Tier 2
study and requesting comments on the draft. The final report to Congress included a summary and analysis of the
comments EPA received.
1-2
-------
Chapter I: Introduction
and oil refiners would provide that cleaner gasoline nationwide. In the proposed rule, EPA
refers to this comprehensive program as the "Tier 2/Gasoline Sulfur Control Program" or simply
as the "Tier 2/Sulfur Program."
1. Vehicle Emission Standards
Today's notice proposes new federal emission standards ("Tier 2 standards") for passenger
cars and light trucks. The program is designed to reduce vehicle emissions of nitrogen oxides
(NOx) and non-methane organic gases (NMOG) (which consist primarily of hydrocarbons (HC)
and volatile organic compounds (VOCs)); NOx and NMOG contribute to the formation of ozone
and particulate matter (PM) which are harmful air pollutants. The program would also, for the
first time, apply the same federal standards to passenger cars and all light trucks ("light LDTs"
and "heavy LDTs").
The proposed Tier 2 standards would reduce new vehicle NOx levels to an average of 0.07
grams per mile (g/mi). For new passenger cars and light LDTs, these standards would phase in
beginning in 2004, with the standards to be fully phased in by 2007.b For heavy LDTs, the
proposed Tier 2 standards would be phased in beginning in 2008, with full compliance in 2009.
During the phase-in period from 2004-2007, all passenger cars and light LDTs not certified to
Tier 2 standards would have to meet an interim average standard of 0.30 g/mi NOx, equivalent to
the current NLEV standards for LDVs.c During the period 2004-2008, heavy LDTs not certified
to Tier 2 standards would phase in an average standard of 0.20 g/mi NOx, with an emissions cap
of 0.60 g/mi NOx.
Manufacturers would be allowed to comply with the very stringent proposed new standards
in a flexible way, assuring that the average emissions of a company's production met the target
emission levels while allowing the manufacturer to choose from several more- and less-stringent
emission categories for certification. The proposed requirements also include more stringent PM
standards, which primarily affect diesel vehicles, and more stringent hydrocarbon controls
(exhaust NMOG and evaporative emissions standards).
2. Gasoline Sulfur Standards
The other major part of today's proposal would significantly reduce average gasoline sulfur
levels nationwide beginning in 2004, and likely earlier due to the proposed incentive program to
By comparison, the NOx standards for the National Low Emission Vehicle (NLEV) program, which will
be in place nationally in 2001, range from 0.30 g/mi for passenger cars to 0.50 g/mi for medium-sized light trucks.
For further comparison, the standards met by today's Tier 1 vehicles range from 0.60 g/mi to 1.53 g/mi.
cThere are also NMOG standards associated with both the interim and Tier 2 standards. The NMOG
standards vary depending on which of various individual sets of emission standards manufacturers choose to use in
complying with the average NOx standard. This "bin" approach is described more fully in section IV.B.
1-3
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
encourage early sulfur reductions. Refiners would generally install advanced refining equipment
to remove sulfur in their refining processes. Importers of gasoline would be required to import
and market only gasoline meeting the proposed sulfur limits. Temporary, less stringent
standards would apply to certain small refiners.
EPA is proposing that gasoline produced by refiners and sold by gasoline importers
generally meet an average sulfur standard of 30 ppm and a cap of 80 ppm. The proposed
program builds upon the existing regulations covering gasoline content as it relates to emissions
performance. It includes provisions for trading of sulfur credits, increasing the flexibility
available to refiners for complying with the new requirements. The proposed credit program is
intended to ease compliance uncertainties by providing refiners the flexibility to phase in early
controls in 2000-03 and use credits generated in these years to delay some control to 2006. As
proposed, the program would achieve expected environmental benefits while providing
substantial flexibility to refiners. The effect of the credit program is that those refiners that
participate would have the opportunity for more overall lead-time to attain the final sulfur levels.
1-4
-------
Chapter II: Health and Welfare Concerns
Chapter II: Health and Welfare Concerns
This chapter describes the public health and welfare concerns associated with the pollutants
impacted by this rulemaking, including ozone, particulate matter, carbon monoxide, air toxics,
and regional haze.
A. Ozone
Ground-level ozone, the main ingredient in smog, is formed by complex chemical reactions
of volatile organic compounds (VOC) and nitrogen oxides (NOx) in the presence of heat and
sunlight. Ozone forms readily in the lower atmosphere, usually during hot summer weather.
VOCs are emitted from a variety of sources, including motor vehicles, chemical plants,
refineries, factories, consumer and commercial products, and other industrial sources. VOCs
also are emitted by natural sources such as vegetation. NOx is emitted largely from motor
vehicles, nonroad equipment, power plants, and other sources of combustion.
The science of ozone formation, transport, and accumulation is complex. Ground-level
ozone is produced and destroyed in a cyclical set of chemical reactions involving NOx, VOC,
heat, and sunlight.d As a result, differences in NOx and VOC emissions and weather patterns
contribute to daily, seasonal, and yearly differences in ozone concentrations and differences from
city to city. Many of the chemical reactions that are part of the ozone-forming cycle are
sensitive to temperature and sunlight. When ambient temperatures and sunlight levels remain
high for several days and the air is relatively stagnant, ozone and its precursors can build up and
produce more ozone than typically would occur on a single high temperature day. Further
complicating matters, ozone also can be transported into an area from pollution sources found
hundreds of miles upwind, resulting in elevated ozone levels even in areas with low VOC or
NOx emissions.
Emissions of NOx and VOC are necessary for the formation of ozone in the lower
atmosphere. For example, small amounts of NOx enable ozone to form rapidly when VOC
levels are high, but ozone production is quickly limited by removal of the NOx. Under these
conditions, NOx reductions are highly effective in reducing ozone while VOC reductions have
little effect. Such conditions are called "NOx limited." Because the contribution of VOC
emissions from biogenic (natural) sources to local ambient ozone concentrations can be
significant, even some areas where man-made VOC emissions are low can be NOx limited.
When NOx levels are high and VOC levels relatively low, NOx forms inorganic nitrates but
little ozone. Such conditions are called "VOC limited." Under these conditions, VOC
d Carbon monoxide also participates in the production of ozone, albeit at a much slower rate than most
VOC and NOx compounds.
II-1
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
reductions are effective in reducing ozone, but NOx reductions can actually increase local ozone.
The highest levels of ozone are produced when both VOC and NOx emissions are present in
significant quantities.
Rural areas are almost always NOx limited, due to the relatively large amounts of biogenic
VOC emissions in such areas. Urban areas can be either VOC or NOx limited, or a mixture of
both, in which ozone levels exhibit moderate sensitivity to changes in either pollutant.
Ozone concentrations in an area also can be lowered by the reaction of nitric oxide with
ozone, forming nitrogen dioxide (NO2); as the air moves downwind and the cycle continues, the
NO2 forms additional ozone. The importance of this reaction depends, in part, on the relative
concentrations of NOx, VOC, and ozone, all of which change with time and location.
1. Health and Welfare Effects of Ozone
Based on a large number of recent studies, EPA has identified several key health effects
caused when people are exposed to levels of ozone found today in many areas of the country.1'2
Short-term exposures (1-3 hours) to high ambient ozone concentrations have been linked to
increased hospital admissions and emergency room visits for respiratory problems. For example,
studies conducted in the northeastern U.S. and Canada show that ozone air pollution is
associated with 10-20 percent of all of the summertime respiratory-related hospital admissions.
Repeated exposure to ozone can make people more susceptible to respiratory infection and lung
inflammation and can aggravate preexisting respiratory diseases, such as asthma. Prolonged
exposure to ozone can cause repeated inflammation of the lung, impairment of lung defense
mechanisms, and irreversible changes in lung structure, which could lead to premature aging of
the lungs and/or chronic respiratory illnesses such as emphysema and chronic bronchitis.
Children are most at risk from ozone exposure because they typically are active outside,
playing and exercising, during the summer when ozone levels are highest. For example, summer
camp studies in the eastern U.S. and southeastern Canada have reported significant reductions in
lung function in children who are active outdoors. Further, children are more at risk than adults
to ozone exposure because their respiratory systems are still developing. Adults who are
outdoors and moderately active during the summer months, such as construction workers and
other outdoor workers, also are among those most at risk. These individuals, as well as people
with respiratory illnesses such as asthma, especially asthmatic children, can experience reduced
lung function and increased respiratory symptoms, such as chest pain and cough, when exposed
to relatively low ozone levels during periods of moderate exertion.
Several recent studies have shown a possible relationship between exposure to ambient
ozone and premature mortality. This literature has been evolving rapidly. Of the 28 time-series
epidemiology studies identified in the literature that report results on a possible association
between daily ozone concentrations and daily mortality3, 21 were published or presented since
1995. In particular, a series of studies published in 1995 through 1997 (after closure on the
current ozone NAAQS Criteria Document) from multiple cities in western Europe has increased
II-2
-------
Chapter II: Health and Welfare Concerns
significantly the body of studies finding a positive association. Fifteen of the 28 studies report a
statistically significant relationship between ozone and mortality; the more recent studies tended
to find statistical significance more often than the earlier studies. The ozone-mortality datasets
also have tended to become larger in more recent studies, as longer series of air quality
monitoring data have become available over time. This suggests that it may take many years of
data before the ozone effect can be separated from the daily weather and seasonal patterns with
which it tends to be correlated.
In 1997, as a part of the Regulatory Impact Analysis (RIA) for the ozone NAAQS
promulgation, EPA staff reviewed this recent literature. They identified nine studies that met a
defined set of selection criteria, and conducted a meta-analysis of the results of the nine studies.
(U.S. EPA, 1997). See Chapter VII.C.3.a. for a further discussion of this meta-analysis.
In addition to human health effects, ozone adversely affects crop yield, vegetation and forest
growth, and the durability of materials. Because ground-level ozone interferes with the ability of
a plant to produce and store food, plants become more susceptible to disease, insect attack, harsh
weather and other environmental stresses. Ozone causes noticeable foliar damage in many
crops, trees, and ornamental plants (i.e., grass, flowers, shrubs, and trees) and causes reduced
growth in plants. Studies indicate that current ambient levels of ozone are responsible for
damage to forests and ecosystems (including habitat for native animal species). Ozone
chemically attacks elastomers (natural rubber and certain synthetic polymers), textile fibers and
dyes, and, to a lesser extent, paints. For example, elastomers become brittle and crack, and dyes
fade after exposure to ozone.
VOC emissions are detrimental not only for their role in forming ozone, but also for their
role as air toxics. Some VOCs emitted from motor vehicles are toxic compounds. At elevated
concentrations and exposures, human health effects from air toxics can range from respiratory
effects to cancer. Other health impacts include neurological, developmental and reproductive
effects. Chapter III contains more information about air toxics.
Besides their role as an ozone precursor, NOx emissions produce a wide variety of health
and welfare effects.45 These problems are caused in part by emissions of nitrogen from motor
vehicles. Nitrogen dioxide can irritate the lungs and lower resistance to respiratory infection
(such as influenza). NOx emissions are an important precursor to acid rain and may affect both
terrestrial and aquatic ecosystems. Atmospheric deposition of nitrogen leads to excess nutrient
enrichment problems ("eutrophication") in the Chesapeake Bay and several nationally important
estuaries along the East and Gulf Coasts. Eutrophication can produce multiple adverse effects
on water quality and the aquatic environment, including increased algal blooms, excessive
phytoplankton growth, and low or no dissolved oxygen in bottom waters. Eutrophication also
reduces sunlight, causing losses in submerged aquatic vegetation critical for healthy estuarine
ecosystems. Deposition of nitrogen-containing compounds also affects terrestrial ecosystems.
Nitrogen fertilization can alter growth patterns and change the balance of species in an
ecosystem. In extreme cases, this process can result in nitrogen saturation when additions of
nitrogen to soil over time exceed the capacity of plants and microorganisms to utilize and retain
the nitrogen.
II-3
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Elevated levels of nitrates in drinking water pose significant health risks, especially to
infants. Studies have shown that a substantial rise in nitrogen levels in surface waters are highly
correlated with human-generated inputs of nitrogen in those watersheds.6 These nitrogen inputs
are dominated by fertilizers and atmospheric deposition.
Nitrogen dioxide and airborne nitrate also contribute to pollutant haze, which impairs
visibility and can reduce residential property values and tourism revenues. Section II.D. further
describes information about visibility impairment and regional haze.
B. Particulate Matter
Particulate matter (PM) is the general term for the mixture of solid particles and liquid
droplets found in the air. Particulate matter includes dust, dirt, soot, smoke, and liquid droplets
that are directly emitted into the air from natural sources (such as windblown dust and fires) and
manmade sources (such as motor vehicles, construction sites, factories, and driving on unpaved
roads). Secondary PM is formed in the atmosphere through a number of physical and chemical
processes that transform gases such as sulfur dioxide, NOx, and VOC into particles.
Particulate matter is distinguished between larger or "coarse" particles (larger than 2.5
micrometers) and smaller or "fine" particles (smaller than 2.5 micrometers). The characteristics,
sources, and potential health effects of coarse and fine particles are very different. Coarse
particles primarily come from natural sources, such as windblown dust and sea salt. They
remain in the atmosphere a relatively short period of time. Fine particles primarily consist of
secondary particles formed by gaseous emissions and often come from human sources, such as
industrial and residential combustion, vehicle exhaust, and agriculture (including silvicultural
prescribed burning). Fine PM consists primarily of sulfate-based particles (produced from sulfur
oxides (SOx)), nitrate-based particles (produced from NOx), and carbon-based particles emitted
directly from combustion processes and created through transformation of VOC emissions.
Particles directly emitted from motor vehicles, and those formed by the transformation of motor
vehicle gaseous emissions, tend to be in the fine particle range.
The formation and fate of secondary PM involves complex processes which are sensitive to
sunlight, temperature, humidity, and other reactants. SOx, NOx, and VOC emissions are
photochemically oxidized and react with water vapor to form sulfuric and nitric acids.6 Sulfuric
acid reacts with ammonia to form mostly ammonium sulfate and some ammonium bisulfate,
while nitric acid reacts with ammonia to form ammonium nitrate. Ammonia gas is emitted from
biogenic sources and biomass burning, both natural and anthropogenic. If ammonia is in limited
supply, it will react to form sulfate rather than nitrate since sulfuric acid has a higher chemical
affinity for ammonia than does nitric acid. Furthermore, ammonium nitrate reacts with ammonia
and nitric acid in an equilibrium reaction, so nitric acid removal processes such as dry deposition
will also lower the concentration of nitrate PM.
eSulfuric acid is a paniculate, while nitric acid is a gas at ambient conditions.
II-4
-------
Chapter II: Health and Welfare Concerns
As a result, a much higher fraction of SOx is converted to PM than is the case for NOx.
Conversion rates vary depending on local meteorology and the amount of ammonia, NOx, and
SOx in the local atmosphere. However, mobile sources reasonably can be estimated to
contribute to ambient secondary sulfate and nitrate in proportion to their contribution to total
NOx and SOx emissions in a given area geographically.
Mobile sources are significant producers of carbonaceous PM, which consists largely of
elemental carbon directly emitted by diesel vehicles and poorly maintained gasoline vehicles.
Secondary carbonaceous PM results when VOCs or their photochemical reaction products
adsorb to existing particles.
In the eastern U.S., based on limited monitoring data, sulfate is the largest single component
of fine PM, closely followed by carbonaceous PM. Nitrate is the third-largest component of fine
PM, accounting for roughly 10 percent of the total. Most of the rest is soil dust. In the West,
again based on limited monitoring data, carbonaceous PM is generally the largest fraction of fine
PM. Sulfate forms a smaller fraction of fine PM than in the East, probably because SOx
emissions are lower. Sulfate still comprises a larger fraction of the total than nitrate, however,
except in parts of California. Soil dust is a more important component of fine PM in the West
than in the East, but is still smaller than nitrate in most places. Throughout the U.S., rural areas
have lower fine PM levels than urban areas.
1. Health and Welfare Effects of Particulate Matter
Scientific studies have linked particulate matter (alone or in combination with other air
pollutants) with a series of health effects.7 Coarse particles can accumulate in the respiratory
system and aggravate health problems such as asthma. Fine particles penetrate deeply into the
lungs and are more likely than coarse particles to contribute to a number of the health effects.
These health effects include premature death and increased hospital admissions and emergency
room visits, increased respiratory symptoms and disease, decreased lung function, and alterations
in lung tissue and structure and in respiratory tract defense mechanisms. Children, the elderly,
and people with cardiopulmonary disease, such as asthma, are most at risk from these health
effects. Chapter III contains a discussion of the toxic health effects from particulate matter in
diesel fuel exhaust.
Particulate matter also causes a number of adverse effects on the environment. Fine
particulate matter is the major cause of reduced visibility in parts of the U.S., including many of
our national parks and wilderness areas. (Section II.D. further describes visibility impairment
and regional haze). Other environmental impacts occur when particles deposit onto soil, plants,
water, or materials. For example, particles containing nitrogen and sulfur that deposit onto land
or water bodies may change the nutrient balance and acidity of those environments, leading to
changes in species composition and buffering capacity.
Particles that are deposited directly onto leaves of plants can, depending on their chemical
composition, corrode leaf surfaces or interfere with plant metabolism. When deposited in
II-5
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
sufficient quantities, such as near unpaved roads, tilled fields, or quarries, particles block
sunlight from reaching the leaves, stressing or killing the plant. Finally, particulate matter
causes soiling and erosion damage to materials, including culturally important objects, such as
carved monuments and statues.
C. Carbon Monoxide
Carbon monoxide (CO) is a colorless, odorless gas produced though the incomplete
combustion of carbon-based fuels. Carbon monoxide enters the bloodstream through the lungs
and reduces the delivery of oxygen to the body's organs and tissues. The health threat from CO
is most serious for those who suffer from cardiovascular disease, particularly those with angina
or peripheral vascular disease. Healthy individuals also are affected, but only at higher CO
levels. Exposure to elevated CO levels is associated with impairment of visual perception, work
capacity, manual dexterity, learning ability and performance of complex tasks.
Several recent epidemiological studies have shown a link between CO and premature
mortality and morbidity (including angina, congestive heart failure, and other cardiovascular
diseases). EPA currently is in the process of reviewing these studies as part of the CO Criteria
Document process.
Since 1979, the number of areas in the nation violating the CO NAAQS has decreased by a
factor of almost ten, from 48 areas in 1979 to five areas (covering seven counties) in 1995 and
1996. In 1997, three counties, with a total population of nine million people, failed to meet the
CO standard.
In addition to the substantial decrease in the number of areas where the NAAQS is
exceeded, the severity of the exceedances also has decreased significantly. Nationally, CO
concentrations decreased 38 percent during the past 10 years/ From 1979 to 1996, the measured
atmospheric concentrations of CO during an exceedance decreased from 20-25 ppm at the
beginning of the period to 10-12 ppm at the end of the period. Expressed as a multiple of the
standard, atmospheric concentration of CO during an exceedance was two to almost three times
the standard in 1979. By 1996, the CO levels present during an exceedance decreased to 10-30
percent over the nine ppm standard.
Unlike the case with ozone and PM, EPA has not made any recent comprehensive
projections of future ambient CO levels and attainment and maintenance of the CO NAAQS.
However, section 202(j) of the CAA requires a separate study of the need for more stringent cold
CO standards. EPA is currently conducting this study.
fThis value of the CO concentration decrease is measured by the composite average of the annual second
highest 8-hour concentration.
II-6
-------
Chapter II: Health and Welfare Concerns
D. Visibility and Regional Haze
Visibility impairment is the haze that obscures what we see, and is caused by the presence
of tiny particles in the air. These particles cause light to be scattered or absorbed, thereby
reducing visibility. Visibility impairment, also called regional haze, is a complex problem that
relates to several pollutants. Visibility in our national parks and monuments, and many urban
areas of the country, continues to be obscured by regional and local haze.
The principle cause of visibility impairment is fine particles, primarily sulfates, but also
nitrates, organics, and elemental carbon and crustal matter. Particles between 0.1 and one
micrometers in size are most effective at scattering light, in addition to being of greatest concern
for human health. Of the pollutant gases, only NO2 absorbs significant amounts of light; it is
partly responsible for the brownish cast of polluted skies. However, it is responsible for less
than ten percent of visibility reduction.
In the eastern U.S., reduced visibility is mainly attributable to secondary particles,
particularly those less than a few micrometers in diameter. Based on data collected by the
Interagency Monitoring of Protected Visual Environments (IMPROVE) network for visibility
monitoring, sulfate particles account for about 50-70 percent of annual average light extinction
in eastern locations. Sulfate plays a particularly significant role in the humid summer months,
most notably in the Appalachian, northeast, and mid-south regions. Nitrates, organic carbon, and
elemental carbon each account for between 10-15 percent of total light extinction in most
eastern locations. Rural areas in the eastern U.S. generally have higher levels of impairment
than most remote sites in the western U.S., generally due to the eastern U.S.'s higher levels of
man-made pollution, higher estimated background levels of fine particles, and higher average
relative humidity levels.
The relative contribution of individual pollutants to visibility impairment vary
geographically. While secondary particles still dominate in the West, direct particulate emissions
from sources such as woodsmoke contribute a larger percentage of the total particulate load than
in the East. In the rural western U.S., sulfates also play a significant role, accounting for about
25-40 percent of total light extinction in most regions. In some areas, such as the Cascades
region of Oregon, sulfates account for over 50 percent of annual average light extinction.
Organic carbon typically is responsible for 15-35 percent of total light extinction in the rural
western U.S. and elemental carbon (absorption) accounts for about 15-25 percent, so the total
carbonaceous contribution is between 30 and 60 percent. Soil dust (coarse PM) accounts for
about 10-20 percent. Nitrates typically account for less than 10 percent of visibility
impairment.8
The CAA requires EPA to protect visibility, or visual air quality, through a number of
programs. These programs include the national visibility program under sections 169a and 169b
of the Act, the Prevention of Significant Deterioration program for the review of potential
impacts from new and modified sources, and the secondary NAAQS for PM10 and PM2 5. The
national visibility program established in 1980 requires the protection of visibility in 156
mandatory Federal Class I areas across the country (primarily national parks and wilderness
II-7
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
areas). The CAA established as a national visibility goal, "the prevention of any future, and the
remedying of any existing, impairment of visibility in mandatory Federal class I areas in which
impairment results from manmade air pollution." The Act also calls for state programs to make
"reasonable progress" toward the national goal. In July 1997, EPA proposed a program to
address regional haze in the nation's most treasured national parks and wilderness areas (see 62
FR 41137, July 31, 1997).
Since mobile sources contribute to visibility-reducing PM, control programs that reduce the
mobile source emissions of direct and indirect PM will have the effect of improving visibility.
Western Governors, in commenting on the Regional Haze Rule and on protecting the 16 Class I
areas on the Colorado Plateau, stated that, "...the federal government must do its part in
regulating emissions from mobile sources that contribute to regional haze in these areas..." and
called on EPA to make a "binding commitment to fully consider the Commission's
recommendations related to the ... federal national mobile source emissions control strategies",
including Tier 2 vehicle emissions standards.9 The Grand Canyon Visibility Transport
Commission's report found that reducing total mobile source emissions is an essential part of
any program to protect visibility in the Western U.S.10 The Commission identifies mobile source
pollutants of concern as VOC, NOX, and elemental and organic carbon.
II-8
-------
Chapter II: Health and Welfare Concerns
Chapter II. References
1. U.S. EPA, 1996, Review of National Ambient Air Quality Standards for Ozone,
Assessment of Scientific and Technical Information, OAQPS Staff Paper, EPA-452/R-
96-007.
2. U.S. EPA, 1996, Air Quality Criteria for Ozone and Related Photochemical Oxidants,
EPA/600/P-93/004aF.
3. U. S. EPA, 1997, Regulatory Impact Analysis for the Ozone NAAQS, Appendix J,
"Assessment and Synthesis of Available Epidemiological Evidence of Mortality
Associated with Ambient Ozone from Daily Time-series Analyses".
4. U.S. EPA, 1995, Review of National Ambient Air Quality Standards for Nitrogen
Dioxide, Assessment of Scientific and Technical Information, OAQPS Staff
Paper,
EPA-452/R-95-005
5. U.S.EPA, 1993, Air Quality Criteria for Oxides of Nitrogen, EPA/600/8-91/049aF.
6. Vitousek, Pert M., John Aber, Robert W. Howarth, Gene E. Likens, et al. 1997. Human
Alteration of the Global Nitrogen Cycle: Causes and Consequences. Issues in Ecology.
Published by Ecological Society of America, Number 1, Spring 1997.
7. U.S. EPA, 1996, Air Quality Criteria for Particulate Matter, EPA/600/P-95/001aF.
8. "National Air Quality and Emissions Trends Report, 1996", EPA Document Number
454/R-97-013.
9. Letter from Governor Michael Leavitt of Utah, on behalf of the Western Governors'
Association, to EPA Administrator Carol Browner, dated June 29, 1998.
10. "Report of the Grand Canyon Visibility Transport Commission to the United States
Environmental Protection Agency", June 1996.
II-9
-------
Chapter III: Environmental Impact
Chapter III: Environmental Impact
A. Inventory Impacts of Tier 2/Sulfur
Today's proposal, if adopted, would reduce NOx, VOC, particulate, SOx, carbon
monoxide, and hazardous air pollutant emissions from cars and light trucks by lowering the
VOC, NOx, and PM emission standards for these vehicles and requiring that gasoline sulfur
levels be reduced. Over time, the projected benefits of today's proposal would grow as vehicles
meeting the new standards replace older, higher-emitting vehicles and as total VMT continues to
grow. The results of our analysis of light-duty inventory levels with and without today's action
are presented and discussed for each pollutant in the following sections. In all cases, "without
Tier 2/Sulfur" refers to continuation of National LEV on in-use fuel as currently specified; sulfur
levels for Conventional Gasoline are estimated at 330 ppm, summertime Phase 2 RFG levels are
estimated at 150 ppm (i.e., baseline case). "With Tier 2/Sulfur" refers to implementation of a 30
ppm sulfur standard nationwide in 2004 and the phase-in of NOx, VOC, and PM standards
proposed under today's action (i.e., control case).8
For this proposal, EPA developed new inventory projections for the United States excluding
California, Alaska, and Hawaii.11 These inventory projections can be divided into three major
types of sources for the purpose of describing the methodologies used: stationary and area
sources, nonroad mobile sources, and highway motor vehicles. To assess air quality need and
the impact of today's proposal on urban areas, separate inventory analyses were also performed
on four high ozone cities: New York, Chicago, Atlanta and Charlotte. Inventory estimates for
each city were developed using the same data sources as the 47-state inventory discussed below,
except where noted. Comprehensive inventories (47-state and four city) are presented in
Appendix A with and without Tier 2/sulfur control, in 2005 (47-state only), 2007, 2010, 2015,
2020 and 2030.
These 47-state inventory projections are described more fully in this section. These
projections differ in some respects from the inventory projections used for the ozone analyses
described in Section B.I. and the inventory projections used for the benefit/cost analyses
gToday's proposal includes a provision for the averaging, banking and trading (ABT) of sulfur levels
which would allow average sulfur levels to be higher than 30 ppm in 2004/2005 in exchange for sulfur control
prior to 2004 (See Section IV.C.S.c.i of the preamble for a detailed discussion of this program). We expect that
overall emission reductions from the ABT program between 2001 and 2005 would be consistent with
implementation of 30 ppm in 2004 without prior sulfur reduction, and hence assumed the latter schedule for the
control case results presented here.
hThe 47-state region comprised of the U.S. minus California, Alaska and Hawaii is interchangeably
referred to as "nationwide" throughout this section. Although excluded from this analysis, emission reductions will
be realized in each of these states. Today's action applies fully to Alaska, Hawaii, and U.S. territories; California,
although subject to a separate vehicle and fuel control program, will benefit from lower-emitting Federal vehicles
migrating to and/or traveling within the state, as well as California vehicles operating on cleaner non-California
fuel.
III-l
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
described in Chapter VII. The methods used to develop the inventory projections presented in
this section are described below. Subsequently, the differences between those methods and the
methods used to develop the inventories used for the ozone analyses and benefit/cost analyses
are described.
The 47-state inventory projections include updated emission estimates for the stationary,
area, nonroad, and highway mobile source sectors.1 For stationary and area sources, we relied on
a set of 47-state projection inventories developed for this analysis by E.H. Pechan and
Associates. Pechan used state inventories developed for the Regional Ozone Transport Rule
(ROTR, 63 FR 57356, October 27, 1998) and Trends inventories for the non-OTAG states to
develop a base year stationary and area source inventory. For sources not covered by emissions
caps, the total emissions were grown using BEA-based growth factors. Emissions for large
electric utilities were held constant at projected 2005 levels, consistent with emissions cap
requirements in OTAG states and projections of shifts in types of fuel used in other states.
For nonroad mobile sources (except locomotives, aircraft, and commercial marine), we
developed 47-state emission inventories using EPA's Draft NONROAD emissions model. This
model is a significant update in data sources and methodologies compared to the NEVES
inventories which have been the basis for nonroad emission estimates since 1992. Although
NONROAD has only been released in draft form, the emissions estimation data and methods
incorporated in it represent our most recent analysis of nonroad emission levels. For this reason,
we chose to use the draft NONROAD model to develop our nonroad mobile source emission
estimates used to evaluate the impact of the Tier 2/Sulfur proposal on emission inventories. The
methods and data used in NONROAD are also consistent with the methods and data used in
recent EPA proposed and final rules on nonroad engine standards and the nonroad emissions
projections used here reflect all final and proposed standards for nonroad engines and equipment.
Growth estimates in NONROAD are based on a linear projection of historical populations of
nonroad equipment.
Because NONROAD does not yet include locomotives, aircraft, or commercial marine
vessels, we had to use different sources to project emission inventories for these sources.
Estimates of projected locomotive emissions were based on estimates in EPA's Final Rule on
locomotive standards, adjusted to reflect the 47-state basis of the inventory described above.
Commercial marine emissions were based on estimates in EPA's proposed commercial marine
rule. Aircraft emissions estimates were based on Trends estimates adjusted to reflect a 47-state
basis and grown using FAA growth estimates.
The most critical piece of our 47-state inventory analysis is the light-duty on-highway
vehicle inventory. We are in the process of updating the on-highway mobile source emission
factor models MOBILE (NOx, VOC and CO) and PART (PM and SOx), and the latest versions
of these models (MOBILE6 and PART6) are not yet available. However, many of the modified
inputs and assumptions which will be used in these models have at least been developed in draft
form; thus, we were able to develop an up-to-date assessment of light-duty vehicle and truck
emission inventory for today's proposal using a model which incorporated available elements
which have or will be proposed as part of the MOBILE6 and PART6 models. This model,
III-2
-------
Chapter III: Environmental Impact
referred to as the Tier 2 Model1, reflects updated assessments of in-use emission deterioration
and off-cycle emissions, fuel sulfur impacts, and the increase in truck sales relative to cars. The
model also reflects existing national and local motor vehicle control programs including National
LEV (NLEV), Supplemental Federal Test Procedure (SFTP), On-Board Diagnostics (OBD),
reformulated gasoline (RFG) and Inspection/Maintenance (I/M) programs. We used this model
to develop baseline emission estimates assuming that NLEV program continued in perpetuity
(.e., that there would be no Tier 2/Sulfur standards implemented) and to develop emission
estimates for various control scenarios.2
The 47-state nonexhaust VOC emission inventory was developed using MOBILESb, since
MOBILE6 estimates of evaporative emissions were not available at the time of the analysis.
However, we incorporated available elements of MOBILE6 where possible, including mileage
accumulation, VMT mix, and age distribution.3 A modified version of MOBILESb was also
developed to estimate the benefits of today's proposed evaporative standards.4
47-state inventory estimates for heavy-duty gasoline, heavy-duty diesel and motorcycles
also incorporated available aspects of MOBILE6 and PART6, including new base emission
rates, defeat device emissions for heavy-duty diesel vehicles, off-cycle emissions, mileage
accumulation and age distribution. New standards recently finalized for heavy-duty diesel
vehicles were accounted for in these inventories, as were standards expected to be proposed for
heavy-duty gasoline vehicles.5'6
To generate inventory projections, we needed to combine the on-highway emission factors
generated using the Tier 2 Model or other means described above with estimates of on-highway
vehicle miles traveled (VMT). For our 47-state inventory analysis, VMT estimates through 2010
were based on EPA's Trends Report through 2010. Beyond 2010, we developed VMT estimates
for light-duty cars and trucks based on current trends in VMT growth as reported by NHTSA.
From 2010 through 2015, we project that 47-state light-duty VMT will grow at a rate of 2.1
percent per year compounded; beyond 2015, we estimate VMT growth will be reduced to a
linear 2.1 percent per year (i.e, 2.1 percent of 2015 VMT added incrementally in successive
years). Projected 47-state VMT levels for heavy-duty gasoline and diesel vehicles were
developed based on data from EPA's Trends Report.
Consistent with EPA's Trends Report, the 47-state inventory estimates were developed on
the basis of annual tons emitted. Annual VMT estimates were used in conjunction with emission
factors which reflected seasonal fuel control (i.e., low sulfur RFG in the summer only). Because
of limitations in the Tier 2 Model, however, seasonal temperature adjustments were not made.
1 The Tier 2 Model is the next generation of Modified MOBILESb (T2AT), the inventory model used in the
Tier 2 Study. Since the study, the model has been transferred to a Microsoft Excel platform, updated extensively
and expanded to include SOx and PM emissions. The development of this model and generation of light-duty
inventory results presented in this section are outlined in the technical report "Development of Light-Duty
Emissions Inventory Estimates in the Notice of Proposed Rulemaking for Tier 2 and Sulfur Standards" contained in
Docket No. A-97-10. The Tier 2 Model is being made available in concurrence with the publication of today's
proposal.
m-3
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Thus, the annual tonnages developed for the 47-state inventories are annual tonnages assuming
summertime temperatures year-round. The effect of using summertime conditions to estimate
year-round emissions is to understate annual NOx emissions by about seven percent. Estimates
of NOx emission reductions from changes in standards will be similarly underestimated. EPA
does not consider this small error to be material.
The urban (four-city) analysis was performed on the basis of summertime (May through
September) emissions. Stationary and area source inventories were provided by E.H. Pechan
and Associates and were based on state inventories developed for the Regional Ozone Transport
Rule (ROTR). Emissions from nonroad equipment were estimated using the NONROAD model,
which includes the capability to allocate emissions to the county level. For nonroad equipment
not included in the NONROAD model (locomotives, aircraft, and commercial marine) we did
not have enough information to directly allocate the 47 inventories described above down to the
county level for these urban areas. However, E.H. Pechan has calculated national and local
inventories for these categories and areas using older assumptions about future emissions
standards. We used those older inventories to calculate the proportion of national emissions
from locomotives, aircraft, and commercial marine engines in the four urban areas. We then
applied those proportions to the our newer national inventories for the three categories to
estimate emissions for locomotives, aircraft, and commercial marine engines in the four urban
areas using our latest assumptions about the effects of new standards.
Summertime VMT estimates for each area used in generation of OTAG inventories were
provided by Pechan for 1995 and 2007; in order to more closely match localized VMT growth
trends, the values were linearly interpolated between these years, and extrapolated linearly
beyond 2007. Emission factors for highway vehicles were derived using the same methods
described above for the 47-state inventories, but with local specific inputs, such as I/M programs
or reformulated gasoline, where applicable.
The emission inventories used for the ozone analyses described in Section B.I. of this
chapter and the benefit/cost analyses described in Chapter VII were developed prior to the 47-
state inventory described in this section. The ozone analysis and benefit/cost analysis
inventories differ from one another and from the 47-state inventories in several respects. It
should be noted, however, that we used the emission inventory analyses described in this section
to determine the change in emissions from the proposed Tier 2/Sulfur standards for both of these
analyses.
The inventories used for the ozone modeling are described more fully in Section B.I. To
develop the car and light truck baseline inventories (the inventories that would result if the Tier
2/Sulfur proposal were not adopted), we used the car and light truck inventories developed for
the ROTR; these inventories were based on MOBILES inputs and emission factors. To estimate
the change in emissions from cars and light trucks that would occur if the proposed Tier 2/Sulfur
standards were implemented, we used the same methods used to develop the 47-state emission
inventories (as described in this section). The inventories for highway heavy-duty engine
emissions and nonroad emissions were based on the emission modeling tools that were available
to the Ozone Transport Assessment Group (OTAG) and were used during that process and the
III-4
-------
Chapter III: Environmental Impact
subsequent rulemaking process that resulted in promulgation of the ROTR. The highway heavy-
duty engine emissions were based on MOBILES inputs and emission factors; the nonroad
emissions were based on NEVES inputs.
The benefit/cost analyses described in Chapter VII used an even earlier set of estimates for
highway, nonroad, and stationary and area source emissions that was developed before the
ROTR was proposed or promulgated. These estimates were developed using the emission
modeling tools available at the time the Regulatory Impact Analyses for the revised ozone and
PM NAAQS rules were developed.7 The inventories used in the benefit/cost analyses are
described more fully in Chapter VII.
1. NOx
a. Light-Duty NOx Trends Without Tier 2/Sulfur
Total NOx emissions produced annually in the 47 states by cars and trucks without Tier
2/Sulfur controls are shown in Table III-l and Figure III-l, broken down by relative contribution
of cars (light-duty vehicles, or LDVs), LDTls and 2s (light pickup trucks, minivans and most
sport utility vehicles), and LDT3s and 4s (heavier pickup trucks and sport utility vehicles). As
shown, total light-duty emissions decline from approximately 3.9 million tons to 3.0 million tons
between 2000 and 2010 due to turnover of Tier 1 and NLEV vehicles and the phase in of off-
cycle standards (SFTP). By 2014, however, the effect of these control programs begins to be
offset by increases in overall VMT, in conjunction with the shift of VMT from cars to trucks.
Light-duty emissions increase to 3.2 million tons by 2020 and 3.8 million tons by 2030, such that
the gains from the Tier 1, NLEV and SFTP control programs are almost completely eradicated
by VMT growth.
m-5
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table III-l. 47-State Light Duty NOx Emissions Without Tier 2/Sulfur (Annual Tons)
Year
2000
2004
2007
2010
2015
2020
2030
Light-Duty
Emissions
3,869,383
3,539,655
3,218,530
3,041,639
3,020,806
3,221,151
3,790,840
Contribution by Vehicle Class
LDV
48.0%
41.4%
36.3%
32.0%
27.7%
25.9%
25.4%
LDT1/2
35.6%
39.9%
42.2%
43.6%
44.3%
44.8%
45.1%
LDT3/4
16.4%
18.7%
21.5%
24.4%
28.0%
29.3%
29.5%
4,000,000
3,000,000
2,000,000
1,000,000
2005
2010
2015
2020
ILDV
ILDT1/2 D LDT3/4
2025
2030
Figure III-l. 47-State Light-Duty NOx Emissions Without Tier 2/Sulfur (Annual Tons)
The impact of steady truck growth on overall light-duty NOx emissions is clearly
m-6
-------
Chapter III: Environmental Impact
demonstrated in the preceding figure. In 2000, we project that trucks will produce 50 percent of
overall NOx emissions. Over the next 30 years, trucks will grow to dominate light-duty NOx
emissions due to the combined effects of sales migration, higher mileage accumulation rates,
longer lifespan, and more relaxed emission standards relative to LDVs. By 2010, we project
trucks will make up two-thirds of light-duty NOx emissions; by 2020, nearly three-quarters of all
light-duty NOx emissions will be produced by trucks. As shown in Figure III-l, the decrease in
overall light-duty emission levels is due solely to reductions in LDV emissions. The benefits
from Tier 1, NLEV and SFTP are not as pronounced for trucks, and are offset almost
immediately by growth in truck VMT. As a result, truck emissions are stable through 2010 and
begin increasing steadily beyond this as VMT growth overtakes the gains from existing control
programs.
The emission trends for the four urban areas we analyzed show similar behavior.
Although the presence of localized control programs (I/M and in some cases, RFG) do serve to
delay the upturn in light-duty emissions, they are not sufficient to counteract the effects of VMT
growth. As shown in Table III-2, light-duty emissions decrease steadily in each city through
2010. Emission trends beyond 2010 depend on the rate of VMT growth in each city. In New
York, which is projected to have relatively low VMT growth, emissions continue to decrease
steadily through 2015 before leveling off and then turning upwards by 2017. In Chicago,
Atlanta, and Charlotte, emissions begin to level off by 2010. Emissions start to increase in 2017
in Chicago, 2015 in Atlanta, and 2013 in Charlotte. For the latter two cities, emissions increase
at a rapid rate beyond these years. We project that Atlanta's emission reductions achieved from
programs currently in place will be almost fully offset by rapid VMT growth by 2030, while we
project Charlotte's rapid VMT growth to cause emissions in 2030 to be over 10 percent higher
than in 2000.
Table III-2. Four-City Light-Duty NOx Emissions Without Tier 2/Sulfur (Summer Tons)
Year
2000
2004
2007
2010
2015
2020
2030
New York
78,287
66,857
57,753
51,811
47,634
48,033
52,280
Chicago
37,037
32,314
28,399
25,958
24,440
25,080
28,165
Atlanta
33,267
30,912
28,313
26,846
26,384
27,721
32,018
Charlotte
4,714
4,526
4,230
4,081
4,109
4,402
5,239
Figures III-2 and III-3 show our projections of the contribution of light-duty vehicles and
trucks to the total NOx inventory (i.e., NOx emission from all sources, including stationary, area,
m-7
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
nonroad) in the 47 states and in Atlanta, in 2020. Table III-3 shows this same contribution
across the 47 states and all four cities from 2007 through 2030. Across the 47 states, cars and
trucks produce nearly one-fifth of total NOx emissions across all years. In urban areas, however,
this contribution can be significantly higher. Atlanta provides the most striking example of this;
we project that roughly 40 percent of all NOx emissions will be produced by cars and trucks
through 2030. While less than Atlanta, the light-duty contribution in New York is significantly
higher than the national estimates; from 2007 through 2030, we project that 27 to 29 percent of
all emissions in this area will be produced by light-duty cars and trucks. We estimate the
contribution in Chicago and Charlotte to be slightly higher but comparable to the 47-state
estimate of one-fifth of the total NOx inventory.
Light-duty NOx contribution in urban areas is generally higher than the 47-state region
because of the increased concentration of VMT, in conjunction with the decreased prevalence of
significant NOx contributors which are largely in non-urban areas (primarily utilities and
agricultural nonroad sources). We expect that this trend will be consistent across many high-
ozone urban areas.
III-8
-------
Chapter III: Environmental Impact
Light-Duty
Vehicles and
Trucks
18%
Stationary and
Area
59%
Other On-
Highway
7%
N on road
16%
Figure III-2: Breakdown of Total 2020 47 State NOx Inventory Without Tier 2
Stationary and
Area
20%
N on road
26%
Light-Duty
Vehicles and
Trucks
40%
Other On-
H ighway
14%
Figure III-3. Breakdown of Total 2020 Atlanta NOx Inventory Without Tier 2
111-9
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table III-3. Light-Duty Contribution to Total NOx Inventory Without Tier 2/Sulfur
Year
2007
2010
2015
2020
2030
47 -state
17%
17%
17%
18%
20%
New York
29%
28%
27%
27%
28%
Chicago
19%
18%
18%
19%
19%
Atlanta
38%
38%
39%
40%
42%
Charlotte
18%
19%
19%
21%
22%
b.
NOx Reductions Due To Tier 2/Sulfur
Today's proposal would provide substantial reductions in NOx emissions from cars and
trucks. The implementation of low sulfur fuel in 2004 would afford an immediate drop in NOx
emissions, while the phase-in of tighter vehicle standards would continue to reduce emissions
over time, serving to mitigate through at least 2030 the projected upward trend in light-duty
NOx emissions that would occur with no control. Table III-4 contains annual tons of NOx we
project would be reduced by today's proposal, encompassing benefits of low sulfur fuel and the
introduction of Tier 2 light-duty vehicle and light-duty truck standards. Figure III-4 shows
annual 47-state light-duty NOx emissions with implementation of the Tier 2/Sulfur program,
broken down by LDV, LDT1/2 and LDT3/4 categories.
Table III-4. 47-State Light-Duty NOx Reductions Due To Tier 2/Sulfur (Annual Tons)
Year
2004
2007
2010
2015
2020
2030
Light-Duty
Emissions Without
Tier 2/Sulfur
3,539,655
3,218,530
3,041,639
3,020,806
3,221,151
3,790,840
Light-Duty
Emissions With
Tier 2/Sulfur
3,037,144
2,422,796
1,859,316
1,241,925
1,023,038
1,004,495
Emissions
Reduced
502,511
795,734
1,182,323
1,778,881
2,198,113
2,786,345
Percent Reduction in
Baseline Inventory
Light-Duty
14%
25%
39%
59%
68%
74%
All Sources*
-
4%
7%
10%
12%
15%
: Includes emission reductions from Heavy-Duty Gasoline Vehicles due to sulfur control
III-10
-------
Chapter III: Environmental Impact
4,000,000 -T
3,000,000
2,000,000
1,000,000
WITHOUT TIER 2/SULFUR CONTROLS
2005
2010
2015
2020
2025
2030
LDV
I LOT 1/2
D LDT3/4
Figure III-4. 47-State Light-Duty NOx Emissions With Tier 2/Sulfur (Annual Tons)
As shown, the implementation of reduced sulfur levels in 2004 would result in an
immediate benefit of over one-half million tons, a 14 percent drop in uncontrolled 2004 light-
duty emissions; this is the equivalent of emissions produced by over 26 million pre-Tier 2 cars
and trucks.d'8 In 2004, nearly all of the benefits would be due to reduced emissions from Tier 0,
Tier 1 and NLEV vehicles.
After 2004, emission are reduced further as the fleet turns over to predominantly Tier 2
vehicles operating on low sulfur fuel, versus NLEVs and Tier 1 trucks operating on current in-
use sulfur levels. By 2020, the projected benefit represents a two-thirds reduction in 2020 light-
duty emissions without Tier 2/Sulfur, equivalent to the emissions from over 166 million pre-Tier
2 cars and trucks. NOx emissions from all sources would be reduced by 12 percent.
We project that light-duty emissions would continue to decrease beyond 2020, reversing
the upward emissions trend in the baseline case brought on by VMT growth. By 2030,
essentially the entire fleet would consist of Tier 2 vehicles. The projected benefit of 2.8 million
tons in this year represents a nearly three-quarters reduction in 2030 light-duty emissions without
Tier 2/Sulfur, equivalent to the emissions from 213 million pre-Tier 2 cars and trucks. These
emission reductions would amount to 15 percent of total man-made NOx emissions in that year
i.e., vehicles that would be on the road in the absence of Tier 2/Sulfur control.
Ill-11
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
in the absence of today's proposal.
The estimated percentage reductions in total inventory presented in Table III-4 include
benefits that would be realized on heavy-duty gasoline vehicles due to sulfur control. We
estimate these heavy-duty emission reductions to be on the order of approximately 30,000 tons
per year for every year starting in 2004, as shown in Appendix A.
NOx reductions due to today's proposal would be of a similar scope in urban areas.
Table III-5 shows NOx emissions reduced due to Tier 2/Sulfur control, and light-duty highway
vehicle emissions remaining, for each of the four cities. Table III-6 presents these reductions in
terms of the percentage of baseline light-duty and total inventory reduced.
Table III-5. Four-City Light-Duty NOx Emissions With Tier 2/Sulfur (Summer Tons)
Year
2004
2007
2010
2015
2020
2030
New York
Reduced
7,368
12,139
18,432
27,544
33,177
39,488
Remain
59,489
45,614
33,380
20,089
14,857
12,792
Chicago
Reduced
3,062
5,546
8,915
14,020
17,296
21,259
Remain
29,252
22,853
17,043
10,421
7,784
6,906
Atlanta
Reduced
4,550
7,346
10,975
16,483
20,188
25,160
Remain
26,362
20,967
15,871
9,901
7,534
6,857
Charlotte
Reduced
666
1,098
1,668
2,567
3,206
4,117
Remain
3,860
3,133
2,413
1,542
1,196
1,122
III-12
-------
Chapter III: Environmental Impact
Table III-6. Percent Reduction From Light-Duty and
Total Baseline NOx Emissions in Four Cities
Year
2004
2007
2010
2015
2020
2030
New York
Light-
Duty
11%
21%
36%
58%
69%
76%
All
Sources*
-
6%
10%
16%
19%
21%
Chicago
Light-
Duty
9%
20%
34%
57%
69%
75%
All
Sources*
-
4%
6%
10%
13%
15%
Atlanta
Light-
Duty
15%
26%
41%
62%
73%
79%
All
Sources*
-
10%
16%
25%
30%
33%
Charlotte
Light-
Duty
15%
26%
41%
62%
73%
79%
All
Sources *
-
5%
8%
12%
15%
18%
* Includes emission reductions from Heavy-Duty Gasoline Vehicles due to sulfur control
The magnitude of reductions in urban areas reflect those nationally. An immediate
reduction in light-duty emission would result from sulfur control, even in RFG areas (New York
and Chicago). Over one-third of baseline light-duty emissions would be reduced by 2010 in
each city. Light-duty emissions would be reduced by roughly 70 percent in 2020 and over 75
percent in 2030. Reductions in Atlanta and Charlotte are consistently larger in percentage
terms than in New York and Chicago because they are not RFG areas; emission reductions in
non-RFG urban areas would be particularly large since these areas would tend to have higher-
sulfur fuel than RFG areas in the absence of today's proposal. We project that emissions would
continue to decrease through at least 2028 in all four cities, indicating that today's program
would be successful in reducing light-duty NOx emissions in the face of high VMT growth rates.
The impact on total inventory would also be significant, particularly in New York and
Atlanta. By 2020, we project that the total NOx inventory would be reduced by nearly one-fifth
in New York and one-third in Atlanta due to Tier 2/Sulfur control.
Concurrently, we project that the light-duty contribution to total NOx emissions would
drop significantly. Figures III-5 and III-6 show our 2020 projections of this contribution in the
47 states and in Atlanta with Tier 2/Sulfur control. Table III-7 shows this same contribution
across the 47 states and all four cities from 2007 through 2030. In 2020, we project that the
light-duty contribution would drop to seven percent nationally, from 18 percent without Tier
2/Sulfur control. This trend is similar across the four cities, depending on the level of
contribution without Tier 2/Sulfur control. We project that with Tier 2/Sulfur control, car and
truck emissions would contribute 10 percent of total emissions in New York (down from 27
percent), seven percent in Chicago and Charlotte (down from 19 percent and 21 percent), and 16
percent in Atlanta (down from 40 percent) in 2020.
Ill-13
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Light-Duty
Vehicles and
Trucks
7%
Other On-
Highway
N on road
18%
Stationary and
Area
67%
Figure III-5. Breakdown of Total 2020 47-State NOx Inventory With Tier 2/Sulfur
Stationary and
Area
29%
Light-Duty
Vehicles and
Trucks
16%
Other On-
H ighway
18%
Nonroad
37%
Figure III-6. Breakdown of Total 2020 Atlanta NOx Inventory With Tier 2/Sulfur
III-14
-------
Chapter III: Environmental Impact
Table III-7. Light-Duty Contribution to Total NOx Inventory With Tier 2/Sulfur
Year
2007
2010
2015
2020
2030
47 State
14%
11%
8%
7%
6%
New York
24%
20%
14%
10%
9%
Chicago
16%
13%
9%
7%
6%
Atlanta
31%
27%
20%
16%
13%
Charlotte
14%
12%
8%
7%
6%
c. NOx Emission Reductions From Other Options
We developed 47-state light-duty emission inventory projections for three alternative
vehicle/fuel control options to allow comparison with the emission reductions projected to result
from today's action. These alternative options are:
1) Car and truck emission standards and implementation schedule as proposed in today's
action in conjunction with sulfur control proposed to EPA by the American Petroleum
Institute (API) and National Petroleum Refiners Association (NPRA). Under this plan,
sulfur would be reduced in 2004 to 150 ppm in the eastern half of the U.S., referred to as
the "API NOx Control Region", and 300 ppm in the remainder of the 49-state region.6
2) Option (1) above with implementation of a "rebuttable" element of the API/NPRA
proposal in which sulfur would be reduced to 30 ppm in 2010 in the API NOx Control
Region, while the remainder of the country remains at 300 ppm.
3) Sulfur control as proposed in today's action in conjunction with the default Tier 2 car
and truck emission standards contained in the Clean Air Act. Under this alternative,
LDVs and LDTls would be required to meet full useful life emission standards of 0.125
g/mi NMHC and 0.20 g/mi NOx, assumed for this analysis to follow the implementation
schedule for Tier 2 standards contained in today's proposal. LDT2s would be subject to
California's applicable LEV I standards in 2004, while LDT3s and LDT4s would remain
at Tier 1 levels.
For Options 1 and 2, the effects of sulfur irreversibility were accounted for using the
methodology described in detail in Appendix B. In short, all cars and trucks complying with the
Supplemental Federal Test Procedure (SFTP) were assigned an irreversibility effect of 50
eThe API/NPRA fuel proposal is discussed in detail in the Preamble, Section IV.C. 1
III-15
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
percent, meaning that vehicles within the API NOx Control Region exposed to higher sulfur
levels outside the region would experience a permanent degradation in emissions performance
equivalent to the average of emissions generated on fuel in and outside of the API Region. It
was assumed that at any given time 25 percent of cars and trucks in the API NOx Control Region
fleet would have traveled outside of the region, and hence been exposed to higher sulfur fuel/
47-state light-duty NOx emissions projected for these options are shown in Tables III-8,
in comparison with today's proposal. Table III-9 provide a direct comparison to today's
proposal in terms of shortfall (i.e., emission reductions "lost" by the three options compared to
today's proposal) and total benefits relative to the Tier 2/Sulfur proposal.
Table III-8. 47-State Light-Duty NOx Reductions From Alternative Control Options
(Annual Tons)8
Option:
Vehicle Program:
Fuel Program:
2007
2010
2020
Today 's
Proposal
Proposed Tier 2
Proposed Tier 2
795,733
1,182,323
2,198,113
1
Proposed Tier 2
API
No 30 ppm
397,886
750,100
1,713,531
2
Proposed Tier 2
API
3 0 ppm API Region
2010
397,886
1,020,812
2,000,129
3
Clean Air Act
Default
Proposed Tier 2
611,020
740,258
1,026,690
fThe baseline emission inventory estimates presented here do not account for sulfur irreversibility effects
in RFG areas. Although vehicles in these areas will likely experience irreversibility effects due to exposure to
higher sulfur levels during winter months, the overall impact on baseline emissions are expected to be small because
a) LDT2/3/4s are less sensitive to sulfur under NLEV than expected under the standards proposed in today's action,
and b) vehicles operating on summertime RFG make up a relatively small portion (less than 15%) of annual VMT
in the 47-state region. Accounting for this effect would serve to increase the estimated benefits of today's proposal.
gAlthough not shown, Options 1 and 2 will also increase emissions from heavy-duty gasoline vehicles
relative to today's action due to higher sulfur levels.
Ill-16
-------
Chapter III: Environmental Impact
Table III-9. NOx Reduction Shortfall From Alternative Control
Options Relative to Today's Proposal
Option:
Vehicle Program:
Fuel Program:
Year
2007
2010
2020
1
Proposed Tier 2
API
No 30 ppm
Shortfall
(Annual
Tons)
397,848
432,223
484,582
Benefit
Relative to
Tier 2/Sulfur
Proposal
50%
63%
78%
2
Proposed Tier 2
API
30 ppm API Region 2010
Shortfall
(Annual
Tons)
397,848
162,012
197,984
Benefit
Relative to
Tier 2/Sulfur
Proposal
50%
86%
91%
3
Clean Air Act Default
Proposed Tier 2
Shortfall
(Annual
Tons)
184,713
442,066
1,171,423
Benefit
Relative to
Tier 2/Sulfur
Proposal
77%
63%
47%
2.
voc
a. Light-Duty VOC Trends Without Tier 2/Sulfur
Total VOC emissions produced nationwide by cars and trucks without Tier 2/Sulfur
control are shown in Table III-10 and Figure III-7, broken down by relative contribution of
evaporative emissions (across all cars and trucks), and exhaust emissions for LDVs, LDTl/2s
and LDT3/4s. We project VOC emissions from light-duty vehicles will decline from
approximately 3.8 million tons to 2.0 million tons between 2000 and 2015 as the fleet becomes
increasingly dominated by cars and trucks complying with NLEV, Enhanced Evaporative control
and SFTP requirements. Beginning in 2016, however, light-duty VOC emissions are projected
to begin an upward trend due to VMT and vehicle fleet growth, increasing to 2.1 million tons by
2020 and 2.5 million tons by 2030.
Ill-17
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table 111-10. 47-State Light-Duty VOC Emissions Without Tier 2/Sulfur (Annual Tons)
Year
2000
2004
2007
2010
2015
2020
2030
Light-Duty
Emissions
3,771,569
2,969,912
2,503,855
2,227,593
2,080,284
2,132,070
2,475,783
Contribution by Emission Source / Vehicle Class
Evaporative
(AllLDV/LDT)
44.3%
46.8%
50.4%
52.1%
54.1%
54.7%
54.8%
Exhaust
LDV
22.6%
18.3%
15.3%
12.6%
9.9%
9.1%
8.8%
LDT1/2
20.1%
21.1%
20.4%
19.8%
18.2%
18.0%
18.1%
LDT3/4
13.0%
13.7%
14.0%
15.5%
17.8%
18.2%
18.2%
4,000,000
3,000,000
2,000,000
1,000,000
2005
2010
2015
2020
2025
2030
Evap (All LD) • LDV Exhaust • LDT1/2 Exhaust p LDT3/4 Exhaust
Figure III-7. 47-State Light-Duty VOC Emissions Without Tier 2/Sulfur (Annual Tons)
III-18
-------
Chapter III: Environmental Impact
Although evaporative emissions are projected to grow to over half of the light-duty
inventory, exhaust emissions from trucks play an increasingly significant role in shaping the
overall VOC trend. In 2000, we project that trucks will produce approximately 60 percent of
exhaust VOC emissions; by 2020, trucks account for 80 percent of these emissions, while overall
emissions produced by trucks increase steadily. The benefits from Tier 1, NLEV and SFTP are
not as pronounced for trucks relative to cars, and are offset almost immediately by growth in
truck VMT. As a result, exhaust VOC emissions from trucks see only modest initial reductions
due to these programs before being offset by VMT growth.
The emission trends for the four urban areas we analyzed show similar behavior to the
national trends. As shown in Table III-l 1, light-duty emissions decrease steadily in each city
through 2010. In all cities, the decline in emissions due to existing vehicle standards essentially
ends by 2020, after which VOC emissions are projected to increase if today's proposal is not
adopted.
Table III-l 1. Four-City Light-Duty VOC Emissions Without Tier 2/Sulfur (Summer Tons)
Year
2000
2004
2007
2010
2015
2020
2030
New York
56,782
40,063
30,609
25,732
22,062
21,124
22,744
Chicago
27,145
19,768
15,404
13,151
11,386
11,061
12,264
Atlanta
28,791
22,166
18,139
15,869
14,239
14,195
16,149
Charlotte
4,080
3,245
2,710
2,412
2,217
2,254
2,642
Figures III-8 and III-9 show our projections of the contribution of light-duty vehicles and
trucks to the total anthropogenic (i.e., human-caused) 2020 VOC inventory in the 47 states and
in Atlanta. Table 111-12 shows this same contribution across the 47 states and all four cities from
2007 through 2030. Nationally, cars and trucks produce nearly one-fifth of total VOC emissions
in 2007; this percentage declines subsequent years before stabilizing at 14 percent by 2015 and
increasing after 2020. The light-duty contribution in New York, Chicago, and Charlotte are
slightly lower than the national average, but significantly higher in Atlanta, where we project
that one-fourth of all VOC emissions will be produced by cars and trucks in 2020.
Ill-19
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Light-Duty
Vehicles and
Trucks otherOn-
140/0 Highway
3%
Nonroad
18%
Stationary and
Area
65%
Figure III-8. Breakdown of Total 2020 47-State VOC Inventory Without Tier 2/Sulfur
Light-Duty
Vehicles and
Trucks
26%
Stationary and
Area
50%
Nonroad
16%
Other On-
Highway
8%
Figure III-9. Breakdown of Total 2020 Atlanta VOC Inventory Without Tier 2/Sulfur
m-20
-------
Chapter III: Environmental Impact
Table 111-12. Light-Duty Contribution to Total VOC Inventory Without Tier 2/Sulfur
Year
2007
2010
2015
2020
2030
47 State
18%
16%
14%
14%
15%
New York
15%
13%
11%
10%
10%
Chicago
12%
11%
10%
9%
9%
Atlanta
33%
31%
28%
26%
26%
Charlotte
15%
14%
12%
12%
12%
b.
VOC Reductions Due To Tier 2/Sulfur
Table III-13 contains annual nationwide tons of VOC we project would be reduced due to
today's proposal, encompassing the effects of low sulfur fuel and the introduction of Tier 2
light-duty vehicle and light-duty truck standards for both exhaust and evaporative emissions.
Figure III-10 shows projected 47-state emissions with Tier 2/Sulfur control, broken down by
light-duty evaporative emissions and exhaust emissions from LDVs, LDTl/2s and LDT3/4s.
Table 111-13. 47-State Light-Duty VOC Reductions Due to Tier 2/Sulfur (Annual Tons)
Year
2004
2007
2010
2015
2020
2030
Light-Duty
Emissions Without
Tier 2/Sulfur
2,969,912
2,503,855
2,227,593
2,080,284
2,132,070
2,475,783
Light-Duty
Emissions With
Tier 2/Sulfur
2,865,843
2,372,427
2,050,465
1,821,904
1,800,394
2,039,802
Emissions
Reduced
104,069
131,428
177,128
258,380
331,676
435,981
Percent Reduction in
Baseline Inventory
Light-Duty
4%
5%
8%
12%
16%
18%
All
Sources*
-
1.0%
1.3%
1.8%
2.3%
2.7%
Includes emission reductions from Heavy-Duty Gasoline Vehicles due to sulfur control
m-21
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
4,000,000
3,000,000
2,000,000
1,000,000
WTTHOUT T ER 2/SULFUR CONTROLS
2005
2010
2015
2020
2025
2030
I Evap (All LD) • LDV Exhaust • LDT1/2 Exhaust D LDT3/4 Exhaust
Figure 111-10. 47-State Light-Duty VOC Emissions With Tier 2/Sulfur (Annual Tons)
We project that lower sulfur levels in 2004 would reduce light-duty emissions four
percent, due almost entirely to reduced emission from Tier 0, Tier 1 and NLEV vehicles; this is
the equivalent of emissions produced by 6.5 million pre-Tier 2 cars and trucks. After 2004, the
introduction of LDT2s, LDT3s, and LDT4s complying with the Tier 2 NMOG standard and
operating on low sulfur fuel reduce emission further. By 2020, baseline light-duty VOC
emissions are reduced 16 percent due to Tier 2/Sulfur control, the equivalent of emissions from
38 million pre-Tier 2 cars and trucks. This represents a 2.3 percent reduction of the total
anthropogenic VOC inventory. With Tier 2/Sulfur, we project that the upturn in light-duty VOC
emissions will begin in 2021, five years later than the baseline case.
In addition to emission benefits on light-duty vehicles and trucks, we project that heavy-
duty gasoline vehicles would decrease emissions by approximately 7,000 tons per year
beginning in 2004, growing to 12,000 tons in 2030. These reductions are shown in Appendix A,
and are included in the estimates of mobile source and all source percent reduction contained in
TableIII-13.
Tables III-14 and III-15 show VOC reductions in the four cities in both tonnage and
percentage terms; the percentage reductions are expressed relative to light-duty emissions and
total anthropogenic emissions if today's proposal were not adopted. VOC reductions would be
larger in these areas in percentage terms than is the average throughout the 47 states. In 2020,
111-22
-------
Chapter III: Environmental Impact
we project that 23 percent of light-duty VOC emissions would be reduced in these cities, versus
16 percent for the 47-state region. This difference is driven by the presence of I/M in each area
and RFG in some of these areas. As modeled, vehicles with malfunctioning emission control
systems do not realize the full benefit of the proposed Tier 2 vehicle standards. With I/M, it is
assumed that a good portion of these vehicles are identified and repaired, thus increasing the
relative benefit of the Tier 2/Sulfur program.11 The reductions in total anthropogenic VOC
inventory are generally consistent with the 47-state results, although in Atlanta the reductions are
larger; by 2020, we project that 6.1 percent of Atlanta's total VOC emission would be reduced
by today's action, versus 2.3 percent nationally.
Table 111-14. Four-City Light-Duty VOC Reductions Due To Tier 2/Sulfur (Summer Tons)
Year
2004
2007
2010
2015
2020
2030
New York
Reduced
1,282
1,870
2,675
3,919
4,919
6,031
Remain
38,781
28,739
23,057
18,143
16,205
16,713
Chicago
Reduced
611
913
1,333
2,007
2,568
3,247
Remain
19,157
14,490
11,818
9,380
8,493
9,017
Atlanta
Reduced
1,110
1,462
1,860
2,592
3,240
4,184
Remain
21,056
16,677
14,009
11,648
10,955
11,965
Charlotte
Reduced
163
218
283
404
515
685
Remain
3,083
2,492
2,130
1,814
1,740
1,958
The approach used for developing I/M benefits for Tier 2 vehicles is discussed in detail in the technical
report "Development of Light-Duty Emission Inventory Estimates in the Notice of Proposed Rulemaking for Tier 2
and Sulfur Standards"
111-23
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table III-15. Percent Reduction From Light-Duty and
Total Baseline VOC Emissions in Four Cities
Year
2004
2007
2010
2015
2020
2030
New York
Light-
Duty
3%
6%
10%
18%
23%
27%
All
Sources*
-
0.9 %
1.4%
2.0%
2.4%
2.8%
Chicago
Light-
Duty
3%
6%
10%
18%
23%
26%
All
Sources*
-
0.8%
1.1%
1.7%
2.1%
2.4%
Atlanta
Light-
Duty
5%
8%
12%
18%
23%
26%
All
Sources*
-
2.8%
3.7%
5.1%
6.1%
6.9%
Charlotte
Light-
Duty
5%
8%
12%
18%
23%
26%
All
Sources *
-
1.3%
1.7%
2.3%
2.8%
3.3%
* Includes emission reductions from Heavy-Duty Gasoline Vehicles due to sulfur control
Figures III-l 1 and III-12 show the contribution of light-duty cars and trucks to total 2020
VOC inventory in the 47 states and in Atlanta with Tier 2/Sulfur control. Table 111-16 shows this
same contribution across the 47 states and all four cities from 2007 through 2030. In 2020, the
light-duty contribution would drop to 12 percent nationally, from 14 percent without Tier
2/Sulfur control. This trend would be similar across the four cities, depending on the level of
light-duty contribution without Tier 2/Sulfur control. We project that with Tier 2/Sulfur control,
car and truck emissions would contribute eight percent of total emissions in New York (down
from 10 percent), seven percent in Chicago (down from nine percent), ten percent in Charlotte
(down from 12 percent), and 22 percent in Atlanta (down from 26 percent).
111-24
-------
Chapter III: Environmental Impact
Light-Duty
Vehicles and
Trucks
12%
Other On-
H ighway
3%
Nonroad
18%
Stationary and
Area
67%
Figure III-ll. Breakdown of Total 2020 47 State VOC Inventory With Tier 2/Sulfur
Light-Duty
Vehicles and
Trucks
22%
Stationary and
Area
54%
Other On-
H ighway
7%
Nonroad
17%
Figure 111-12. Breakdown of Total 2020 Atlanta VOC Inventory With Tier 2/Sulfur
111-25
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table III-16. Light-Duty Contribution to Total VOC Inventory With Tier 2/Sulfur
Year
2007
2010
2015
2020
2030
47 State
17%
15%
13%
12%
13%
New York
14%
12%
9%
8%
8%
Chicago
12%
10%
8%
7%
7%
Atlanta
32%
28%
24%
22%
21%
Charlotte
14%
12%
10%
10%
10%
c. VOC Emission Reductions From Other Options
VOC reductions for the three alternative options discussed under Section III.A.l.c above
are shown in Tables 111-17 and 111-18, in comparison to reductions projected from today's
proposal. It it assumed for this analysis that the evaporative controls contained in today's action
would be included in each option.
Table 111-17. 47-State Light-Duty VOC Reductions From Alternative Control Options
(Annual Tons)
Option:
Vehicle Program:
Fuel Program:
2007
2010
2020
Today 's
Proposal
Proposed Tier 2
Proposed Tier 2
131,428
177,128
331,676
1
Proposed Tier 2
API
No 30 ppm
74,331
118,809
264,220
2
Proposed Tier 2
API
3 0 ppm API Region
2010
74,331
155,750
305,361
3
Clean Air Act
Default
Proposed Tier 2
101,706
107,955
131,552
m-26
-------
Chapter III: Environmental Impact
Table 111-18. VOC Reduction Shortfall From Alternative
Control Options Relative to Today's Action
Scenario:
Vehicle Program:
Fuel Program:
Year
2007
2010
2020
1
Proposed Tier 2
API
No 30 ppm
Shortfall
(Annual
Tons)
57,097
58,319
67,456
Benefit
Relative to
Tier 2/Sulfur
Proposal
57%
67%
80%
2
Proposed Tier 2
API
30 ppm API Region 2010
Shortfall
(Annual
Tons)
57,097
21,378
26,315
Benefit
Relative to
Tier 2/Sulfur
Proposal
57%
88%
92%
3
Clean Air Act Default
Proposed Tier 2
Shortfall
(Annual
Tons)
29,722
69,173
200,123
Benefit
Relative to
Tier 2/Sulfur
Proposal
77%
61%
40%
3.
SOx
a. Light-Duty SOx Trends Without Sulfur Control
Gaseous SOx emissions are formed by the combustion of fuel sulfur, and hence depend
entirely on the level of sulfur in the fuel. SOx emissions without sulfur control are shown in
Table 111-19 and Figure 111-13, broken down by LDV, LDT1/2 and LDT3/4. As shown, we
project that SOx emission levels will increase unabated through 2030 in conjunction with VMT
growth in the absence of any action to reduce fuel sulfur levels. In 2000, we project light-duty
vehicles and trucks will emit 194,000 tons of SOx; by 2020, this level is projected to be nearly
300,000 tons, an increase of 55 percent.
Ill-
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table 111-19. 47-State SOx Emissions Without Sulfur Control (Annual Tons)
Year
2000
2004
2005
2010
2015
2020
2030
Emissions
From All
Sources
-
-
18,045,277
18,350,974
18,773,428
19,161,564
20,099,769
Light-Duty
Emissions
193,467
211,072
215,659
240,694
270,174
299,959
357,611
Light-Duty
Contribution
to All
Sources
-
-
1.2%
1.3%
1.4%
1.6%
1.8%
Contribution by Vehicle Class
LDV
48%
41%
40%
33%
29%
28%
27%
LDT1/2
39%
44%
45%
50%
53%
54%
55%
LDT3/4
13%
15%
15%
17%
18%
18%
18%
400,000
300,000
200,000
100,000
2005
2010
2015
2020
2025
LDV
LPT 1/2 DLDT3/4
2030
Figure 111-13. 47-State Light-Duty SOx Emissions Without Sulfur Control (Annual Tons)
Trucks, primarily LDTls and LDT2s, are responsible for the steady increase in light-duty
SOx emissions. While LDV SOx emissions are relatively stable, SOx emissions from trucks
(and hence the contribution to light-duty inventory produced by trucks) are projected to increase
111-28
-------
Chapter III: Environmental Impact
steadily. In 2000, trucks account for roughly half of light-duty SOx emissions, growing to over
70 percent by 2020.
b.
SOx Reductions Due To Sulfur Control
We project that today's proposal would immediately and substantially reduce SOx
emissions from cars and trucks once its fuel sulfur provisions take effect. Table 111-20 contains
annual nationwide tons of gaseous SOx we project will be reduced from light-duty vehicles and
trucks due to sulfur control. Figure 111-14 shows SOx emissions after sulfur control, broken
down by LDV, LOT 1/2 and LDT3/4.
Table 111-20. 47-State Light-Duty SOx Reductions Due To Sulfur Control (Annual Tons)
Year
2004
2005
2010
2015
2020
2030
Light-Duty
Emissions
Without Sulfur
Control
211, Ql 2
215,659
240,694
270,174
299,959
357,611
Light-Duty
Emissions
With Sulfur
Control
21,426
21,899
24,257
27,210
30,203
36,002
Emissions
Reduced
189,646
193,760
216,437
242,964
269,756
321,609
Percent Reduction in
Baseline Inventory
Light-Duty
90%
90%
90%
90%
90%
90%
All Sources*
-
1.3%
1.4%
1.5%
1.6%
1.8%
* Includes reductions from Heavy-Duty Gasoline Vehicles, Motorcycles and Nonroad Sources
111-29
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
400,000
300,000
200,000
100,000
WITHOUT SULFUR CONTROL
2005
2010
2015
2020
2025
2030
ILDV
ILDT1/2
DLDT3/4
Figure 111-14. 47-State Light-Duty SOx Emissions With Tier 2/Sulfur (Annual Tons)
As shown, a 90 percent reduction in light-duty SOx emissions would be realized
beginning in 2004. This relative reduction remains constant beyond 2004, since SOx emissions
are not reduced further as new Tier 2 VOC, NOx, and PM standards are phased in. The absolute
level of emission reductions would become larger with time, however, due to VMT growth.
SOx emission reductions will also occur from heavy-duty gasoline vehicles and
motorcycles due to sulfur control; we estimate this reduction to be approximately 10,000 tons in
2005, growing to 14,000 tons by 2020. In addition, emissions from all gasoline-powered
nonroad equipment would be reduced due to sulfur control. Based on our NONROAD model,
we estimate this benefit would be approximately 25,000 tons per year on average between 2005
and 2020. These reductions, shown in Appendix A, are included in the percent reductions from
all sources in Table 111-20.
4.
Particulate Matter
Trends in particulate matter emissions will depend very strongly on the prevalence of
diesel vehicles in the light-duty fleet. Currently, diesels make up a very small portion (less than
one percent) of overall car and truck sales. However, sharp increases in diesel sales are a
reasonable possibility given the focus on diesel technology for improving fuel economy under
111-30
-------
Chapter III: Environmental Impact
the Partnership for a New Generation of Vehicles (PNGV). Thus, we assessed PM emissions
under two sales scenarios: a "no growth" scenario, for which current diesel sales trends were
assumed to continue, and an "increased growth" scenario, for which diesels grow to 50 percent
of light-duty truck sales by 2010. The effects of Tier 2/Sulfur control were assessed for both
scenarios. The results presented here are for direct exhaust PM, comprising carbonaceous PM
and sulfate emitted directly from the tailpipe and a subset of Total PM (which also includes
direct non-exhaust PM from tire and brake wear, and indirect PM caused by secondary reactions
to emitted NOx and SOx in the atmosphere). Direct PM2 5 and PM10 emissions are presented
separately for the "no growth" scenario.
a.
"No Growth" Diesel Sales Scenario
/'. Light-Duty Direct Exhaust PM25 Trends Without Tier 2/Sulfur
In general, gasoline vehicles emit PM at rates much lower than their diesel counterparts.
Under the no growth scenario, direct PM emissions are driven largely by sulfate emissions from
gasoline vehicles, which depend primarily on gasoline fuel sulfur level. Without Tier 2/Sulfur
control, these emissions increase at a steady rate in conjunction with VMT, as shown in Tables
111-21 and Figure 111-15. In 2005, we project that approximately 35,000 tons will be emitted
annually by light-duty cars and trucks. This level is projected to exceed 47,000 tons in 2020 and
56,000 tons in 2030.
Table 111-21. 47 State Light-Duty Direct Exhaust PM2 5 Emissions Without Tier 2/Sulfur
No Growth in Diesel Sales
(Annual Tons)
Year
2000
2004
2005
2010
2015
2020
2030
Emissions
From All
Sources*
-
-
2,071,897
2,108,058
2,217,074
2,318,805
2,544,434
Light-
Duty
Exhaust
Emissions
34,072
34,612
35,051
38,409
42,724
47,397
56,505
Light-Duty
Contribution
to All
Sources
-
-
1.7%
1.8%
1.9%
2.0%
2.2%
Contribution by Fuel Type / Vehicle Class
Diesel
LDV/LDT
5%
3%
3%
2%
2%
2%
2%
Gas
LDV
45%
42%
40%
34%
30%
28%
28%
Gas
LDT1/2
34%
41%
42%
48%
51%
53%
53%
Gas
LDT3/4
17%
14%
15%
16%
17%
17%
17%
* Excludes natural and miscellaneous sources (e.g., fugitive dust), but includes indirect sources such as tire and
brake wear.
III-31
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
60,000
50,000
40,000
30,000
20,000
10,000
2000
2005
2010
2015
2020
2025
2030
Diesel
LDVGas BLDT1/2 Gas
LDT3/4 Gas
Figure 111-15. 47-State Light-Duty Direct Exhaust PM2 5 Emissions Without Tier 2/Sulfur
-No Diesel Growth (Annual Tons)
As expected, the diesel contribution to overall emissions in the no growth scenario is
relatively small. Rather, gasoline trucks (primarily LDTls and LDT2s) are responsible for the
steady increase in PM emissions. Under this scenario, we project the contribution of gasoline
trucks to light-duty PM2 5 inventory to grow to 70 percent by 2020.
/'/'. Direct Exhaust PM25 Reductions Due To Tier 2/Sulfur Control
Under the no growth scenario, today's proposal would provide an immediate and
substantive reduction in direct PM emissions from cars and trucks, due primary to sulfur control.
Table 111-22 contains annual nationwide tons of direct exhaust PM25 we project would be
reduced from light-duty vehicles and trucks due to Tier 2/Sulfur control. Figure III-16 shows
PM2 5 emissions after Tier 2/Sulfur control broken down by diesel (all light-duty cars and trucks)
and gasoline LDV, LOT 1/2 and LDT3/4.
111-32
-------
Chapter III: Environmental Impact
Table 111-22. 47-State Light-Duty Direct Exhaust PM2 5 Reductions Due To Tier 2/Sulfur
No Growth in Diesel Sales
(Annual Tons)
Year
2004
2005
2010
2015
2020
2030
Light-Duty
Emissions Without
Tier 2/Sulfur
34,612
35,051
38,409
42,724
47,397
56,505
Light-Duty
Emissions With
Tier 2/Sulfur
14,703
14,509
14,999
16,129
17,690
20,956
Emissions
Reduced
19,909
20,542
23,410
26,595
29,707
35,549
Percent Reduction in
Baseline Inventory
Light-Duty
58%
59%
61%
62%
63%
63%
All Sources*
-
1.0%
1.1%
1.2%
1.3%
1.4%
* Includes emission reductions from Heavy-Duty Gasoline Vehicles due to sulfur control
60,000
50,000
40,000
2000
WITHOUT TIER 2/SULFUR CONTROLS
2005
2010
2015
2020
2025
2030
Diesel
LDV Gas
• LDT1/2 Gas
DLDT3/4 Gas
Figure 111-16. 47-State Light-Duty Direct Exhaust PM2 5 Emissions With Tier 2/Sulfur -
No Diesel Growth (Annual Tons)
m-33
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Reductions from gasoline vehicles would result almost entirely from sulfur control, rather
than the proposed PM2 5 exhaust standards. PM2 5 emissions on current technology gasoline
vehicles are much lower than diesel vehicles, and gasoline vehicle emissions are not expected to
be reduced in response to the PM25 standards contained in today's proposal. As such, we project
that an immediate emission reduction of 58 percent from baseline levels would be realized due to
sulfur control, increasing to 63 percent by 2020.
In addition to light-duty PM benefits, sulfur control would reduce PM2 5 emissions from
heavy-duty gasoline vehicles. We estimate these benefits would be approximately 700 tons per
year beginning in 2004, increasing to 1,000 tons by 2020. Across all sources, we project Tier
2/Sulfur control would reduce direct PM2 5 from all non-natural sources by about one percent.
/'/'/'. Direct Exhaust PM10 Reductions Due To Tier 2/Sulfur Control
Direct exhaust PM10 emissions exhibit similar trends to PM2 5, and are thus shown here
only for the no growth diesel case; PM10 emissions with and without Tier 2/Sulfur control are
shown in Table 111-23.
Table 111-23. 47-State Light-Duty PM10 Emissions With and Without Tier 2/Sulfur Control
No Growth in Diesel Sales
(Annual Tons)
Year
2004
2005
2010
2015
2020
2030
Emissions
From All
Sources
Without
Tier 2/Sulfur**
-
2,985,623
3,060,154
3,207,687
3,345,810
3,659,928
Light-Duty
Exhaust
Emissions
Without
Tier 2/Sulfur
37,323
37,794
41,412
46,064
51,102
60,922
Light-Duty
Contribution
to All
Sources
-
1.3%
1.4%
1.4%
1.5%
1.7%
Light-Duty
Exhaust
Emissions
With
Tier 2/Sulfur
15,861
15,649
16,173
17,390
19,071
22,591
Emissions
Reduced
21,462
22,145
25,239
28,674
32,031
38,331
Percent Reduction
in Baseline
Inventory
Light-
Duty
58%
59%
61%
62%
63%
63%
All
Sources*
-
0.8%
0.9%
0.9%
1.0%
1.1%
* Includes emission reductions from Heavy-Duty Gasoline Vehicles due to sulfur control
** Excludes natural and miscellaneous sources (e.g., fugitive dust), but includes indirect sources such as tire and
brake wear.
b.
'Increased Growth" Sales Scenario
The "increased growth" scenario was developed with the intent of analyzing an upper
111-34
-------
Chapter III: Environmental Impact
bound for diesel growth. We developed this scenario by assuming that the percent of diesels
making up total light-duty truck sales increase to five percent in 2001, adding five percentage
points per subsequent year until diesels represent 50 percent of light-duty truck sales in 2010;
beyond 2010, the diesel engine share of the light truck market was assumed to stay at 50 percent.
Within the period of diesel sales growth, we assumed that light duty truck classes were
"converted" to diesels in a sequential manner starting with the heaviest trucks; i.e., LDT4s
became diesels first, then LDT3s, etc. This methodology resulted in the diesel sales penetrations
shown in Table 111-24.
Table 111-24. Diesel LDT Sales Penetration Under Increased Growth Scenario
Model Year
2001
2002
2003
2004
2005
2006
2007
2008
2009
20 10 and later
Diesel Sales Penetration
All LDT LDT2 LDT3 LDT4
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
0%
0%
0%
0%
0%
9%
18%
26%
35%
44%
0%
12%
41%
71%
100%
100%
100%
100%
100%
100%
63%
100%
100%
100%
100%
100%
100%
100%
100%
100%
/'. Light-Duty Direct Exhaust PM25 Trends Without Tier 2/Sulfur
Our projections for light-duty direct exhaust PM2 5 under the increased diesel sales
scenario are down in Table 111-25 and Figure III-17. As expected, this scenario is projected to
result in dramatic increases in light-duty PM2 5 emissions. 2005 baseline emissions are
approximately 43,000 tons, 23 percent higher than the 35,000 tons projected in the no growth
diesel case from Table 111-21. However, by 2020, we project this scenario would result in direct
PM emissions of 138,000 tons, nearly three times the emissions projected for the no growth
scenario in the same year.
m-35
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table 111-25. 47 State Light-Duty Direct Exhaust PM2 5 Emissions Without Tier 2/Sulfur
Increased Diesel Growth Scenario
(Annual Tons)
Year
2000
2004
2005
2010
2015
2020
2030
Light-Duty
Emissions Without
Tier 2/Sulfur
34,072
39,932
43,439
72,626
109,622
138,177
175,068
Contribution by Fuel Type
Diesel LDV/LDT
5%
19%
25%
56%
72%
77%
80%
Gasoline LDV/LDT
95%
81%
75%
44%
28%
23%
20%
2000
2005
2010
2015
2020
2025
Diesel
Gasoline
2030
Figure 111-17. 47-State Light-Duty Direct Exhaust PM2 5 Without Tier 2/Sulfur
Increased Diesel Sales (Annual Tons)
m-36
-------
Chapter III: Environmental Impact
As shown, the rapid growth of diesels in conjunction with high per-vehicle PM emissions
from diesels drive overall direct PM emissions under this scenario. In 2005, we project diesels
would already account for 25 percent of all light-duty emissions. Diesel contribution grows to
over 50 percent by 2010 and over 75 percent by 2020.
/'/'. Direct Exhaust PM25 Reductions Due To Tier 2/Sulfur
Tier 2/Sulfur control would effectively neutralize excess PM emissions generated under
our increased diesel penetration scenario. Table 111-26 contains reductions in direct exhaust
PM2 5 emissions due to Tier 2/Sulfur standards for the increased diesel sales penetration case.
Figure III-18 shows these emissions with Tier 2/Sulfur control, broken down by diesel and
gasoline.
Table 111-26. 47-State Light-Duty Direct Exhaust PM2 5 Reductions Due To Tier 2/Sulfur
Increased Diesel Growth Scenario
(Annual Tons)
Year
2004
2005
2010
2015
2020
2030
Light-Duty
Emissions Without
Tier 2/Sulfur
39,932
43,439
72,626
109,622
138,177
175,068
Light-Duty
Emissions With
Tier 2/Sulfur
19,700
20,696
22,542
23,275
24,754
28,393
Emissions
Reduced
20,232
22,743
50,084
86,347
113,423
146,675
Percent Reduction in
Baseline Inventory
Light-Duty
51%
52%
69%
79%
82%
84%
All Sources*
-
1.1%
2.4%
3.8%
4.7%
5.6%
m-37
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
180,000
150,000
120,000
90,000
60,000
30,000
WITHOUT TIER 2/SULFUR CONTROLS
2000
2005
2010
2015
2020
2025
2030
Diesel
Gasoline
Figure 111-18. 47-State Light-Duty Direct Exhaust PM2 5 Emissions With Tier 2/Sulfur
Increased Diesel Growth (Annual Tons)
In 2005, the fleet would still be comprised primarily of gasoline vehicles under this
scenario; thus, significant benefits from gasoline sulfur control would be realized immediately,
as with the no growth case. The rapid growth of diesel market penetration in conjunction with
implementation of the proposed Tier 2 PM standards would result in a diesel fleet comprised
almost exclusively of vehicles compliant with Tier 2. Thus, a large share of the baseline
inventory would be reduced very soon after implementation of the Tier 2/Sulfur standards. In
2010, nearly 70 percent of baseline light-duty exhaust PM25 inventory is reduced; by 2020, we
project 82 percent of baseline emissions would be reduced. Today's proposal would serve to
mitigate the large increases in direct PM emissions which would occur without control due to
increased growth in diesel penetration, effectively stabilizing these emissions through at least
2030.
B. Air Quality Measures
This section describes the analyses performed to evaluate the impact of the Tier 2/Sulfur
proposal on ozone and visibility levels, as discussed in Section III of the preamble. These
analyses were performed using different emission inventories, control assumptions, ozone and
visibility models, and analysis years than the air quality modeling we conducted for the
benefit/cost analysis described in Chapter VII. As a result, the ozone and visibility modeling
results presented in Section III of the preamble and described more fully in this section are not
111-38
-------
Chapter III: Environmental Impact
directly comparable to the ozone and visibility modeling results used for the benefit/cost
analysis.
This section does not discuss the impact of the Tier 2/Sulfur proposal on PM levels, since
the PM air quality modeling we performed was conducted for the benefit/cost analysis described
in Chapter VII. This section also does not discuss the impact of the Tier 2/Sulfur proposal on
CO levels since we have not quantified the proposal's impact on CO emission levels at this time.
1. Ozone
Current air quality with respect to ozone can be expressed in terms of formal designation
of attainment or nonattainment of the 1-hour standard (there is as yet no such designation for the
8-hour standard) and in terms of measured ambient design values (defined below) for both the 1-
hour and 8-hour standards. Estimates of the ozone impact of today's proposal and the expected
future ozone concentrations after its implementation can be obtained by modeling a base case
(before control measures) and a control case (after control measures). The outputs of these and
other model runs are used in combination with measured design values to project future design
values. Other metrics described in this section are also used to compare one model run to
another. The structure of this section is as follows:
Subsection B. 1 .a. presents the data supporting the discussion in the preamble of
current nonattainment status, including an explanation of design values.
Subsection B. 1 .b. describes in general terms the ozone modeling that we used to
estimate the effects of Tier 2/Sulfur controls and the way we used that modeling
to estimate the resulting design values.
Subsection B. 1 .c. explains the "rollback method," used to estimate future design
values based on measured historical ozone levels and ozone modeling results.
Subsection B. 1 .d. describes the ozone modeling simulations used to evaluate the
impact of Tier 2/Sulfur controls on future ozone levels.
Subsection B.l.e. presents the results of two ozone simulations that were used to
explore the relative effects of NOX and VOC controls on ozone levels.
Subsection B. 1 .f describes two ozone simulations that were used to estimate the
effects of today's proposal on ozone levels.
a. Measures of Current Attainment and Non-attainment
Measures of attainment and non-attainment consist of both the formal attainment and
nonattainment designations and the most recent set of ambient design values, which are based on
measurements from 1995 to 1997. Formal attainment/nonattainment status applies only to the 1-
hour standard, since such designations have not yet been made for the 8-hour standard.
Outside of California, the 1990 census showed 72 million people living in areas that were
formally designated as non-attainment for the 1-hour standard as of August 10, 1998. The
111-39
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
individual areas, their populations, and their nonattainment classifications are listed in Table C-l
in Appendix C.
Design values
An ozone design value is the concentration or average of concentrations that determines
whether a monitoring site meets the NAAQS for ozone. Because of the way they are defined,
design values can only be determined for three-year monitoring periods. We estimate the design
values, and therefore the attainment effects, resulting from control programs by using a
combination of modeling results and measured design values. Air quality model runs for a base
year and a future year are used to determine the relative change in ozone levels produced by the
controls that would be implemented between the base and future years. This relative change is
used to adjust the measured historical design values in the region being analyzed, as described in
detail below.1
A 1-hour design value is the fourth highest daily maximum 1-hour average ozone
concentration measured over a three-year period at a given monitor. An 8-hour design value is
the three-year average of the annual fourth highest daily maximum 8-hour average ozone
concentration at a given monitor. The full details of these determinations (including accounting
for missing values and other complexities) are given in Appendices H and I of 40 CFR Part 50.
As discussed in these appendices, design values are truncated to whole ppb. Due to the precision
with which the standards are expressed (0.12 ppm for the 1-hour, 0.08 ppm for the 8-hour),
nonattainment of the 1-hour standard is defined as a design value greater than or equal to 125
ppb and nonattainment of the 8-hour standard is defined as a design value greater than or equal
to 85 ppb.
For the 1-hour standard, the design value for a county is the highest design value of the
monitors within that county. Typically, there is one or zero monitors per county. If a county
does not contain an ozone monitor, it cannot have a design value. For most of our analyses,
county design values are consolidated where possible into design values for metropolitan areas.
The design value for a metropolitan area is the highest design value among the included
counties. Counties that are not in metropolitan areas are listed separately. For the purposes of
the analyses described in this section, we have assumed that the definition of county and
metropolitan area design values for the 8-hour standard will be the same as for the 1-hour
standard. It should be noted, however, that we have not yet determined how county and
metropolitan area design values will be defined for the 8-hour standard.
To simulate the air quality effects of today's proposal, design values are estimated or
projected from measured 1995-1997 county design values by the method described in Subsection
1 The procedure described in this paragraph can also be used to isolate the effects of a particular control
program on design values. To do so, one needs to apply the procedure twice: once for a future year without
applying the control program of interest, and once for a future year with the control program of interest applied.
The impact of the control program on ozone design values is given by the difference between the design values
calculated for the two different cases.
111-40
-------
Chapter III: Environmental Impact
B. 1 .c. Projected design values are determined only for counties that have measured design
values. The measured 1995-1997 design values that exceeded the 1-hour standard in
metropolitan areas and rural counties in the 37 states that participated in OTAG are shown in
Table C-2 in Appendix C. Similar measured design values for the 8-hour standard are shown in
Table C-3.
b. General Description of Ozone Modeling in the OTAG Domain to Estimate
the Effect of Tier 2/Sulfur Controls
We used the considerable development work done for the ROTR as a foundation to
estimate the impact of the proposed Tier 2/Sulfur controls on ozone levels in the OTAG domain.
A method for estimating the design values that result from a given control scenario was also
developed for the ROTR and has been extended to estimate the effects of Tier 2/Sulfur controls.
Further details of the modeling work are presented below and in a technical memorandum to
Docket A-97-10, "Photochemical Air Quality Simulations in Support of Tier 2/Sulfur," by
Harvey Michaels.
The basic method of using modeling
The modeling methodology requires running two simulations, a base case (without Tier
2/Sulfur controls) and a control case (with Tier 2/Sulfur controls). The effects of the control
program are then evaluated by comparing the modeling results of the control case with those of
the base case. The base case for our Tier 2/Sulfur ozone analysis is the 2007 post-ROTR
scenario. We used two versions of this base case; the first was published with the ROTR's
Supplemental Notice of Proposed Rulemaking (SNPR) and used the SNPR emission inventories,
while the second used the ROTR Final Inventory (September 1998), which was updated based
on public comments. Table 111-27 indicates which of our Tier 2/Sulfur simulations correspond to
which base case.
Emissions inventories and meteorology were developed for four historical ozone
episodes, each about 10 days long, from 1988, 1991, 1993, and 1995. When the photochemical
grid model was judged to satisfactorily reproduce the historical episodes, the meteorology was
retained and emission inventories for the base and control cases were substituted for the
historical emission inventories. The base and control cases were then run for all four episodes.
The model output is hourly average ozone concentrations in all grid cells of the modeling
domain for all hours of the simulation. A large number of different metrics have been developed
to compare the ozone concentrations in the base case with those in the control case. One of the
most useful of these, because of its relationship to measured design values used to determine
attainment and nonattainment, is projected design values. Design values were projected using
111-41
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
the modeling results from three episodes: 1991, 1993, and 1995.j The calculation of projected
design values is described in the next section.
To facilitate the ozone modeling, the emission reduction due to Tier 2/Sulfur controls
was expressed as a percentage reduction from the 2007 post-ROTR emission inventory for all
highway mobile sources. The procedure used to do this is described in a technical memo to
Docket No. A-97-10 ("Methodology for Developing Inventory Reductions Used in Ozone
Modeling," by John Koupal). These percentage reductions were applied everywhere in the
modeling domain to all on-highway emissions in the base case. The proposed Tier 2/Sulfur
program would achieve almost all of its emission reductions from cars and light trucks, but
converting these reductions to a percentage of all on-highway emissions greatly streamlined the
process of modeling the proposed Tier 2/Sulfur controls.
The standard ozone metrics applied to the modeling results are relatively simple and self-
explanatory. For example, "Grid Cell Days Above the Standard" is a count of all the grid cells
on all simulation days (except for 2 or 3 startup days in each episode) that the daily maximum
ozone concentration (either 1- or 8-hour average, depending on the specific metric) exceeded the
standard. The "rollback method" of projecting design values is considerably more complicated,
because it uses both measured design values and simulations. This method is described below.
c. The "Rollback Method" for Estimating Design Values Resulting from
Control Measures
Because of the way they are defined, design values can only be determined for three-year
monitoring periods. We estimate the design values resulting from a given control program by
beginning with the measured design values and then using two model runs to determine the
relative change produced by the control program. The first model run is the base year case and
uses an emissions inventory that corresponds to the measured design values. For 1995-1997
design values, we used the 1995/96 Base Year emissions inventory. The second model run is the
control case and employs the inventories for which we are projecting resulting design values.
We projected design values for three control cases: 2007 ROTR, 2007 Tier 2/Sulfur (OMS4) and
2020 Tier 2/Sulfur (OMS3). The relative change between the base year case and control case is
used to adjust the measured design values, as described in this section. This process, called the
"rollback method," was used in the ROTR rulemaking and is more fully described in the
document: "Procedures for Estimating the Impact of OTAG Strategy Run 5 on Attainment of the
J This method, and the reasons why 1991, 1993, and 1995 were used, are discussed in "Procedures for
Estimating the Impact of the OTAG Strategy Run 5 on Attainment of the 8-Hr Ozone NAAQS." Draft: October
1997. Staff Report, Air Quality Modeling Group, Emissions, Monitoring and Analysis Division, Office of Air
Quality Planning and Standards, U.S. EPA. EPA Air Docket A-96-56, II-A-24.
111-42
-------
Chapter III: Environmental Impact
8-Hr Ozone NAAQS", Staff Report, Draft October 1997, (EPA Air Docket A-96-56, II-A-24).k
There are three ozone episodes that are used in the modeling. Each corresponds to the
meteorological conditions of a historical air pollution episode and is approximately 10 days long.
The actual episodes occurred in 1991, 1993, and 1995. When we refer to running those three
episodes for the ROTR in 2007, for example, we mean using the meteorology of those three
episode with the emissions inventory we expect in 2007 following the ROTR.
The following procedure, referred to as the rollback method, was used to estimate the
effects of control strategies on 1-hr and 8-hr ozone design values. Note that, except for Step
l(a), the procedures for treating 1-hr and the 8-hr design values are the same. The base year case
refers to the 1995/96 Base Year inventory, which corresponds to the 1995-1997 period used to
determine measured design values. The control case refers to one of the three cases for which
we are projecting design values: either the 2007 post-ROTR scenario or the 2007 Tier 2/Sulfur
case or the 2020 Tier 2/Sulfur case.
Step 1: Calculate ambient design values
(a) For each monitor in a county determine the monitor-specific 1-hr design values
by taking the 4th highest daily maximum value from ozone data collected at the
monitoring site for the period 1995-1997. For determining an 8-hr design value,
calculate the 3-year average of the 4th highest daily maximum 8-hr value in each
year at the monitor.
(b) Select the highest design value from all monitors within the county as the
county-specific design value.
Step 2: Generate model predictions for three OTAG episodes (July 1991, 1993 and 1995) for the
base year case and for the control case.
(a) The base year case model predictions reflect emissions levels in the 1995-1997
time period.
(b) The control case model predictions reflect a future year control scenario.
Step 3: Calculate an adjustment factor for each grid cell
Notes:
(1) The adjustment factor is based on the percent difference in ozone predictions
between the base year case and the control case. These factors will be used in
k In keeping with Appendices H and I of 40 CFR Part 50, projected design values are truncated to whole
ppb. Nonattainment of the 1-hour standard is defined as a design value greater than or equal to 125 ppb.
Nonattainment of the 8-hour standard is defined as a design value greater than or equal to 85 ppb.
111-43
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Step 5 to "rollback" ambient design values to reflect the impacts of the control
case.
(2) Step 3 must be followed separately for the base year case and the control case.
For each grid cell:
(a) Calculate daily maximum ozone concentrations for every day simulated
(excluding the ramp up days for each episode) for the three OTAG episodes
identified in Step 2. The ramp up days are one and two for the 1993 episode and
one, two, and three for the other episodes.
(b) For each episode select the 1st, 2nd, and 3rd highest daily maximum values.
(c) For each of these "ranks" (i.e., 1st, 2nd, and 3rd ranked values), average the
concentrations across the episodes (e.g., sum all 1st ranked values and divide by
number of episodes). This yields an average value for each rank (i.e., the average
of the highest, the average of the 2nd highest, and the average of the 3rd highest
concentrations).
(d) For each of the average ranks, calculate the ratio of the control case to the base
year case. This yields a set of three ratios, one for each rank.
(e) Average these three ratios to get the Adjustment Factor, which is multiplied times
the 1995-1997 design value for a given grid cell to get the new design value for
that grid cell.
Step 4: Assign a unique grid cell adjustment factor to each individual county
(a) The cell with the largest portion of its area in the county is assigned to that
county. If more than one cell is completely contained in the county, the cell with
the highest base year case value is assigned to the county. The 1990 Base Year
OTAG model predictions were used in those cases where it was necessary to
chose among multiple grid cells for assigning a grid cell to a county.
(b) The step of assigning a unique grid cell to each county yields the county-specific
adjustment factor. Note that only one grid cell is assigned to a county. Thus,
there is no spatial averaging or spatial weighting of adjustment factors using
multiple grid cells in determining the county-specific factors.
Step 5: Calculate the rollback ambient design value
(a) This step adjusts the ambient design values in each county to reflect the ozone
reductions estimated to result from the control case.
(b) Multiply the county-specific ambient design value from Step 1 times the
111-44
-------
Chapter III: Environmental Impact
county-specific adjustment factor from Step 4.
d. Specific Simulations Used to Evaluate Tier 2/Sulfur
Four simulations were performed in support of the Tier 2/Sulfur rulemaking. They are
listed in the following table with the percentage reduction in highway vehicle emissions that
were modeled for each run.
Table 111-27. Percent Reductions for Tier 2/Sulfur Ozone Modeling Runs
Run
OMS1
OMS2
OMS3
OMS4
% Reduction from Base Case
NOX
0%
54.2%
50.2%
18.5%
VOC
30.3%
0%
10.5%
4.3%
Base Case
2007 Post-ROTR published
with SNPR
2007 Post-ROTR Final
Inventory, September 1998
OMS1 and OMS2 were intended to explore the relative effect on ozone of VOC and NOX
reductions. OMS3 was intended to model the effect of Tier 2/Sulfur in 2020, when it would
affect a large portion of the fleet. OMS4 was intended to model Tier 2/Sulfur in 2007, which is
an important year for ozone attainment and is also a year for which a large body of ROTR-
related modeling results are available.
As mentioned previously, the percentage reductions in Table 111-27 are those that when
applied to the whole highway mobile source fleet will reproduce the reductions expected to
result from Tier 2/Sulfur controls. The emissions modeling used to obtain the percent reductions
from the base case are described in a technical memo to the Docket A-97-10 from John Koupal,
titled "Methodology for Developing Inventory Reductions Used in Ozone Modeling."
e. Results of the NOx-only and VOC-only Runs (OMS1 and OMS2)
While today's proposal decreases both NOx and VOC, NOx is decreased preferentially
because it has a far greater effect on ozone. Most areas are NOx-limited—their ozone
concentrations respond more to decreases in NOX than to decreases in VOCs. Only a few, highly
localized areas are VOC-limited. For this reason, the Ozone Transport Assessment Group
reached a broad consensus that regional ozone reductions in the eastern U.S. are best
accomplished by reducing NOX. This consensus is reflected in the ROTR, which only reduces
NOY.
111-45
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
We have demonstrated that this conclusion is still valid even after the large NOx
reductions resulting from the ROTR are taken into account. The demonstration used two
OTAG-domain modeling runs that simulated the separate effects of mobile source VOC and
NOX reductions. OMS2 simulated a 54 percent reduction in highway mobile source NOX and
OMS1 a 30 percent reduction in highway mobile source VOC.1
The results of the OMS1 and OMS2 runs demonstrate that mobile source NOX reductions
are much more effective at reducing ozone than are mobile source VOC reductions. The number
of grid cell days on which the daily maximum 1-hour average ozone concentration exceeded 124
ppb fell 46% for the NOX reductions but only 2% for the VOC reductions. The number of grid
cell days on which the daily maximum 8-hour average ozone concentration exceeded 84 ppb fell
40% for the NOX reductions but only 1% for the VOC reductions.
f. Details of the Tier 2/Sulfur Ozone Modeling Runs (OMS3 and OMS4)
The results of our modeling of Tier 2/Sulfur have been summarized in the preamble. In
this section, we discuss the detailed methods and results that were not covered there. The
design-value results for all counties are listed in Appendix C.
As stated above, OMS3 and OMS4 were intended to model Tier 2/Sulfur in 2020 and
2007, respectively. For these two runs, the emission inventory for the base case, to which they
are compared, is the ROTR budget case. This base case inventory uses the OTAG nonroad,
highway heavy-duty, and highway light-duty emission estimates, as updated for the ROTR final
rule based on public comment as of September 1998. This inventory is a newer version than the
base case inventory used for the OMS1 and OMS2 runs. For both base case inventories, the
nonroad inventories are based on the NEVES study, and the highway mobile source inventories
are based on MOBILES emission factors, vehicle distributions, and mileage accumulation
patterns. The fact that these inventories do not reflect the more recent information incorporated
in the emission inventory analyses presented in Section A. of this chapter creates some
uncertainty as to the absolute values of design values and which counties are in attainment or
nonattainment under the base case.
Emissions and ozone levels were modeled for 2007 and 2020. The 2007 case is
straightforward because we produced a full emission inventory for 2007 for the ROTR. For
2020, we assumed that total emissions (under ROTR plus current vehicle standards and fleet
turnover) would be the same as in 2007, i.e., that emission reductions from fleet turnover and
emission increases from growth in all sectors balance each other. This assumption is not exactly
correct, but is close. Our best estimate is that without Tier 2/Sulfur NOx emissions from all
human sources actually would be about 3% lower in 2020, and VOC emissions would be about
5% higher. The details of these estimates are in Section A of this chapter, "Inventory Impacts of
1 For comparison, we estimate that today's proposal will actually reduce mobile source NOX 50.0 percent
and VOC 10.2 percent in 2020.
m-46
-------
Chapter III: Environmental Impact
Tier 2/Sulfur."
A relatively minor caveat is that the Tier 2/Sulfur NOX and VOC reductions were
distributed evenly over all highway mobile sources, which have a different spatial distribution
from light-duty vehicles.
After OMS3 and OMS4 had been run and design values calculated by the rollback
method, our proposal was refined, resulting in slightly different percent reductions from the base
case. These two sets of percent reductions are shown in Table 111-28.
Table 111-28. Percentage Reductions from the 2007 Post-ROTR Inventory of NOX and
NMHC for OMS3 (2020), OMS4 (2007), and for Today's Proposal in 2007, 2010, and 2020.
Year
2007
2010
2020
OMS3andOMS4
NMHC
4.3%
-
10.5%
NOX
18.5%
-
50.2%
Today 's Proposal
NMHC
4.0%
5.4%
10.2%
NOX
18.1%
26.9%
50.0%
Because OMS3 and OMS4 were so close to today's proposal, we obtained the design
values for today's proposal by linearly interpolating or extrapolating slightly based on the
differences in NOX alone. Interpolations and extrapolations were also used to determine design
values in 2010, which was not modeled, and to estimate design values in 2010 without Tier
2/Sulfur. The percentage reduction in total highway NOX emissions between the 2007 baseline
and 2010 without Tier 2/Sulfur was 4.3%.
For discussing the effects of today's proposal on one- and eight-hour design values, we
projected county design values using the rollback method for three modeling runs: 2007 post
ROTR, OMS3, and OMS4. All other projected design values have been linearly interpolated
based on NOX. As we have discussed previously, the primary effect on ozone has been produced
byNOx.
All measured and projected county design values are in Appendix C. In addition, the
preamble indicates counts of metropolitan areas and rural counties whose measured or projected
design values meet various criteria with respect to the one and eight-hour standards. Appendix C
contains the lists of these metropolitan areas and counties together with their design values and
populations.
2. Visibility/Regional Haze
m-47
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
The Northern Front Range Air Quality Study (NFRAQS) report collected numerous
ambient PM2 5 samples in various areas around Denver, including urban areas such as Welby and
rural areas such as Brighton, during the winter of 1997. The samples were analyzed for their
composition, including the contribution of carbon-based, sulfate, nitrate, and crustal matter
particles to each sample. The results of that analysis are summarized in Table 111-29.
Table 111-29. NFRAQS Compositional Analysis of PM25 Samples
Site
Welby
Brighton
Carbon-based
PM25
49%
42%
Sulfate-based
PM25
10%
15%
Nitrate-based
PM25
25%
32%
Crustal Matter
PM25
16%
11%
The study used a variety of techniques to determine how much of the carbon-based,
sulfate, and nitrate PM found in the PM2 5 samples came from gasoline vehicles. Organic tracer
compounds were used to determine how much of the carbonaceous PM25 came from gasoline
vehicles and to separate the contribution of normal emitting vehicles and higher emitting
vehicles. A combination of inventory analysis, dispersion modeling, atmospheric chemistry, and
analysis of compositional variation over time were used to determine the contribution of gasoline
vehicles to sulfate and nitrate PM2 5. The study reported the following average percentages of
sulfates and nitrates coming from gasoline vehicles. The proportion of each type of PM25
determined to come from gasoline vehicles is shown in Table 111-30.
Table 111-30. Percentage of PM2 5 Coming from Gasoline Vehicles
Site
Welby
Brighton
Carbon-Based
57%
62%
Sulfate-Based
20%
14%
Nitrate-Based
36%
38%
From these two sets of numbers, one can calculate the contribution of each type of PM25
from gasoline vehicles to total PM2 5, as shown in the middle three columns of Table 111-34. The
results can be summed to derive the contribution of gasoline vehicles to total PM25, as shown in
the last column in Table III-31.
Table 111-31. Percentage of Total PM2 5 From Gasoline Vehicles
Site
Welby
Brighton
Carbon-Based
28%
26%
Sulfate-Based
2%
2%
Nitrate-Based
9%
12%
Total
39%
40%
m-48
-------
Chapter III: Environmental Impact
This section presents the analytic basis for the preamble discussion of the impact of
mobile sources on visibility impairment in the U.S. In this context, "visibility impairment"
refers to the reduction in the distance that one can see as the result of air pollution. As discussed
in the preamble, fine particles suspended in the atmosphere are the primary cause of visibility
impairment.
As discussed in the preamble, the Grand Canyon Visibility Transport Commission
examined visibility impairment on the Colorado Plateau. Figures II-4 and II-5 in the
Commission's June 10, 1996 report titled "Recommendations for Improving Western Vistas"
contain estimates for the contribution of 11 different sources to the man-made visibility
impairment at Hopi Point. Figure II-4 is for annual average light extinction"1 and Figure II-5 for
the worst days. Each figure gives estimates for 1990, 2000, 2010, and 2040. In 2000, for both
annual average and worst days, the contribution from "Mobile" to light extinction is about 10
percent. EPA understands this category to consist of highway vehicles only, since there is a
separate category for "Non Road Diesel." Furthermore, the "Mobile" category must exclude
dust caused by highway vehicle travel since there is a separate category for "Road Dust." The
road dust category is estimated to be responsible for about 30 percent of light extinction at Hopi
Point.
It is generally recognized that the traditionally-used emission factors and transport
assumptions for road dust have considerable uncertainty. Therefore, the contribution of road
dust may be overstated in these figures. If light extinction from highway vehicles is expressed as
a percentage of all light extinction not attributable to road dust, the highway vehicle contribution
is 14 percent. Hence efforts to reduce highway vehicle emissions that cause light extinction can
contribute significantly to improved visibility on the Colorado Plateau.
The benefit/cost analysis in Chapter VII includes the visibility-related economic benefits
that would result from implementation of the Tier 2/Sulfur proposal.
C. Air Toxics
This section summarizes our analysis of the impact of the proposed Tier 2/Sulfur
standards on emissions of and exposure to air toxics. Section C. 1. reviews the effects of selected
air toxics emissions on human health. Section C.2. describes our analysis of air toxics emissions
and exposure and the effect that the proposed Tier 2/Sulfur standards may have on air toxics
emissions and exposure.
1. Health Effects
Our assessment of motor vehicle toxics focused on the following compounds with cancer
m Light extinction is a measure of visibility impairment.
111-49
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
potency estimates that have or could have significant emissions from cars and light trucks:
benzene, 1,3-butadiene, formaldehyde, acetaldehyde, and diesel PM. It should be noted,
however, that the EPA does not have an official quantitative estimate of diesel emissions
potency at present. The current estimate is still draft, as discussed below. A brief summary of
health effects information on these compounds follows. The information in this section is based
on our preliminary study of motor vehicle toxics emissions. The study will be peer reviewed in
the near future. We anticipate updating our estimates once the study completes peer review.
a. Benzene
Benzene is an aromatic hydrocarbon which is present as a gas in both exhaust and
evaporative emissions from motor vehicles. Benzene in the exhaust, expressed as a percentage
of total organic gases (TOG), varies depending on control technology (e.g., type of catalyst) and
the levels of benzene and aromatics in the fuel, but is generally about three to five percent. The
benzene fraction of evaporative emissions depends on control technology (i.e., fuel injector or
carburetor) and fuel composition (e.g., benzene level and Reid Vapor Pressure, or RVP) and is
generally about one percent.
The EPA has recently reconfirmed that benzene is a known human carcinogen by all
routes of exposure.9 Respiration is the major source of human exposure. At least half of this
exposure is by way of gasoline vapors and automotive emissions (EPA 1998a). Long-term
exposure to high levels of benzene in air has been shown to cause cancer of the tissues that form
white blood cells. Among these are acute nonlymphocytic" leukemia, chronic lymphocytic
leukemia and possibly multiple myeloma (primary malignant tumors in the bone marrow),
although the evidence for the latter has decreased with more recent studies.10'11 Leukemias,
lymphomas, and other tumor types have been observed in experimental animals that have been
exposed to benzene by inhalation or oral administration (EPA 1985, Clement 1991). Exposure
to benzene and/or its metabolites has also been linked with genetic changes in humans and
animals12 and increased proliferation of mouse bone marrow cells.13 Furthermore, the occurrence
of certain chromosomal changes in individuals with known exposure to benzene may serve as a
marker for those at risk for contracting leukemia.14
"Leukemia is a blood disease in which the white blood cells are abnormal in type or number. Leukemia
may be divided into nonlymphocytic (granulocytic) leukemias and lymphocytic leukemias. Nonlymphocytic
leukemia generally involves the types of white blood cells (leukocytes) that are involved in engulfing, killing, and
digesting bacteria and other parasites (phagocytosis) as well as releasing chemicals involved in allergic and immune
responses. This type of leukemia may also involve erythroblastic cell types (immature red blood cells).
Lymphocytic leukemia involves the lymphocyte type of white bloods cell that are responsible for the immune
responses. Both nonlymphocytic and lymphocytic leukemia may, in turn, be separated into acute (rapid and fatal)
and chronic (lingering, lasting) forms. For example; in acute myeloid leukemia (AML) there is diminished
production of normal red blood cells (erythrocytes), granulocytes, and platelets (control clotting) which leads to
death by anemia, infection, or hemorrhage. These events can be rapid. In chronic myeloid leukemia (CML) the
leukemic cells retain the ability to differentiate (i.e., be responsive to stimulatory factors) and perform function; later
there is a loss of the ability to respond.
111-50
-------
Chapter III: Environmental Impact
The latest assessment by EPA places the excess risk of developing acute nonlymphocytic
leukemia at 2.2 x 10"6 to 7.7 x 10"6/|ig/m3. In other words, there is a risk of two to eight excess
acute nonlymphocytic leukemia cases in one million people exposed to 1 |ig/m3 benzene over a
lifetime (70 years). These numbers represent the maximum likelihood (MLE) estimate of risk,
not an upper confidence limit (UCL).
A number of adverse noncancer health effects, blood disorders such as preleukemia and
aplastic anemia, have also been associated with low-dose, long-term exposure to benzene (EPA
1985, Clement 1991, 15). People with long-term exposure to benzene may experience harmful
effects on the blood-forming tissues, especially the bone marrow. These effects can disrupt
normal blood production and cause a decrease in important blood components, such as red blood
cells and blood platelets, leading to anemia (a reduction in the number of red blood cells),
leukopenia (a reduction in the number of white blood cells), or thrombocytopenia (a reduction in
the number of blood platelets, thus reducing the ability for blood to clot). Chronic inhalation
exposure to benzene in humans and animals results in pancytopenia°,a condition characterized by
decreased numbers of circulating erythrocytes (red blood cells), leukocytes (white blood cells),
and thrombocytes (blood platelets).16'17 Individuals that develop pancytopenia and have
continued exposure to benzene may develop aplastic anemia,p whereas others exhibit both
pancytopenia and bone marrow hyperplasia (excessive cell formation), a condition that may
indicate a preleukemic state.18'19 The most sensitive noncancer effect observed in humans is the
depression of absolute lymphocyte counts in the circulating blood.20 A draft reference
concentration (RfC) has been developed for benzene. The reference concentration (RfC) is an
estimate of a continuous inhalation exposure to the human population (including sensitive
subgroups) that is likely to be without an appreciable risk of deleterious noncancer effects during
a lifetime; these estimates frequently have uncertainty levels that span perhaps an order of
magnitude. The benzene RfC is 9 |ig/m3, which means that long-term exposures to benzene
should be kept below 9 |ig/m3 to avoid appreciable risks of these non-cancer effects.21
b. 1,3-Butadiene
1,3-Butadiene is formed in vehicle exhaust by the incomplete combustion of the fuel. It
is not present in vehicle evaporative and refueling emissions, because it is not present in any
appreciable amount in gasoline. 1,3-Butadiene accounts for 0.4 to 1.0 percent of total exhaust
Pancytopenia is the reduction in the number of all three major types of blood cells (erythrocytes, or red blood
cells, thrombocytes, or platelets, and leukocytes, or white blood cells). In adults, all three major types of blood cells are
produced in the bone marrow of the vertebra, sternum, ribs, and pelvis. The bone marrow contains immature cells,
known as multipotent myeloid stem cells, that later differentiate into the various mature blood cells. Pancytopenia results
from a reduction in the ability of the red bone marrow to produce adequate numbers of these mature blood cells.
p Aplastic anemia is a more severe blood disease and occurs when the bone marrow ceases to function, i.e.,
these stem cells never reach maturity. The depression in bone marrow function occurs in two stages - hyperplasia, or
increased synthesis of blood cell elements, followed by hypoplasia, or decreased synthesis. As the disease progresses, the
bone marrow decreases functioning. This myeloplastic dysplasia (formation of abnormal tissue) without acute leukemia
is known as preleukemia. The aplastic anemia can progress to AML (acute mylogenous leukemia).
m-51
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
TOG, depending on control technology and fuel composition.
EPA recently prepared a draft assessment that would determine sufficient evidence exists
to consider 1,3-butadiene a known human carcinogen.22 However, the Environmental Health
Committee of EPA's Scientific Advisory Board (SAB), in reviewing the draft document, issued
a majority opinion that 1,3-butadiene should instead be classified as a probable human
carcinogen.23 In the draft EPA assessment, the MLE estimate of a lifetime extra cancer risk from
continuous 1,3-butadiene exposure is about 3.9 x 10"6/|ig/m3. In other words, it is estimated that
approximately 4 persons in one million exposed to 1 |ig/m3 1,3-butadiene continuously for their
lifetime (85 years in this case) would develop cancer as a result of their exposure. Lower
exposures are expected to result in risks that are lower.
1,3-Butadiene also causes a variety of reproductive and developmental effects in mice
and rats (no human data) when exposed to long-term, low doses of butadiene (EPA 1998c). The
most sensitive effect was reduced litter size at birth and at weaning. These effects were observed
in studies in which male mice exposed to 1,3-butadiene were mated with unexposed females. In
humans, such an effect might manifest itself as an increased risk of spontaneous abortions,
miscarriages, still births, or very early deaths. Long-term exposures to 1,3-butadiene should be
kept below its reference concentration of 0.33 |ig/m3 to avoid appreciable risks of these
reproductive and developmental effects (EPA 1998c).
c. Formaldehyde
Formaldehyde is the most prevalent aldehyde in vehicle exhaust. It is formed from
incomplete combustion of both gasoline and diesel fuel and accounts for one to four percent of
total exhaust TOG emissions, depending on control technology and fuel composition. It is not
found in evaporative emissions.
Formaldehyde exhibits extremely complex atmospheric behavior.24 It is present in emis-
sions and is also formed by the atmospheric oxidation of virtually all organic species, including
biogenic (produced by a living organism) hydrocarbons. Mobile sources contribute both primary
formaldehyde (emitted directly from motor vehicles) and secondary formaldehyde (formed from
photooxidation of other VOCs emitted from vehicles). The mobile source contribution is
difficult to quantify, but it appears that at least 30 percent of formaldehyde in the ambient air
may be attributable to motor vehicles (EPA 1993a).
EPA has classified formaldehyde as a probable human carcinogen25 based on limited
evidence for carcinogenicity in humans and sufficient evidence of carcinogenicity in animal
studies, rats, mice, hamsters, and monkeys. Epidemiological studies in occupationally exposed
workers suggest that long-term inhalation of formaldehyde may be associated with tumors of the
nasopharyngeal cavity (generally the area at the back of the mouth near the nose), nasal cavity,
and sinus (Clement 1991, EPA 1993a). Studies in experimental animals provide sufficient
evidence that long-term inhalation exposure to formaldehyde causes an increase in the incidence
of squamous (epithelial) cell carcinomas (tumors) of the nasal cavity (Clement 1991, EPA
111-52
-------
Chapter III: Environmental Impact
1993a, EPA 1987). The distribution of nasal tumors in rats suggests that not only regional
exposure but also local tissue susceptibility may be important for the distribution of
formaldehyde-induced tumors (Clement 1991, EPA 1993a). Research has demonstrated that
formaldehyde produces mutagenic activity in cell cultures.
The MLE estimate of a lifetime extra cancer risk from continuous formaldehyde
exposure is about 1.3 x 10"6/|ig/m3. In other words, it is estimated that approximately 1 person in
one million exposed to 1 |ig/m3 formaldehyde continuously for their lifetime (70 years) would
develop cancer as a result of their exposure. Lower exposures are expected to result in risks that
are lower.
Formaldehyde exposure also causes a range of noncancer health effects. At low
concentrations (0.05-2.0 ppm), irritation of the eyes (tearing of the eyes and increased blinking)
and mucous membranes is the principal effect observed in humans. At exposure to 1-11 ppm,
other human upper respiratory effects associated with acute formaldehyde exposure include a dry
or sore throat, and a tingling sensation of the nose. Sensitive individuals may experience these
effects at lower concentrations. Forty percent of formaldehyde-producing factory workers
reported nasal symptoms such as rhinitis (inflammation of the nasal membrane), nasal
obstruction, and nasal discharge following chronic exposure.26 In persons with bronchial
asthma, the upper respiratory irritation caused by formaldehyde can precipitate an acute
asthmatic attack, sometimes at concentrations below 5 ppm;27 formaldehyde exposure may also
cause bronchial asthma-like symptoms in nonasthmatics.28'29 However, it is unclear whether
asthmatics are more sensitive than nonasthmatics to formaldehyde's effects.30
Immune stimulation may occur following formaldehyde exposure, although conclusive
evidence is not available. Also, little is known about formaldehyde's effect on the central
nervous system. Several animal inhalation studies have been conducted to assess the
developmental toxicity of formaldehyde: The only exposure-related effect noted was decreased
maternal body weight gain at the high-exposure level but no adverse effects on reproductive
outcome of the fetuses that could be attributed to treatment were noted. An inhalation reference
concentration (RfC), below which long-term exposures would not pose appreciable non-cancer
health risks, is not available for formaldehyde at this time.
d. Acetaldehyde
Acetaldehyde is a saturated aldehyde that is found in vehicle exhaust and is formed as a
result of incomplete combustion of both gasoline and diesel fuel. It is not a component of
evaporative emissions. Acetaldehyde comprises 0.4 to 1.0 percent of exhaust TOG, depending
on control technology and fuel composition.
The atmospheric chemistry of acetaldehyde is similar in many respects to that of
formaldehyde (Ligocki et al., 1991,31). Like formaldehyde, it can be both produced and
destroyed by atmospheric chemical transformation, so mobile sources contribute to ambient
acetaldehyde levels both by their primary emissions and by secondary formation resulting from
m-53
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
their VOC emissions. Data from emission inventories and atmospheric modeling indicate that
roughly 40 percent of the acetaldehyde in ambient air may be attributable to mobile sources.
Acetaldehyde emissions are classified as a probable human carcinogen. The MLE
estimate of a lifetime extra cancer risk from continuous acetaldehyde exposure is about 0.78 x
10"6/|ig/m3. In other words, it is estimated that less than 1 person in one million exposed to 1
|ig/m3 acetaldehyde continuously for their lifetime (70 years) would develop cancer as a result of
their exposure.
Non-cancer effects in studies with rats and mice showed acetaldehyde to be moderately
toxic by the inhalation, oral, and intravenous routes.32'3334 The primary acute effect of exposure
to acetaldehyde vapors is irritation of the eyes, skin, and respiratory tract. At high
concentrations, irritation and pulmonary effects can occur, which could facilitate the uptake of
other contaminants. Little research exists that addresses the effects of inhalation of acetaldehyde
on reproductive and developmental effects. The in vitro and in vivo studies provide evidence to
suggest that acetaldehyde may be the causative factor in birth defects observed in fetal alcohol
syndrome, though evidence is very limited linking these effects to inhalation exposure Long-
term exposures should be kept below the reference concentration of 9 |ig/m3 to avoid appreciable
risk of these non-cancer health effects.35
e. Diesel Particulate Matter
The paniculate matter (PM) from diesel exhaust typically consists of a solid core,
composed mainly of elemental carbon, which has a coating of various organic and inorganic
compounds. The diameter of diesel particles is very small with typically 75-95 percent of the
particle mass having a diameter smaller than 1.0 jim. The characteristically small particle size
increases the likelihood that the particles and the attached compounds will reach and lodge in the
deepest and more sensitive areas of the human lung. Both the diesel particle and the attached
compounds may be influential in contributing to a potential for human health hazard from long
term exposure.
The heavy-duty highway and off-road diesel engines, as a group, account for most of the
diesel particulate emissions currently released into ambient air.36 Diesel particulate matter is
mainly attributable to the incomplete combustion of fuel hydrocarbons, though some may be due
to engine oil or other fuel components.
In two human studies on railroad workers, and one on truckers, occupationally exposed
to diesel exhaust (EPA 1998d), it was observed that long-term inhalation of diesel exhaust
produced an excess risk of lung cancer. Taken together, the human studies show a positive
association between diesel exhaust exposure and lung cancer. Studies in experimental animals
provide additional evidence that long-term inhalation exposure to high levels of diesel
particulate may pose a significant cancer risk. Research has demonstrated that exposure to high
diesel exhaust levels causes an increase in lung tumors in two strains of rats and two strains of
mice (EPA 1998d). Also, as a result of extensive studies, the direct-acting mutagenic activity of
111-54
-------
Chapter III: Environmental Impact
both the particle and gaseous fractions of diesel exhaust has been shown (EPA 1998d).
EPA's draft Diesel Health Assessment identifies both lung cancer as well as several other
adverse respiratory health effects including respiratory tract irritation, immunological changes,
and changes in lung function, as possible concerns for long term exposure to diesel exhaust. The
evidence in both cases comes from the studies involving occupational exposures and or high
exposure animal studies mentioned above, and the Health Assessment, when completed, will
recommend how the data should be interpreted for lower environmental levels of exposure. The
draft Health Assessment is currently being revised to address comments from a peer review
panel of the Clean Air Science Advisory Committee. Based on human epidemiology studies, the
draft MLE estimate of a lifetime extra cancer risk from continuous diesel exhaust particulate
exposure ranges from 3.0 x 10-4 to 1.0 x 10-3/|ig/m3. In other words, it is estimated that
approximately 300 to 1000 persons in one million exposed to 1 |ig/m3 diesel exhaust particulate
continuously for their lifetime (70 years) would develop cancer as a result of their exposure.
The California Air Resources Board has identified diesel exhaust PM as a "toxic air
contaminant" under the state's air toxics program, based on the information available on cancer
and non-cancer health effects. California is in the process of determining the need for, and
appropriate degree of, control measures for diesel exhaust particulate matter. Note that
California limited its finding to diesel particulate matter, as opposed to diesel exhaust. EPA's
assessment activities of diesel exhaust PM are coincident with, but independent from,
California's evaluation.
Particulates (i.e, particulate matter, PM) are a prominent part of diesel exhaust and play a
role in contributing to total ambient PM, especially PM 2 5 (PM less than 2.5 jim in diameter).
This means that EPA's new National Ambient Air Quality Standard for PM2 5 provides another
health-based reference point. Diesel exhaust particles may pose a particularly serious health risk
since more than 75 percent of the particles can be less than Ijim and the smaller diesel particles
can be inhaled and deposited deeper in the lung. Diesel particles also have a large surface area
per unit mass and carry a coating of organic compounds with them which may contribute to the
health effects observed from particles. At the present time, EPA believes that for many people,
keeping long term exposures to diesel particulate matter at or below 5 |ig/m3 provides an
adequate margin of safety for the noncancer respiratory hazards.37
2. Assessment of Emissions and Exposure
In 1993, EPA released the "Motor Vehicle-Related Air Toxics Study" to meet the
requirements of Section 202(1)(1) of the Clean Air Act, which required EPA to complete a study
of the need for, and feasibility of, controlling emissions of toxic air pollutants associated with
motor vehicles and motor vehicle fuels (EPA 1993a). In 1998, EPA updated the emissions and
exposure analyses done for this study to account for new information38'39 Base scenarios for
1990, 1996, 2007, and 2020 were included in the assessment, as well as several control scenarios
in 2007 and 2020. Toxic emissions and exposure were modeled for the following urban areas:
Chicago, Denver, Houston, Minneapolis, New York, Philadelphia, Phoenix, Spokane, and St.
m-55
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Louis. Results for these urban areas were extrapolated nationwide. As mentioned previously,
EPA has assessed emissions, and exposure from the following air toxics: benzene,
formaldehyde, acetaldehyde, 1,3-butadiene, and diesel particulate matter. An assessment of the
cancer and non-cancer effects of mobile source emissions of these compounds has not yet been
completed as part of the updated analyses.
This subsection describes the analysis we have conducted to update our 1993 study.
Subsection C.2.a. discusses the emission modeling conducted for mobile source gaseous air
toxics (including both exhaust and nonexhaust air toxics) and diesel PM. Subsection C.2.b.
describes how we calculated nationwide air toxic emissions for our baseline scenario, which
assumed continuation of the National Low Emission Vehicle program indefinitely. Subsection
C.2.c. describes our analysis of air toxics exposure for our baseline scenario. Subsection C.2.d.
describes our analysis of the effects of various vehicle and fuel control scenarios on air toxics
emissions and exposure. It also describes how we used those analyses to estimate the effect of
the proposed Tier 2/Sulfur standards on air toxics emissions. This subsection also reviews our
analysis of the potential impact of increased diesel engine use in light trucks on diesel PM
emissions and exposure.
a. Emissions Modeling
/'. Gaseous Air Toxics Emissions Modeling
In these analyses, emissions of benzene, formaldehyde, acetaldehyde, and 1,3-butadiene
were estimated using a toxic emission factor model, MOBTOXSb. This model is based on a
modified version of MOBILESb, which estimates emissions of regulated pollutants, and
essentially applies toxic fractions to TOG estimates. The TOG basic emission rates used in this
modeling were similar, but not identical, to the rates used for previous modeling studies. The
model accounted for differences in toxic fractions between technology groups, driving cycles,
and normal versus high emitters. Impacts of fuel formulations were also addressed in the
modeling. Motor vehicle toxic emissions were modeled for the following urban areas: Chicago,
Denver, Houston, Minneapolis, New York, Philadelphia, Phoenix, Spokane, and St. Louis.
Exhaust Emissions
Analysis of speciation data from 1990 technology light-duty gasoline vehicles done for
the EPA Complex Model for Reformulated Gasoline showed that the fraction of toxic emissions
relative to TOG differs among eight technology groups within the Complex Model as well as
between normal emitters and high emitters.40 This difference is especially significant for 1,3-
butadiene; its TOG fraction is about three times larger for high emitters than for normal emitters.
If this difference is not taken into account, the impact of I/M programs and fleet turnover to
vehicles with lower deterioration rates will be underestimated. Thus, the input format for
exhaust toxic adjustment factors in MOBTOXSb was structured to allow input of high and
normal emitter toxic emission rates for a given "target" fuel. These toxic emission rates were
then weighted to come up with a composite toxic emission factor, based on a distribution of
111-56
-------
Chapter III: Environmental Impact
normal and high emitters. This distribution is not supplied directly by the MOBILE model.
Instead, this distribution was determined from the fleet average TOG emission rate on baseline
fuel as determined by MOBILE and average normal and high TOG emission rates on baseline
fuel derived from the Complex Model. Essentially, "toxic-TOG curves" were developed that
plot the target fuel toxic emission rate against the base fuel TOG emission rate.
To construct these curves, the distribution of normal and high emitters was determined in
the following manner for each model year. A TOG gram per mile emission rate for normal
emitters (TOG-N) and a TOG emission rate for high emitters (TOG-H) on baseline fuel were
input into MOBTOXSb. TOG-N from newer technology light-duty gasoline vehicles and trucks
were obtained from an unconsolidated version of the Complex Model, which provides output for
normal emitters in each of eight technology groups. The Complex Model provides estimates for
mass of exhaust VOC, which is TOG minus the mass of methane and ethane. TOG was
estimated by applying a conversion factor which accounts for the mass of these compounds. The
conversion factor was derived by analysis of weight percent emissions of methane and ethane
from available speciation data. Based on the distribution of technology groups in given model
year, the individual TOG estimates were weighted appropriately to obtain a composite estimate
for all normal emitters. Since the unconsolidated model's TOG-N emission rates are applicable
only to Tier 0 light duty vehicles, they had to be adjusted for Tier 1 and later vehicles. This
adjustment was performed by multiplying the unconsolidated model results by the ratio of the
emission standard for these later vehicles to the Tier 0 emission standard. TOG-H was also
obtained from the unconsolidated version of the Complex Model. TOG-H was assumed to be
the same for all Tier 0 and later vehicles.
For benzene, 1,3-butadiene, formaldehyde, and acetaldehyde, milligram per mile toxic
emission rates for normal and high emitters running on a given fuel formulation were also
entered into MOBTOXSb, using output from the unconsolidated version of the Complex Model.
An example of the data file format is provided in Table 111-32. Using the information in
the data file, an overall FTP toxic emission rate for each vehicle class in a given model year is
calculated. This overall rate takes into account the distribution of normal and high emitters by
calculating the slope and intercept of a straight line (the "toxic-TOG" curve), where the FTP
toxic emission rates for a vehicle class in a given model year are a linear function of the baseline
fuel TOG emission rate:
TOXnt Fuel AI FTP — A + B TOGBaseline &el] FTP (1)
A and B are determined as follows:
A = (TOG-H*TOX-N - TOG-N*TOX-H)/(TOG-H - TOG-N) (2)
B = (TOX-H - TOX-N)/(TOG-H - TOG-N) (3)
where:
111-57
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
TOX-N = toxic emission rate for normal emitters derived from the Complex Model
TOX-H = toxic emission rate for high emitters derived from the Complex Model
Table 111-32. Example of Data File Format for Toxic Adjustment Factors
IV
1
1
1
1
MYA
1965
1975
1981
1988
MYB
1974
1980
1987
1999
TOG-N
0.000
0.000
0.640
0.570
TOG-H
10.00
10.00
4.03
4.03
BZ-N
0.00
0.00
28.63
17.49
BZ-H
276.93
263.61
113.23
116.45
AC-N
0.00
0.00
5.07
4.02
AC-H
109.72
108.70
32.89
28.65
FR-N
0.00
0.00
7.16
5.67
FR-H
224.28
173.41
44.59
36.68
BD-N
0.00
0.00
2.14
2.04
BD-H
93.15
44.57
25.84
30.82
IV = vehicle class, MYA = initial model year, MYB = final model year, TOG-N = TOG for
normal emitters running on baseline fuel in g/mi, TOG-H = TOG for high emitters on baseline
fuel in g/mi, BZ = benzene in mg/mi for vehicles running on fuel A, AC = acetaldehyde in
mg/mi on fuel A, FR = formaldehyde in mg/mi on fuel A, BD = 1,3-butadiene in mg/mi on fuel
A
These relationships can be thought of graphically, as illustrated in Figure III-19, below.
Hypothetical Benzene-TOG Curve
140
0.5
1 1.5
Baseline Fuel TOG (g/mi)
2.5
Figure 111-19. Example Plot of Target Fuel Benzene Versus
Baseline Fuel TOG under FTP Conditions
An issue related to the above methodology is whether the linear assumption is valid for baseline
111-58
-------
Chapter III: Environmental Impact
TOG values above the high emitter point and below the normal emitter point. This is
particularly relevant in cases where A and B values are determined from Tier 0 vehicles (e.g., the
Complex model), but the results are applied to Tier 1 and LEV-category vehicles. For the simple
example presented above, negative benzene emissions are estimated for the target fuel when the
baseline fleet-average TOG emission rate falls below 0.295 g/rni. Thus, for fleet-average
emission rates below (and above) the normal (and high) emitter values, a different methodology
was needed. In those cases, it was assumed that the toxic emission rate was the same on a
fractional basis. For the example above, the benzene emission rate for a baseline TOG value of
0.1 g/mi would be calculated as follows:
BZ(TOG=0 ! g/mi) = 0.1 g/mi * (1 6 mg/mi BZ / 0. 5 g/mi TOG) = 3.2 mg/mi
This has the effect of forcing the toxic-TOG curve from the normal-emitter point back through
the origin and thus avoids negative toxic emission rate estimates for Tier 1 and LEV-category
vehicles. The same approach is used in cases where the fleet-average baseline TOG emission
rate is above the high emitter point.
For non-light duty vehicle classes and older technology light-duty vehicles, such as non-
catalyst and oxidation catalyst vehicles, adequate toxic emissions data were not available to
distinguish between emission rates of normal and high emitters. In such cases, the toxic fraction
was assumed to be constant.
Next, aggressive driving corrections were applied to the FTP toxic emission rates for
light duty vehicles. These corrections were provided in an external data file and were
multiplicative in form. Several recent studies suggest that toxic fractions of TOG differ between
FTP and aggressive driving conditions41'42'43 Thus, another adjustment to the toxic emission rates
was applied to take into account this difference in toxic fractions. This adjustment took the form
of the ratio of the toxic mass fraction over the unified cycle (FTP and off-cycle) to the toxic mass
fraction over the FTP. The adjustment was obtained from an analysis of unpublished CARB
data as described in Sierra Research et al. (Sierra 1998). The toxic emission rate under the
unified cycle (FTP and off-cycle) was calculated in the model as follows:
TO^ = TO^ * ADT * ADT (d\
1 \jy^vc 1 Wy"S.FTp fLLJ J Aggressive Driving •rt-L/ J TOX UC/FTP V V
where
TOXUC = Unified Cycle toxic emission rate
TOXFTp = FTP toxic emission rate
sive Driving = Adjustment to TOG emissions for aggressive driving
C/FTP = Adjustment for difference in toxic mass fraction over the UC versus FTP
Next, toxic emission rates were adjusted in the model to take into account air
conditioning effects on light duty vehicles. In the absence of data, we assumed that FTP-based
toxic fractions will apply to the increased TOG mass as a result of air conditioning usage. Thus,
the increase in TOG mass as a result of air conditioner usage was estimated from model year-
specific corrections for air conditioner use on TOG emissions. The corrections were provided in
111-59
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
an external data file in the model. The model calculates the increase in toxic emissions as a
result of air conditioner use as follows:
TOXA/C = TOXFTP*ADJA/C (5)
where
-, = increase in toxic emissions as a result of air conditioning usage
P = toxic emission rate under FTP conditions
. = Air conditioner usage adjustment for TOG
This result was then added to the TOXUC estimate for an overall in-use toxic emission rate.
MOBTOXSb then applies temperature, speed, humidity and load corrections.
Evaporative, Refueling, Running Loss, and Resting Loss Emissions
MOBTOXSb estimated evaporative, refueling, running loss, and resting loss toxic
emissions for benzene. (1,3-Butadiene, formaldehyde, and acetaldehyde are not found in fuel
and hence are not found in nonexhaust emissions. Because their nonexhaust emissions are zero,
they were not included in the portions of MOBTOXSb used to estimate nonexhaust emissions.)
Benzene fractions of total hydrocarbons were entered in an external data file. Separate fractions
were entered for hot soak, diurnal, refueling, running loss, and resting loss. Toxic fractions for
evaporative, refueling and running loss benzene from gasoline vehicles were obtained from the
Complex Model (EPA 1994). The Complex Model does not estimate resting loss emissions.
EPA assumed that the benzene fractions of diurnal and resting loss emissions were the same.
/'/'. Diesel PM Emissions Modeling
To estimate diesel PM emissions, we used EPA's PARTS model. PARTS is similar in
structure and function to the MOBILE series of models. It calculates exhaust and non-exhaust
(e.g., road dust) particulate emissions for each vehicle class included in the MOBILE models.
Only primary exhaust PM emission rates from diesel vehicles were included in these analyses
since cancer potencies are not available for PM emissions such as tire and brake wear or for
secondary PM formed through transformation of diesel engine emissions of SOx, NOx, and
VOC. A particle size cut-off of 10 jim was specified in the model inputs since essentially all
primary exhaust PM from diesel engines is smaller than 10 |im.
b. Nationwide Toxic Emissions Estimates - Baseline Scenario
Nationwide urban emission estimates were developed by mapping each county in the
United States to one of the modeled urban areas, based primarily on geographic considerations
(Sierra, 1998). The resulting county level emission rates were weighted by VMT estimates to
come up with average nationwide rates. Average nationwide emission rates for baseline
scenarios in 1990, 1996, 2007, and 2020 are given in Table 111-33. The baseline scenario
111-60
-------
Chapter III: Environmental Impact
assumed implementation of NLEV standards (0.09 g/mi) for light-duty gasoline vehicles and
light duty trucks under 6000 Ibs. gross vehicle weighting, Tier 1 standards for light-duty trucks
over 6000 Ibs., and a mix of conventional gasoline and Phase 2 reformulated gasoline with no
additional sulfur control.
Table 111-33. Average Nationwide Highway Vehicle Toxic Emission Rates (mg/mi)
In 1990,1996, 2007, and 2020, for Baseline Scenarios.
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3 -Butadiene
Diesel PM
CY 1990
126
19
61
17
93
CY 1996
62
14
35
9
62
CY2007
26
6
14
3
23
CY2020
16
O
8
2
17
A number of rough approximations had to be made due to the small number of cities
actually modeled. For instance, most of the South was mapped to Houston, a reformulated
gasoline area, even though most Southern cities do not require reformulated gasoline. Also, all
of California was mapped to Phoenix, which does not take into account the California LEV
program. EPA plans to perform additional modeling prior to the final rulemaking to improve the
national estimate. Despite these limitations, however, the nationwide exposure estimates should
provide reasonable approximations.
c. Exposure - Baseline Scenario
Exposure modeling was done for 1990 using the Hazardous Air Pollutant Exposure
Model for Mobile Sources, Version-3, or HAPEM-MS3.44 This model uses CO as a tracer for
toxics. Since most ambient CO comes from cars and light trucks, we believe CO exposure is an
reasonable surrogate for exposure to other motor vehicle emissions, including toxics emissions.
The HAPEM model links human activity patterns with ambient CO concentration to arrive at
average exposure estimates for 22 different demographic groups (e.g., outdoor workers, children
0 to 17, working men 18 to 44, women 65+, etc.) and for the total population. The model
simulates the movement of individuals between home and work and through a number of
different microenvironments. The CO concentration in each microenvironment is determined by
multiplying ambient concentration by a microenvironmental factor.
With the 1990 CO exposure estimates generated by HAPEM model for each urban area,
EPA determined the fraction of exposure that was a result of on-road motor vehicle emissions.
This calculation was accomplished by scaling the exposure estimates (which reflect exposure to
111-61
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
total ambient CO) by the fraction of the 1990 CO emissions inventory from on-road motor
vehicles, determined from the EPA Emission Trends database.45 Nationwide urban CO exposure
from on-road motor vehicles was estimated by first calculating a population-weighted average
CO exposure for the nine modeled areas. This number was adjusted by applying a ratio of
population-weighted annual average CO for urban areas in the entire country versus average
ambient CO concentration for the modeled areas. To estimate rural exposure, the urban estimate
was scaled downward using rough estimates of urban versus rural exposure from the 1993 Motor
Vehicle-Related Air Toxics Study (EPA 1993a).
Modeled onroad CO exposure for 1990 was divided by 1990 CO grams per mile emission
estimates to create a conversion factor. The conversion factor was applied to modeled toxic
emission estimates (in grams per mile terms) to determine exposure to onroad toxic emissions, as
shown in Equation 6:
TOX = FCO /CO 1 x TOX (6<\
Exposure(^ig/m3) L^^v^Exposure(^g/m3) EF(g/mi)J1990 EF(g/nu) V /
where TOX reflects one of the six toxic pollutants considered in this study.
The exposure estimates for calendar years 1996, 2007, and 2020 were adjusted for VMT
growth relative to 1990. 1,3-Butadiene exposure was adjusted for atmospheric transformation.
The multiplicative factors used were 0.44 for summer, 0.70 for spring and fall, and 0.96 for
winter.46 In contrast, estimated exposure to formaldehyde and acetaldehyde was based on direct
emissions. For these pollutants, removal of direct emissions in the afternoon was assumed to be
offset by secondary formation. Table 111-34 presents annual average exposure estimates for the
entire population. Estimates were also developed for outdoor workers, and children 0-17 years
of age. Exposure among outdoor workers was higher than for the entire population, and among
children it was slightly lower.
Table 111-34. Average Nationwide Highway Vehicle Toxic Exposure (jig/m3)
In 1990,1996, 2007, and 2020, for Baseline Scenarios.
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3-Butadiene
Diesel PM
CY 1990
1.35
0.20
0.66
0.14
0.99
CY 1996
0.90
0.20
0.51
0.09
0.89
CY2007
0.46
0.10
0.26
0.04
0.42
CY2020
0.36
0.08
0.20
0.04
0.40
It should be noted that recent California-EPA studies estimated a population-weighted
average outdoor diesel exhaust PM10 (particulate matter < 10 jim) exposure for 1995.47
111-62
-------
Chapter III: Environmental Impact
California also estimated indoor and total exposure concentrations for 1995. The 1995 indoor
and total air exposure concentrations were estimated to be 1.47 |ig/m3 and 1.54 |ig/m3,
respectively. This estimate compares to the estimated annual average nationwide diesel PM
1996 exposure estimate of 0.89 |ig/m3 in Table 111-37. The difference may be due to differences
in estimates of emission rates, exposure patterns, the concentration of diesel vehicle traffic, or
the spatial distribution of diesel engine emissions.
d. Impact of Potential Vehicle and Fuel Controls
The following control scenarios for 2007 and 2020 were assessed:
• base fuels and emissions with NLEV and a 40 ppm sulfur standard.
NLEV, 40 ppm sulfur, and 0.055 NMHC standard in 2004 for light duty gasoline
vehicles and trucks.
NLEV, 40 ppm sulfur, 0.055 NMHC standard, and light duty diesel trucks 50 percent of
light duty truck sales in 2004 (phased in starting in 2001)
Although none of these scenarios represent the standards actually being proposed, the
assessment shows that VOC emission reductions would reduce the health risk posed by many of
the hazardous air pollutants emitted by light-duty vehicles and trucks beyond what was projected
under baseline conditions. Estimates of the impact of VOC reductions from a 0.055 gram per
mile NMHC standard for the full useful life of the vehicle, combined with a 40 ppm sulfur
standard, on toxics emissions and exposure, are provided in Tables 111-35 through 111-39. Actual
reductions under the standards being proposed would be smaller, since the VOC emission
standards being proposed are less stringent. Under the proposed standards, VOC emissions
would be about 20 percent larger than under the 0.055 NMHC/40 ppm sulfur scenario modeled;
thus a similar difference would be expected for gaseous toxics emissions and exposure.q Table
111-39 presents gaseous toxics exposure under the proposed standards, assuming the impact on
toxics exposure is equivalent to the impact on VOC emissions.
The 1998 revision to the 1993 Motor Vehicle-Related Air Toxics Study also evaluated the
potential increase in diesel PM emissions and exposure due to increased use of diesel engines in
light trucks. Diesel engines are used in a very small portion of the cars and light-duty trucks in
service today. However, engine and vehicle manufacturers have projected that diesel engines are
likely to be used in an increasing share of light trucks. Some manufacturers have announced
capital investments to build such engines. The 1998 study evaluated the potential increase in
diesel PM emissions and exposure associated with introducing more diesel engines into the light-
duty fleet, absent any action by EPA to mitigate those risks. An extreme case was modeled,
with light duty diesel trucks accounting for 50 percent of light-duty truck sales in 2004, phased
qThe difference in toxic reductions would not be exactly the same, since fleet average toxic emissions are
affected differently by such factors the distribution of normal and high emitters, and the mix of vehicle control
technologies.
m-63
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
in starting in 2001.
The impact of such increased diesel penetration on emissions and exposure are provided
in Tables 111-35 through 111-39. Based on the exposure estimates for 2020, the potential
nationwide cancer risk from diesel particulate matter would increase by 137 percent under this
scenario. Beyond 2020, the health risks would be even greater for two reasons. First, the
proportion of light trucks equipped with diesel engines would continue to increase as the older,
gasoline-powered trucks are replaced by a mix of gasoline and diesel trucks. Second, continued
growth in the total number of miles driven would increase diesel PM emissions.
It should be noted that this increase in diesel sales is more rapid than the increase in
diesel sales analyzed for its effect on direct and secondary PM levels, which assumes that diesel
engines do not reach 50 percent of light truck sales until 2010. However, both analyses assume
that diesel engines' share of light truck sales eventually reach the same level, and the two
analyses' estimates of the total number of diesel trucks on the road tend to converge after 2010.
Under this more gradual phase-in schedule, the increase in nationwide cancer risk would be
slightly lower, about 128 percent. This estimate was developed by adjusting the estimated
potential increase in risk for the more rapid phase-in to reflect the approximately four percent
decrease in projected diesel PM emissions in 2020 that would result from the more gradual
phase-in schedule.
Under both phase-in scenarios, we have estimated that the proposed Tier 2 standards for
PM emissions from light-duty gasoline vehicles and trucks would reduce the potential increase
in diesel PM cancer risk from cars and light trucks by over 85 percent. The potential number of
cancers avoided would be even larger in future years as the proportion of diesel-powered light-
duty trucks, and the number of miles they are driven, increased.
111-64
-------
Chapter III: Environmental Impact
Table 111-35. Average Nationwide Highway Vehicle
Toxic Emission Rates (mg/mi) in 2007, for Various Scenarios
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3-Butadiene
Diesel PM
No New
Controls
Scenario
25.54
5.54
14.30
3.26
23.36
40 ppm Sulfur
Scenario
24.43
5.43
14.38
2.96
23.36
0.055NMHC
Standard w/40
ppm Sulfur
Scenario
23.44
5.29
14.04
2.86
23.36
0.055NMHC
Standard, 40
ppm Gasoline
Sulfur, & High
Diesel Sales
Scenario (50%
of 2004 Sales)
20.89
5.23
14.15
2.82
38.69
Table 111-36. Average Nationwide Highway Vehicle
Toxic Emission Rates (mg/mi) in 2020, for Various Scenarios
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3-Butadiene
Diesel PM
No New
Controls
Scenario
15.61
3.39
8.42
2.29
17.42
40 ppm Sulfur
Scenario
14.68
3.29
8.45
2.04
17.42
0.055NMHC
Standard w/40
ppm Sulfur
Scenario
11.36
2.82
7.29
1.69
17.42
0.055NMHC
Standard, 40
ppm Gasoline
Sulfur, & High
Diesel Sales
Scenario (50%
of 2004 Sales)
9.02
2.77
7.45
1.53
41.29
m-65
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table 111-37. Average Nationwide Highway Vehicle Toxic Exposures for the Entire
Population (ug/m3) in 2007, for Various Scenarios
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3-Butadiene
Diesel PM
No New
Controls
Scenario
0.46
0.10
0.26
0.044
0.42
40 ppm Sulfur
Scenario
0.44
0.10
0.26
0.040
0.42
0.055NMHC
Standard w/40
ppm Sulfur
Scenario
0.42
0.095
0.26
0.038
0.42
0.055NMHC
Standard, 40
ppm Gasoline
Sulfur, & High
Diesel Sales
Scenario (50%
of 2004 Sales)
0.37
0.094
0.26
0.038
0.70
Table 111-38. Average Nationwide Highway Vehicle Toxic Exposures for the Entire
Population (ug/m3) in 2020, for Various Scenarios
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3-Butadiene
Diesel PM
No Tier 2/Sulfur
Scenario
0.36
0.078
0.20
0.039
0.40
40 ppm Sulfur
Scenario
0.34
0.075
0.20
0.035
0.40
0.055NMHC
Standard w/40
ppm Sulfur
Scenario
0.26
0.064
0.17
0.029
0.40
0.055NMHC
Standard, 40
ppm Gasoline
Sulfur, & High
Diesel Sales
Scenario (50%
of 2004 Sales)
0.21
0.064
0.17
0.026
0.96
m-66
-------
Chapter III: Environmental Impact
Table 111-39. Average Nationwide Highway Vehicle Gaseous Toxic Exposures for the
Entire Population (jig/m3) in 2020, under proposed Tier 2 standards.
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3-Butadiene
Exposure (jug
/m3)
0.25
0.077
0.21
0.031
e.
Limitations
The analysis referenced above was conducted by Sierra Research for the Office of
Mobile Sources. It will undergo formal scientific peer review in the near future. Once that
review is complete and the peer review comments are addressed, OMS expects to conduct a
formal risk assessment on health risk of toxic emissions from mobile sources for the final rule.
111-67
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Chapter III. References
1. "Development of Nonroad, Stationary, and Area Source Emissions for the Tier 2/Sulfur
NPRM", Memorandum from Gary Dolce to Docket No. A-97-10
2. Koupal, J., "Development of Light-Duty Emission Inventory Estimates in the Notice of
Proposed Rulemaking for Tier 2 and Sulfur Standards", EPA Report No. EPA420-R-99-
005, Docket No. A-97-10
3. Ibid.
4. "A Modified Version of MOBILES for Evaluation of Proposed Tier 2 Evaporative
Emission Standards", Memorandum from David Brzezinksi to Docket No. A-97-10
5. "Development of Heavy-Duty Gasoline Emissions Inventories for the Tier 2/Sulfur
NPRM", Memorandum from John Koupal to Docket No. A-97-10
6. "Development of Motorcycle and Heavy-Duty Diesel Emissions Inventories for the Tier
2/Sulfur NPRM", Memorandum from Gary Dolce to Docket No. A-97-10
7. "Regulatory Impact Analyses for the Particulate Matter and Ozone National Ambient Air
Quality Standards and Proposed Regional Haze Rule," Innovative Strategies and
Economics Group, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, NC, July 17, 1997.
8. "Determination of Tier 2/Sulfur Emission Reductions in Terms of Equivalent Baseline
Vehicles", Memorandum from John Koupal to Docket No. A-97-10
9. EPA 1998a. Environmental Protection Agency, Carcinogenic Effects of Benzene: An
Update, National Center for Environmental Assessment, Washington, DC. 1998.
10. EPA 1985. Environmental Protection Agency, Interim quantitative cancer unit risk
estimates due to inhalation of benzene, prepared by the Office of Health and
Environmental Assessment, Carcinogen Assessment Group, Washington, DC. for the
Office of Air Quality Planning and Standards, Washington, DC., 1985.
11. Clement Associates, Inc., Motor vehicle air toxics health information, for U.S. EPA
Office of Mobile Sources, Ann Arbor, MI, September 1991.
12. International Agency for Research on Cancer, 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, p. 345-389, 1982.
13. Irons, R.D., W.S. Stillman, D.B. Colagiovanni, and V.A. Henry, Synergistic action of the
benzene metabolite hydroquinone on myelopoietic stimulating activity of
111-68
-------
Chapter III: Environmental Impact
granulocyte/macrophage colony-stimulating factor in vitro, Proc. Natl. Acad. Sci.
89:3691-3695, 1992.
14. Lumley, M., H. Barker, and J.A. Murray, Benzene in petrol, Lancet 336:1318-1319,
1990.
15. EPA 1993a. Motor Vehicle-Related Air Toxics Study, U.S. Environmental Protection
Agency, Office of Mobile Sources, Ann Arbor, MI, EPA Report No. EPA 420-R-93-005,
April 1993.
16. Aksoy, M. 1991. Hematotoxicity, leukemogenicity and carcinogenicity of chronic
exposure to benzene. In: Arinc, E.; Schenkman, J.B.; Hodgson, E., Eds. Molecular
Aspects of Monooxygenases and Bioactivation of Toxic Compounds. New York:
Plenum Press, pp. 415-434.
17. Goldstein, B.D. 1988. Benzene toxicity. Occupational medicine. State of the Art
Reviews. 3: 541-554.
18. Aksoy, M., S. Erdem, and G. Dincol. 1974. Leukemia in shoe-workers exposed
chronically to benzene. Blood 44:837.
19. Aksoy, M. and K. Erdem. 1978. A follow-up study on the mortality and the development
of leukemia in 44 pancytopenic patients associated with long-term exposure to benzene.
Blood 52: 285-292.
20. Rothman, 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. Hematotoxi city among Chinese workers heavily exposed to benzene.
Am. J. Ind. Med. 29: 236-246.
21. EPA 1998b. Environmental Protection Agency, Toxicological Review of Benzene (Non-
Cancer Effects), July 1998 draft. National Center for Environmental Assessment,
Washington, DC.
22. EPA 1998c. Environmental Protection Agency, Health Risk Assessment of
1,3-Butadiene. EPA/600/P-98/001A, February 1998
23. Scientific Advisory Board. 1998. An SAB Report: Review of the Health Risk
Assessment of 1,3-Butadiene. EPA-SAB-EHC-98, August, 1998.
24. Ligocki, M.P., G.Z. Whitten, R.R. Schulhof, M.C. Causley, and G.M. Smylie,
Atmospheric transformation of air toxics: benzene, 1,3-butadiene, and formaldehyde,
Systems Applications International, San Rafael, CA (SYSAPP-91/106), 1991.
25. EPA 1987. Environmental Protection Agency, Assessment of health risks to garment
workers and certain home residents from exposure to formaldehyde, Office of Pesticides
111-69
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
and Toxic Substances, April 1987.
26. Wilhelmsson, B. and M. Holmstrom. 1987. Positive formaldehyde PAST after prolonged
formaldehyde exposure by inhalation. The Lancet: 164.
27. Burge, P.S., M.G. Harries, W.K. Lam, I.M. O'Brien, and P.A. Patchett. 1985.
Occupational asthma due to formaldehyde. Thorax 40:225-260.
28. Hendrick, D.J., RJ. Rando, DJ. Lane, and MJ. Morris. 1982. Formaldehyde asthma:
Challenge exposure levels and fate after five years. J. Occup. Med. 24:893-897.
29. Nordman, H., H. Keskinen, and M. Tuppurainen. 1985. Formaldehyde asthma - rare or
overlooked? J. Allergy Clin. Immunol. 75:91-99.
30. EPA 199la. Environmental Protection Agency. Formaldehyde risk assessment update.
June 11, 1991. Office of Toxic Substances, U.S. Environmental Protection Agency,
Washington, DC. External review draft, June 11, 1991.
31. Ligocki, M.P., and G.Z. Whitten, Atmospheric transformation of air toxics: acetaldehyde
and polycyclic organic matter, Systems Applications International, San Rafael, CA,
(SYSAPP-91/113), 1991.
32. EPA. 1987. Health Assessment Document for Acetaldehyde — External Review Draft.
Office of Health and Environmental Assessment, Research Triangle Park, NC. Report
No. EPA 600/8-86/015A.
33. California Air Resources Board, Preliminary Draft: Proposed identification of
acetaldehyde as a toxic air contaminant, Part B Health assessment, California Air
Resources Board, Stationary Source Division, August, 1992.
34. EPA 1997b. Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office, Cincinnati, OH, 1997.
35. EPA 1991b. Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office, Cincinnati, OH.
36. EPA 1998d. Environmental Protection Agency, Health assessment document for diesel
emissions; National Center for Environmental Assessment, Washington, DC., EPA
Report No. EPA-600/8-90/057A and B, (External Review Draft), 1998.
37. EPA 1993b. Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office, Cincinnati, OH.
111-70
-------
Chapter III: Environmental Impact
38. Sierra Research, Inc., Radian International Corp., and Energy & Environmental Analysis,
Inc. 1998. Estimation of Motor Vehicle Toxic Emissions and Exposure in Selected
Urban Areas. Prepared for U. S. EPA, Office of Mobile Sources, Assessment and
Modeling Division, Ann Arbor, MI, October 15, 1998.
39. Sierra Research, Inc. 1998. On-Road Motor Vehicle National Toxics Exposure
Estimates. Memorandum from Philip Heirigs to Rich Cook, U.S. EPA. October 15,
1998.
40. EPA. 1994. Regulatory Impact Analysis for the Final Rule on Reformulated and
Conventional Gasoline, February, 1994.
41. Auto/Oil Air Quality Improvement Research Program. Technical Bulletin No. 19:
Dynamometer Study of Off-Cycle Exhaust Emissions; April, 1996.
42. Black, F.; Tejada, S.; Gurevich, M. "Alternative Fuel Motor Vehicle Tailpipe and
Evaporative Emissions Composition and Ozone Potential", J. Air & Waste Manage.
Assoc. 1998,48,578-591.
43. CARB. 1998. Unpublished data.
44. Glen, G. and Shadwick, D., "Final Technical Report on the Analysis of Carbon
Monoxide Exposure for Fourteen Cities Using HAPEM-MS3," Prepared by Mantech
Environmental Technology, Inc. for the U.S. Environmental Protection Agency, March
1998.
45. E. H. Pechan and Associates, Inc. 1997. Determination of Annual Average CO
Inventories and the Mobile Source Contribution in Selected Areas Using the 1990
OAQPS Trends Database. Prepared for U. S. EPA, Office of Mobile Sources,
Assessment and Modeling Division, September, 1997.
46. Systems Applications International. 1994. Projected Emission Trends and Exposure
Issues for 1,3-Butadiene. Prepared for the American Automobile Manufacturers
Association, March, 1994.
47. California-EPA and the California Air Resources Board, Proposed Identification of
Diesel Exhaust as a Toxic Air Contaminant, Appendix III, Part A, Exposure Assessment,
April 22, 1998.
m-71
-------
Chapter IV: Technological Feasibility
Chapter IV: Technological Feasibility
A. Feasibility of Tier 2 Exhaust Emission Standards for LDVs and LDTs
1. NMOG and NOx Emissions from Gasoline-Fueled Vehicles
Emission control technology has evolved rapidly since the passing of the CAA
Amendments of 1990. Emission standards applicable to 1990 model year vehicles required
roughly 90 percent reductions in exhaust HC and CO emissions and a 75 percent reduction in
NOx emissions compared to the uncontrolled emissions. Some vehicles currently in production
show overall reductions in these three pollutants of more than 99 percent. These vehicles
emissions are well below those necessary to meet the current federal Tier 1 and even California
LEV standards.
A number of technological advances and breakthroughs have allowed these significant
emission reductions to occur without the need for expensive, exotic equipment and fuels. For
example, ARB originally projected that many vehicles would require electrically heated catalysts
to meet their LEV program requirements. Today, no manufacturer is expected to use these
devices to comply with the LEV program requirement. EPA projected that alternative fuels,
such as methanol or natural gas, may be needed to meet these low emission levels. Today, while
vehicles using these alternative fuels are capable of meeting the California LEV requirements,
so are vehicles fueled with gasoline.
The most significant improvements which have facilitated these low emission levels have
been to traditional catalysts, which now warm up very rapidly and are substantially more durable
than past technology, and to fuel metering, which is more precise and accurate than previous
systems. Improvements have also been made to base engine designs, which have resulted in
lower engine-out emissions. Reduction of combustion chamber crevice volumes and oil
consumption are examples of improvements to base engine designs.
Table IV-1 below lists specific types of emission controls which EPA projects will be
needed in order for LDVs and LDTs to meet the proposed Tier 2 standards. It is important to
point out that the use of all of the following technologies is not necessarily required to meet the
proposed Tier 2 standards. The choices and combinations of technologies will depend on several
factors, such as current engine-out emission levels, effectiveness of existing emission control
systems, and individual manufacturer preferences. In some cases, such as projected increases in
catalyst volume and precious metal loading, EPA believes that most, if not all, LDVs and LDTs
will use the specified emission control technique.
IV-1
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table IV-1. Emission Control Hardware and Techniques
Projected to Meet Proposed Tier 2 Vehicle Standards
Emission Control Technologies
Fast Light-Off Exhaust Gas Oxygen Sensor
Universal Exhaust Gas Oxygen Sensor
Retarded Spark Timing at Start-Up
More Precise Fuel Control
Faster Microprocessor
Individual Cylinder Air-Fuel Control
Manifold with Low Thermal Capacity
Air-Assisted Fuel Injection
Injection of Air into Exhaust
Heat Optimized Exhaust Pipe
Leak-Free Exhaust System
Close-Coupled Catalyst
Improved Catalyst Washcoats
Increased Catalyst Volume and
Loading
PGM
Full Electronic Exhaust Gas Recirculation
Engine Modifications
a. Technology Description
The following descriptions provide an overview of the latest technologies capable of
reducing exhaust emissions. The descriptions will also discuss the state of development and
current production usage of the various technologies. The technology descriptions are divided
into four categories - base engine improvements, improved fuel control, improved fuel
atomization, and improved catalyst performance.
/'. Base Engine Improvements
There are several design techniques that can be used for reducing engine-out emissions,
especially for HC and NOx. The main causes of excessive engine-out emissions are unburned
HC's and high combustion temperatures for NOx. Methods for reducing engine-out HC
emissions include the reduction of crevice volumes in the combustion chamber, reducing the
combustion of lubricating oil in the combustion chamber and developing leak-free exhaust
systems. Leak-free exhaust systems are considered to be base engine improvements because any
modifications or changes made to the exhaust manifold can directly affect the design of the base
engine. Base engine control strategies for reducing NOx include the use of "fast burn"
combustion chamber designs, multiple valves with variable-valve timing, and exhaust gas
recirculation.
Combustion Chamber Design
IV-2
-------
Chapter IV: Technological Feasibility
Unburned fuel can be trapped momentarily in crevice volumes (i.e., the space between
the piston and cylinder wall) before being subsequently released. Since trapped and re-released
fuel can increase engine-out HC, the reduction of crevice volumes is beneficial to emission
performance. One way to reduce crevice volumes is to design pistons with reduced top "land
heights" (The distance between the top of the piston and the first piston ring). The reduction of
crevice volume is especially desirable for vehicles with larger displacement engines, since they
typically produce greater levels of engine-out HC than smaller displacement engines.
Another cause of excess engine-out HC emissions is the combustion of lubricating oil
that leaks into the combustion chamber, since heavier hydrocarbons in oil do not oxidize as
readily as those in gasoline. Oil in the combustion chamber can also trap gaseous HC from the
fuel and release it later unburned. In addition, some components in lubricating oil can poison the
catalyst and reduce its effectiveness. To reduce oil consumption, vehicle manufacturers will
tighten tolerances and improve the surface finishes of cylinders and pistons, improve piston ring
design and material, and improve exhaust valve stem seals to prevent excessive leakage of
lubricating oil into the combustion chamber.
As discussed above, engine-out NOx emissions result from high combustion
temperatures. Therefore, the main control strategies for reducing engine-out NOx are designed
to lower combustion temperature. The most promising techniques for reducing combustion
temperatures, and thus engine-out NOx emissions, are the combination of increasing the rate of
combustion, reducing spark advance, and adding a diluent to the air-fuel mixture, typically via
exhaust gas recirculation (EGR). The rate of combustion can be increased by using "fast burn"
combustion chamber designs. A fast burn combustion rate provides improved thermal efficiency
and a greater tolerance for dilution from EGR resulting in better fuel economy and lower NOx
emissions. There are numerous ways to design a fast burn combustion chamber. However, the
most common approach is to induce turbulence into the combustion chamber which increases the
surface area of the flame front and thereby increases the rate of combustion, and to locate the
spark plug in the center of the combustion chamber. Locating the spark plug in the center of the
combustion chamber promotes more thorough combustion and allows the ignition timing to be
retarded, decreasing the dwell time of hot gases in the combustion chamber and reducing NOx
formation. Many engine designs induce turbulence into the combustion chamber by increasing
the velocity of the incoming air-fuel mixture and having it enter the chamber in a swirling
motion (known as "swirl").
Improved EGR Design
One of the most effective means of reducing engine-out NOx emissions is exhaust gas
recirculation. By recirculating spent exhaust gases into the combustion chamber, the overall air-
fuel mixture is diluted, lowering peak combustion temperatures and reducing NOx. As
discussed above, the use of high swirl, high turbulence combustion chambers can allow the
amount of EGR to be increased from current levels of 15 to 17 percent to levels possibly as high
IV-3
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
as 20 to 25r percent, resulting in a 15 to 20 percent reduction in engine-out NOx emissions.
Many EGR systems in today's vehicles utilize a control valve that requires vacuum from
the intake manifold to regulate EGR flow. Under part-throttle operation where EGR is needed,
engine vacuum is sufficient to open the valve. However, during throttle applications near or at
wide-open throttle, engine vacuum is too low to open the EGR valve. While EGR operation
only during part-throttle driving conditions has been sufficient to control NOx emissions for
most vehicles in the past, more stringent NOx standards and emphasis on controlling off-cycle
emission levels may require more precise EGR control and additional EGR during heavy throttle
operation to reduce NOx emissions. Many manufacturers now use electronic EGR in place of
mechanical back-pressure designs. By using electronic solenoids to open and close the EGR
valve, the flow of EGR can be more precisely controlled.
While most manufacturers agree that electronic EGR gives more precise control of EGR
flow rate, not all manufacturers are using it. Numerous LEV vehicles certified for the 1998
model year still use mechanical EGR systems, and in some cases, no EGR at all. Nonetheless,
the use of EGR remains a very important tool in reducing engine-out NOx emissions, whether
mechanical or electronic.
Multiple Valves and Variable-Valve Timing
Conventional engines have two valves per cylinder, one for intake of the air-fuel mixture
and the other for exhaust of the combustion products. The duration and lift (distance the valve
head is pushed away from its seat) of valve openings is constant regardless of engine speed. As
engine speed increases, the aerodynamic resistance to pumping air in and out of the cylinder for
intake and exhaust also increases. By doubling the number of intake and exhaust valves,
pumping losses are reduced, improving the volumetric efficiency and useful power output.
In addition to gains in breathing, the multiple-valve (typically 4-valve) design allows the
spark plug to be positioned closer to the center of the combustion chamber (as discussed above)
which decreases the distance the flame must travel inside the chamber. In addition, the two
streams of incoming gas can be used to achieve greater mixing of air and fuel, further increasing
combustion efficiency which lowers engine-out HC emissions.
Even greater improvements to combustion efficiency can be realized by using valve
timing and lift control to take advantage of the 4-valve configuration. Conventional engines
utilize fixed-valve timing and lift across all engine speeds. Typically the valve timing is set at a
level that is a compromise between low speed torque and high engine speed horsepower. At light
engine loads it would be desirable to close the intake valve earlier to reduce pumping losses.
r Some manufacturers have stated that EGR impacts the ability to control net air-fuel ratios tightly due to
dynamic changes in exhaust back pressure and temperature, and that the advantages of increasing EGR flow rates
are lost partly in losses in air-fuel ratio control even with electronic control of EGR. Higher EGR flow rates can be
tolerated by modern engines with more advanced combustion chambers, but EGR cooling may be necessary to
achieve higher EGR flow rates within acceptable detonation limits without significant loss of air-fuel control.
IV-4
-------
Chapter IV: Technological Feasibility
Variable valve timing can enhance both low speed torque and high speed horsepower with no
necessary compromise between the two. Variable valve timing can allow for increased swirl and
intake charge velocity, especially during low load operating conditions where sufficient swirl
and turbulence tend to be lacking. By providing a strong swirl formation in the combustion
chamber, the air-fuel mixture can mix sufficiently, resulting in a faster, more complete
combustion, even under lean air-fuel conditions, thereby reducing emissions. Variable valve
technology by itself may have somewhat limited effect on reducing emissions. Several vehicle
manufacturers estimated emission reductions of 3 percent-10 percent for both NMHC and NOx,
but reductions could be increased when variable valve timing is combined with optimized spark
plug location and additional EGR.
Multi-valve engines already exist in numerous federal and California certified vehicles
and are projected by ARE to become even more common. ARB also projects that in order to
meet LEV II and ULEVII standards, more vehicles will have to make improvements to the
induction system, including the use of variable valve timing.
Leak-Free Exhaust System
Leaks in the exhaust system can result in increased emissions, but not necessarily from
emissions escaping from the exhaust leak to the atmosphere. With an exhaust system leak,
ambient air is typically sucked into the exhaust system by the pressure difference created by the
flowing exhaust gases inside the exhaust pipe. The air that is sucked into the exhaust system is
unmetered and, therefore, unaccounted for in the fuel system's closed-loop feedback control.
The excess air in the exhaust causes the computer to increase fuel to the engine, resulting in
erratic and/or overly rich fuel control. This results in increased emission levels and potentially
poor driveability. In addition, an air leak can cause an oxidation environment to exist in a three-
way catalyst at low speeds that would hamper reduction of NOx and lead to increased NOx
emissions.
Some vehicles currently use leak-free exhaust systems today. These systems consist of an
improved exhaust manifold/exhaust pipe interface plus a corrosion-free flexible coupling
inserted between the exhaust manifold flange and the catalyst to reduce stress and the tendency
for leakage to occur at the joint. In addition, improvements to the welding process for catalytic
converter canning could ensure less air leakage into the converter and further reduce emissions.
ARB and MECA project that vehicle manufacturers will continue to incorporate leak-free
exhaust systems as emission standards become more stringent.
/'/'. Improvements in Air-Fuel Ratio Control
Modern three-way catalysts require the air-fuel ratio (A/F) to be as close to stoichiometry
(the amount of air and fuel just sufficient for nearly complete combustion) as possible. This is
because three-way catalysts simultaneously oxidize HC and CO, and reduce NOx. Since HC and
CO are oxidized during A/F operation slightly lean of stoichiometry, while NOx is reduced
during operation slightly rich of stoichiometry, there exists a very small A/F window of
IV-5
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
operation around stoichiometry where catalyst conversion efficiency is maximized for all three
pollutants (i.e., less than 1 percent deviation in A/F or roughly ± 0.15). Contemporary vehicles
have been able to maintain stoichiometry, or very close to it, by using closed-loop feedback fuel
control systems. At the heart of these systems has been a single heated exhaust gas oxygen
(HEGO) sensor. The HEGO sensor continuously switches between rich and lean readings. By
maintaining an equal number of rich readings with lean readings over a given period, and by
limiting the degree to which the exhaust is rich or lean at any point in time, the fuel control
system is able to maintain stoichiometry. While this fuel control system is capable of
maintaining the A/F with the required accuracy under steady-state operating conditions, the
system accuracy is challenged during transient operation where rapidly changing throttle
conditions occur. Also, as the sensor ages, its accuracy decreases.
Dual Oxygen Sensors
Many vehicle manufacturers have placed a second FIEGO sensor(s) downstream of one
or more catalysts in the exhaust system as a method for monitoring the catalyst effectiveness of
the federally and California mandated on-board diagnostic (OBD II) system. In addition to
monitoring the effectiveness of the catalyst, the downstream sensors can also be used to monitor
the primary control sensor and adjust for deterioration, thereby maintaining precise A/F control
at higher mileages. Should the front primary HEGO sensor, which operates in a higher
temperature environment, begin to exhibit slow response or drift from its calibration point, the
secondary downstream sensor can be relied upon for modifying the fuel system controls to
compensate for the aging effects. By placing the second sensor further downstream from the hot
engine exhaust, where it is also less susceptible to poisoning, the rear sensor is less susceptible to
aging over the life of the vehicle. As a result, the use of a dual oxygen sensor fuel control
system can ensure more robust and precise fuel control, resulting in lower emissions.
Currently, all vehicle manufacturers use a dual oxygen sensor system for monitoring the
catalyst as part of the OBD II system. As discussed above, most manufacturers also utilize the
secondary HEGO sensor for trim (i.e., adjustments to) of the fuel control system. It is
anticipated that all manufacturers will soon use the secondary sensor for fuel trim.
Universal Oxygen Sensors
The universal exhaust gas oxygen (UEGO) sensor, also called a "linear oxygen sensor",
could replace conventional HEGO sensors. Conventional HEGO sensors only determine if an
engine's A/F is richer or leaner than stoichiometric, providing no indication of the exact level of
the A/F. In contrast, UEGO's are capable of recognizing both the direction and magnitude of
A/F transients since the voltage output of the UEGO is "proportional" with changing A/F (i.e.,
each voltage value corresponds to a certain A/F). Therefore, proportional A/F control is possible
with the use of UEGO sensors, facilitating faster response of the fuel feedback control system
and tighter control of A/F.
Although some manufacturers are currently using UEGO sensors, discussions with
various manufacturers suggest that some manufacturers are of mixed opinion as to the future
IV-6
-------
Chapter IV: Technological Feasibility
applicability of UEGO sensors. Because of their high cost, manufacturers claim that it may be
cheaper to improve HEGO technology rather than utilize UEGO sensors. An example of this is
the use of a "planar" design for HEGO sensors. Planar HEGO sensors (also known as "fast
light-off HEGO sensors) have a thimble design that is considerably lighter than conventional
designs. The main benefits are shorter heat-up time and faster sensor response.
Individual Cylinder A/F Control
Another method for tightening fuel control is to control the A/F in each individual
cylinder. Current fuel control systems control the A/F for the entire engine or a bank of
cylinders. By controlling A/F for the entire engine or a bank of cylinders, any necessary
adjustments made to fuel delivery for the engine are applied to all cylinders simultaneously,
regardless of whether all cylinders need the adjustment. For example, there is usually some
deviation in A/F between cylinders. If a particular cylinder is rich, but the "bulk" A/F indication
for the engine is lean, the fuel control system will simultaneously increase the amount of fuel
delivered to all of the cylinders, including the rich cylinder. Thus, the rich cylinder becomes
even richer having a potentially negative effect on the net A/F.
Individual cylinder A/F control helps diminish variation among individual cylinders.
This is accomplished by modeling the behavior of the exhaust gases in the exhaust manifold and
using sophisticated software algorithms to predict individual cylinder A/F. Individual cylinder
A/F control requires use of an UEGO sensor in lieu of the traditional HEGO sensor, and requires
a more powerful engine control computer.
Adaptive Fuel Control Systems
The fuel control systems of virtually all current vehicles incorporate a feature known as
"adaptive memory" or "adaptive block learn." Adaptive fuel control systems automatically
adjust the amount of fuel delivered to compensate for component tolerances, component wear,
varying environmental conditions, varying fuel compositions, etc., to more closely maintain
proper fuel control under various operating conditions.
For most fuel control systems in use today, the adaption process affects only steady-state
operation conditions (i.e., constant or slowly changing throttle conditions). Because transient
operating conditions have always provided a challenge to maintaining precise fuel control, the
use of adaptive fuel control for transient operation would be extremely valuable. Accurate fuel
control during transient driving conditions has traditionally been difficult because of
inaccuracies in predicting the air and fuel flow under rapidly changing throttle conditions. Air
and fuel dynamics within the intake manifold (fuel evaporation and air flow behavior), and the
time delay between measurement of air flow and the injection of the calculated fuel mass, result
in temporarily lean A/F during transient operation. Variation in fuel properties, particularly
distillation characteristics, also increases the difficulty in predicting A/F during transients. These
can all lead to poor driveability and an increase in NOx emissions.
IV-7
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Electronic Throttle Control Systems
As mentioned above, the time delay between the air mass measurement and the
calculated fuel delivery presents one of the primary difficulties in maintaining accurate fuel
control and good driveability during transient driving conditions. With the conventional
mechanical throttle system (a metal linkage connected from the accelerator pedal to the throttle
blade in the throttle body), quick throttle openings can result in a lean A/F spike in the
combustion chamber. Although algorithms can be developed to model air and fuel flow
dynamics to compensate for these time delay effects, the use of an electronic throttle control
system, known as "drive-by-wire" or "throttle-by-wire," may better synchronize the air and fuel
flow to achieve proper fueling during transients (e.g., the driver moves the throttle, but the fuel
delivery is momentarily delayed to match the inertial lag of the increased airflow).
While this technology is currently used on several vehicle models, it is considered
expensive and those vehicles equipped with the feature are expensive, higher end vehicles.
Because of its high cost, it is not anticipated that drive-by-wire technology will become
commonplace in the near future.
Hi. Improvements in Fuel Atomization
In addition to maintaining a stoichiometric A/F ratio, it is also important that a
homogeneous air-fuel mixture be delivered at the proper time and that the mixture is finely
atomized to provide the best combustion characteristics and lowest emissions. Poorly prepared
air-fuel mixtures, especially after a cold start and during the warm-up phase of the engine, result
in significantly higher emissions of unburned HC since combustion of the mixture is less
complete. By providing better fuel atomization, more efficient combustion can be attained,
which should aid in improving fuel economy and reducing emissions. Sequential multi-point
fuel injection and air-assisted fuel injectors are examples of the most promising technologies
available for improving fuel atomization.
Sequential Multi-Point
Typically, conventional multi-point fuel injection systems inject fuel into the intake
manifold by injector pairs. This means that rather than injecting fuel into each individual
cylinder, a pair of injectors (or even a whole bank of injectors) fires simultaneously, sending fuel
into several cylinders. Since only one of the cylinders is actually ready for fuel at the moment of
injection, the other cylinder(s) gets too much or too little fuel. With this less than optimum fuel
injection timing, fuel puddling and intake manifold wall wetting can occur, both of which can
hinder complete combustion. Sequential injection, on the other hand, delivers a more precise
amount of fuel that is required by each cylinder to each cylinder at the appropriate time.
Because of the emission reductions and other performance benefits "timed" fuel injection offers,
sequential fuel injection systems are very common on today's vehicles and are expected to be
incorporated in all vehicles soon.
Air-Assisted Fuel Injectors
IV-8
-------
Chapter IV: Technological Feasibility
Another method to further homogenize the air-fuel mixture is through the use of air-
assisted fuel injection. By injecting high pressure air into the fuel injector, and subsequently, the
fuel spray, greater atomization of the fuel droplets can occur. Since achieving good fuel
atomization is difficult when the air flow into the engine is low, air-assisted fuel injection can be
particularly beneficial in reducing emissions at low engine speeds. In addition, industry studies
have shown that the short burst of additional fuel needed for responsive, smooth transient
maneuvers can be reduced significantly with air-assisted fuel injection due to a decrease in wall
wetting in the intake manifold.
iv. Improvements to Exhaust Aftertreatment Systems
Over the last five years or so, there have been tremendous advancements in exhaust
aftertreatment systems. Catalyst manufacturers are progressively moving to palladium as the
main precious metal in automotive catalyst applications. Improvements to catalyst thermal
stability and washcoat technologies, the design of higher cell densities, and the use of two-layer
washcoat applications are just some of the advancements made to catalyst technology. There has
also been much development in HC and NOx absorber technology. The advancements to
exhaust aftertreatment systems are probably the single most important area of emission control
development.
Catalysts
As previously mentioned, significant changes in catalyst formulation, size and design
have been made in recent years and additional advances in these areas are still possible.
Palladium (Pd) is likely to continue as the precious metal of choice for close-coupled
applications and will start to see more use in underfloor applications. Some manufacturers, for
example, have suggested that they will use Pd/Rh in lieu of tri-metal or conventional Pt/Rh
catalysts for underfloor applications. Palladium catalysts, however, are less resistant to poisoning
by oil-and fuel-based additives than conventional platinum/rhodium (Pt/Rh) catalysts. Based on
current certification trends and information from vehicle manufacturers and catalyst suppliers, it
is expected that Pd-only and Pd/Rh catalysts will be used in the close-coupled locations while
conventional or tri-metal (Pd/Pt/Rh) catalysts will continue to be used in underfloor applications.
As palladium technology continues to improve, it may be possible for a single close-coupled
catalyst to replace both catalysts. If fact, at least one vehicle manufacturer currently uses a
single Pd-only catalyst for one of their models. According to MECA, new Pd-based catalysts are
now capable of withstanding exposure to temperatures as high as 1100°C and, as a result, can be
moved very close to the exhaust manifold to enhance catalyst light-off performance.
In addition to reliance on Pd and tri-metal applications, catalyst manufacturers have
developed "multi-layered" washcoat technologies. Automotive catalysts consist of a cylindrical
or oval shaped substrate, typically made of ceramic or metal. The substrate is made up of
hundreds of very small, but long cells configured in a shape similar to a honey-comb. The
substrate is coated with a substance containing precious metals, rare earth metals, and base-
metal oxides, that is known as the catalyst washcoat. Typical washcoat formulations consist of
IV-9
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
precious metals which either oxidize or reduce pollutants, base-metal oxides, such as alumina,
which provide the surface area support for the precious metals to adhere to, and base components
(rare earth metals) such as lanthanum, ceria, and zirconia, which act as promoters, stabilizers,
and encourage storage and reduction of oxygen. Conventional catalysts have a single layer of
washcoat and precious metals applied to the catalyst substrate. More advanced catalysts use
multi-layered washcoats with two or more layers of different combinations of washcoat and
precious metals. The washcoat can be applied to the substrate such that one layer can be applied
on top of another. The use of multi-layered washcoat technology allows precious metals that
have adverse reactions together to be separated such that catalyst durability and emission
reduction performance are significantly enhanced. For example, Pd and Rh can have adverse
reactions when combined together in a single washcoat formulation. A multi-layer washcoat
architecture that uses Pd and Rh could have the Pd on the bottom layer and the Rh on the top
layer. Rh is particularly used at reducing NOx. It is generally preferable to reduce NOx in the
top layer while CO and HC are still present and then oxidize CO and HC in the bottom layer.
Figure IV-1 illustrates the impact coating architecture (multi-layered washcoat technology) can
have on emission performance.
SAE 960802: 1.8 liter 4 cyl; 100 h aged; Pd/Rh=5/l @ 50 g/cu. ft.
Single layer Pd/Rh
layer - Pdtop
layer - Pd bottom
THC
NOx
Figure IV-1. Impact of Coating Architecture on HC and NOx Emissions.
Manufacturers have also been developing catalysts with substrates which utilize thinner
walls in order to design higher cell density, low thermal mass catalysts for close-coupled
applications (improves mass transfer at high engine loads and increases catalyst surface area as
well as speeding up light-off during cold starts). The greater the number of cells there are, the
more surface area that exists for washcoat components and precious metals to adhere to,
resulting in more precious metal sites available for oxidizing and reducing pollutants. Cell
densities of 600 cells per square inch (cpsi) have already been commercialized, and research on
900 cpsi catalysts has been progressing. Typical cell densities for conventional catalysts are 400
cpsi.
IV-10
-------
Chapter IV: Technological Feasibility
We have projected that in order to meet the proposed Tier 2 emission standards catalyst
volumes will increase. Current California LEV and ULEV passenger car catalyst volume to
engine displacement ratios are approximately 0.7 to over 1.0 while many trucks only have ratios
of 0.6 or less. We believe that in order to comply with proposed Tier 2 standards, most vehicle
will likely need catalyst volumes equal to the displacement of the engine, or in some cases, even
greater. As mentioned above, higher cell density substrates effectively provide more surface
area for pollutant conversion, therefore catalyst volumes may not need to be increased as
significantly if higher cell density substrates are used.
We have also projected that some level of increased catalyst loading will be necessary to
meet proposed Tier 2 standards. Typical catalyst loading for current LEVs and ULEVs range
from 50 g/cu ft to 300 g/cu ft. We believe that, based on input from catalyst suppliers and
vehicle manufacturers, depending on the vehicle, catalysts meeting Tier 2 standards will need
loadings in the 100 - 300 g/cu ft range. However, catalyst suppliers have also indicated to us that
they and vehicle manufacturers are constantly working on ways to reduce the amount of precious
metal loading ( a process they refer to as "thrifting"). Thrifting is achieved in several ways. One
of the most common is matching the catalyst to the attributes of the vehicle. By working in
unison, vehicle manufacturers and catalyst suppliers are able to thrift or reduce the amount of
precious metal used in a given application by attempting to optimize the vehicle fuel control
strategy, exhaust mass flow rate, and exhaust temperature with various catalyst parameters, such
as catalyst location, substrate design, cell density, oxygen storage capability, and precious metal
and base metal dispersion, to name a few. Other methods of thrifting are the constant
improvements being made to washcoat architecture - that is, constant improvement to the
materials used in the washcoat formulation so that the precious metals and other components
better adhere to the substrate surface. Finally, improvements to washcoat application processes
can also significantly improve catalyst performance while allowing thrifting of precious metals.
Improvements to processes consist of advancements to the process used to coat the substrate
with washcoat materials - allowing precious metals, base metals, and ceria to better dispersed.
Better dispersion means that rather than relatively large "clumps" of precious metals unevenly
dispersed throughout the catalyst surface, many smaller precious metal sites are dispersed
uniformly throughout the catalyst surface increasing the chance for pollutants to come into
contact with the precious metal and react into a harmless emission. Therefore, as thrifting
continues, it is possible that precious metal loading may actually decrease rather than increase.
The largest source of HC emissions continues to be cold start operation where the
combination of rich A/F operation and the ineffectiveness of a still relatively cool catalyst result
in excess HC emissions. One of the most effective strategies for controlling cold start HC
emissions is to reduce the time it takes to increase the operating temperature of the catalyst
immediately following engine start-up. The effectiveness or efficiency of the catalyst increases
as the catalyst temperature increases. One common strategy is to move the catalyst closer to the
exhaust manifold where the exhaust temperature is greater (e.g., a close-coupled catalyst). In
addition to locating the catalyst closer to the engine, retarding the spark timing and increasing
idle speed are other possible approaches. Retarding spark timing, causes combustion to occur
later in the power stroke, allowing more heat to escape into the exhaust manifold during the
exhaust stroke. Increased idle speed leads to a greater amount of combustion per unit time,
IV-11
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
providing a greater quantity of heat for heating the exhaust manifold, headpipe, and catalyst.
Another strategy is to use an electrically-heated catalyst (EHC). The EHC consists of a
small electrically heated catalyst placed directly in front of a conventional catalyst. Both
substrates are located in a single can or container. The EHC is powered by the alternator, or
solely from the vehicle's battery, or from a combination of the alternator and battery. The EHC
is capable of heating up almost immediately, assisting the catalyst that directly follows it to also
heat up and obtain light-off temperature (e.g., the catalyst temperature where catalyst efficiency
is 50 percent) quickly. Manufacturers have indicated that EHC's will probably only be
necessary for a limited number of LEV II/ULEV II engine families, mostly larger displacement
V-8's where cold start emissions are difficult to control.
Adsorbers/Traps
Other potential exhaust aftertreatment systems that are used in conjunction with a catalyst
or catalysts, are the HC and NOx adsorbers or traps. Hydrocarbon adsorbers are designed to trap
HC while the catalyst is cold and unable to sufficiently convert the HC. They accomplish this by
utilizing an adsorbing material which holds onto the HC. Once the catalyst is warmed up, the
trapped HC are automatically released from the adsorption material and are converted by the
fully functioning downstream three-way catalyst. There are three principal methods for
incorporating an adsorber into the exhaust system. The first is to coat the adsorber directly on
the catalyst substrate. The advantage is that there are no changes to the exhaust system required,
but the desorption process cannot be easily controlled and usually occurs before the catalyst has
reached light-off temperature. The second method locates the adsorber in another exhaust pipe
parallel with the main exhaust pipe, but in front of the catalyst and includes a series of valves
that route the exhaust through the adsorber in the first few seconds after cold start, switching
exhaust flow through the catalyst thereafter. Under this system, mechanisms to purge the
adsorber are also required. The third method places the trap at the end of the exhaust system, in
another exhaust pipe parallel to the muffler, because of the low thermal tolerance of adsorber
material. Again a purging mechanism is required to purge the adsorbed HC back into the
catalyst, but adsorber overheating is avoided.
NOx adsorbers are also being developed, but according to MECA, are generally
recognized as a control for NOx resulting from reduced EGR. They are typically used for lean-
burn applications and are not applicable to engines that attempt to maintain stoichiometry all the
time.
Secondary Air Injection
Secondary injection of air into exhaust ports after cold start (e.g., the first 40-60 seconds)
when the engine is operating rich, coupled with spark retard, can promote combustion of
unburned HC and CO in the exhaust manifold and increase the warm-up rate of the catalyst. By
means of an electrical pump, secondary air is injected into the exhaust system, preferably in
close proximity of the exhaust valve. Together with the oxygen of the secondary air and the hot
exhaust components of HC and CO, oxidation ahead of the catalyst can bring about an efficient
IV-12
-------
Chapter IV: Technological Feasibility
increase in the exhaust temperature which helps the catalyst to heat up quicker. The exothermic
reaction that occurs is dependent on several parameters (secondary air mass, location of
secondary air injection, engine A/F ratio, engine air mass, ignition timing, manifold and
headpipe construction, etc.), and ensuring reproducibility demands detailed individual
application for each vehicle or engine design.
Insulated or Dual Wall Exhaust System
Insulating the exhaust system is another method of furnishing heat to the catalyst to
decrease light-off time. Similar to close-coupled catalysts, the principle behind insulating the
exhaust system is to conserve heat generated in the engine to aid the catalyst warm-up. Through
the use of laminated thin-wall exhaust pipes, less heat will be lost in the exhaust system,
enabling quicker catalyst light-off
v. Improvements in Engine Calibration Techniques
Of all the technologies discussed above, one of the most important emission control
strategies is not hardware-related. Rather, it is software related and, more specifically, involves
the algorithms and calibrations contained within the software that are used in the power-train
control module (PCM) which control how the various engine and emission control components
and systems operate. Advancements in software along with refinements to existing algorithms
and calibrations can have a major impact in reducing emissions. Confidential discussions
between manufacturers and EPA have suggested that manufacturers believe emissions can be
further reduced by improving and updating their calibration techniques. As computer
technology and software continues to advance, so does the ability of the automotive engineer to
use these advancements in ways to better optimize the emission control systems. For example,
as processors become faster, it is possible to perform calculations quicker, thus allowing for
faster response times for controlling engine parameters, such as fuel rate and spark timing. As
the PCM becomes more powerful with greater memory capability, algorithms can become more
sophisticated. Manufacturers have found that as computer processors, engine control sensors
and actuators, and computer software become more advanced, and, in conjunction with their
growing experience with developing calibrations, as time passes, their calibration skills will
continue to become more refined and robust, resulting in even lower emissions.
Manufacturers have suggested to EPA that perhaps the single most effective method for
controlling NOx emissions will be tighter A/F control which could be accomplished with
advancements in calibration techniques without necessarily having to use advanced technologies,
such as UEGO sensors. Manufacturers have found ways to improve calibration strategies such
that meeting federal cold CO requirements, as well as, complying with LEV standards, have not
required the use of advanced hardware, such as EHCs or adsorbers.
Since emission control calibrations are typically confidential, it is difficult to predict
what advancements will occur in the future, but it is clear that improved calibration techniques
and strategies are a very important and viable method for further reducing emissions.
IV-13
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
b. Data Supporting Tier 2 Technical Feasibility
Automobile manufacturers generally design vehicles to meet emission targets which are
50-70 percent of the emission standards, even after the catalytic converter has been aged to full
useful life (e.g., 100,000-120,000 miles). The manufacturer desires this 30-50 percent safety
margin in order to reduce the probability that in-use vehicles will exceed the standard to an
acceptable level. Thus, the emission design targets for proposed Tier 2 standards for full useful
life would be 0.045-0.063 g/mi NMOG and 0.035-0.049 g/mi NOx at full useful life.
With this in mind, we will present data from several sources that establish our proposed
Tier 2 standards to be feasible. The data ranges from certification emission levels to feasibility
evaluation programs undertaken in the last year by ARB and MECA. Even though these
programs were directed towards the LEV II program, the data and information resulting from
these programs are useful to EPA in establishing feasibility of Tier 2 emission standards since
our proposed Tier 2 standards are the same as the LEV II standards.
/'. Certification Emission Levels
Manufacturers report certification results for engine families. Those engine families are
used in a variety of vehicle models and configurations. Manufacturers are required to report
certification test results for at least two vehicle configurations and often report results for five or
six or more models or configurations within an engine family. Manufacturers, for example, will
report certification test results for both automatic and manual transmission versions of a vehicle.
Table IV-2 below indicates the number of engine families with at least one vehicle configuration
at or below full-life NOx levels of 0.04, 0.07, and 0.1 g/mile. Of those at or below 0.04 and 0.07
g/mile NOx, 16 and 35, respectively, also have HC levels below 0.09 g/mile. There are
approximately 400 engine families total.
Table IV-2. Number of Engine Families with One or More Engine/Vehicle Configurations
with Low Full-life NOx Levels
NOx level
< 0.04
< 0.07
< 0.1
Vehicles Below 6,000 pounds
(LDVs, LDTls, LDT2s)
20
45
150
Vehicles Above 6,000 pounds
(LDT3s, LDT4s)
2
O
11
Table IV-3 provides a listing of engine families with one or more vehicle configurations
at or below 0.07 g/mile NOx. The table also provides the HC certification levels for those
configurations. Where a range is shown, there is more than one configuration within the engine
IV-14
-------
Chapter IV: Technological Feasibility
family with full-life NOx certification levels at or below 0.07. The same vehicle models appear
in the table more than once because multiple engine families are often certified for the same
vehicle models. EPA assembled this list by reviewing 1999 model year certification data for
engine families certified to nationwide Tier 1 standards, NLEV program standards, and the
California program standards.
IV-15
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table IV-3. 1999 Engine Families with One or More Vehicle Configurations
with Full-life NOx Certification Levels at or below 0.07 g/mile NOx
Manufacturer
Models
NOx level
HC level
Standard
LDVs (passenger cars)
Hyundai
Ford
Ford
Volkswagen, Audi
Volvo
Volvo
Hyundai
Daimler Chrysler
Mitsubishi
Mitsubishi
Suzuki
Ford
Ford
Daimler Chrysler
Hyundai
Volkswagen
Nissan
Ford
Ford
Ford
Daimler Chrysler
Daimler Chrysler
Hyundai
Hyundai
Volkswagen
Mazda
Elantra Wagon, Tiburon
Contour, Mystique, Cougar
Contour, Mystique, Cougar
Passat, Passat wagon
A4
V70, S70
S70, V70, C70
Elantra, Tiburon
Cirrus, Stratus, Breeze
Diamante
Gallant, Mirage
Metro
Mustang
Contour, Mystique, Cougar
S320
Sonata
Jetta, Golf, Cabriolet
Altima
Sable, Taurus
Mustang
Contour, Mystique, Cougar
E430, SL500
SL600
Accent
Sonata
New Beetle, New Golf, New Jetta
MX-5 Miata
0.01-0.02
0.02-0.05
0.02*
0.03 - 0.07*
0.03
0.03 - 0.04
0.03-0.04*+
0.04*+
0.04*+
0.04 +
0.04
0.04
0.04 - 0.05
0.04 +
0.04-0.06*
0.04 - 0.06
0.05
0.05-0.06
0.05-0.06
0.06
0.06 - 0.07
0.06
0.06*
0.06*
0.06*
0.07
0.05
0.12-0.17
0.12
0.06 - 0.07
0.06 - 0.08
0.05 - 0.07
0.04 - 0.06
0.06
0.05
0.03
0.03
0.17-0.21
0.07 - 0.08
0.07
0.07
0.04 -0.07
0.03
0.13-0.14
0.07
0.07
0.02
0.12
0.08-0.1
0.04 - 0.05
0.06
0.07
LEV
Tierl
Tierl
TLEV
TLEV
LEV
Tierl
LEV
LEV
LEV
TLEV
Tierl
TLEV
Tierl
TLEV
TLEV
LEV
Tierl
TLEV
Tierl
LEV
Tierl
TLEV
TLEV
LEV
TLEV
IV-16
-------
Chapter IV: Technological Feasibility
Mitsubishi
Volvo
Volvo
Daimler Chrysler
Honda
Honda
Honda
Infiniti
Mirage
S80
S80
C230 Kompressor
Accord
Civic HX
Civic
Q45
0.07
0.06-0.07*
0.04 - 0.05
0.07
0.07*
0.07*
0.07*
0.07*
0.05
0.07 - 0.08
0.11
0.03
0.04 - 0.05
0.09
0.07 - 0.08
0.11
LEV
TLEV
TLEV
TLEV
LEV
TLEV
TLEV
Tierl
LDT1
Daimler Chrysler
Ford
Mazda
Ford
Jeep Cherokee 2WD, 4WD
Ranger
B2500, B3000
Ranger
0.03*+
0.04 - 0.07
0.04 - 0.06
0.05*
0.06
0.09-0.18
0.08-0.13
0.11
Tierl
Tierl
Tierl
Tierl
LDT2
Ford
Ford, Mazda
Ford
Mazda
Ford, Mazda
Daimler Chrysler
Nissan
Explorer
Ranger, B3000
F-150
B3000
Ranger, B3000
Caravan, Voyager
Frontier
0.03-0.04
0.04-0.07
0.05*
0.05*
0.05-0.07
0.07
0.07*
0.07-0.10
0.12-0.15
0.08-0.10
0.06 - 0.07
0.07-0.12
0.07
Tierl
Tierl
Tierl
Tierl
Tierl
LEV
LEV
LDT3
Ford
Ford
F-150
F-150
0.04 - 0.06
0.05 - 0.06
0.07 - 0.08
0.11-0.12
Tierl
Tierl
LDT4
Ford
Expedition, Navigator, F-250
0.04*
0.16-0.17
Tierl
* Other model configurations have NOx certification levels above 0.07 g/mile
+ The official NOx certification result reported was 0 for these vehicles due to rounding. The
values shown are the unrounded results.
A review of the Table above shows that most of the engine families with configurations
IV-17
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
certified at 0.07 g/mile NOx or less are passenger cars and lighter weight LDTs . This is
understandable since all LDT classes except LDT1 have emission standards considerably higher
than LDVs. Thus, to this point, there has been no motivation for vehicle manufacturers to design
and produce light-duty trucks with emission control systems on par with light-duty vehicle
systems. Even so, there are several light-duty trucks with certification levels at or very close to
the proposed Tier 2 requirements.
/'/'. Industry Sulfur Test Program
The Coordinating Research Council (CRC), automobile manufacturers and the American
Petroleum Institute (API) all tested a number of vehicles capable of complying with the
California LEV or ULEV standards. The primary purpose of these test programs was to estimate
how higher fuel sulfur levels affected emissions. However, the test results with low sulfur fuel
(i.e., 30-40 ppm sulfur) provide an indication of the emission control potential of these vehicles.
Of the 20 unique vehicle models tested in these programs, four models met both of the Tier 2
NMOG and NOx design targets mentioned above. An additional three models had NMOG
levels below the design targets and NOx levels above the design targets, but below the proposed
NOx standard. All of these low emitting models were LDVs with 100K catalyst systems.
Hi. MECA Test Program
The Manufacturers of Emission Controls Association (MECA) sponsored vehicle
emission testing at the Southwest Research Institute (SwRI)1 for the purpose of demonstrating
the performance of advanced emission control systems in meeting proposed California LEV II
and our proposed Tier 2 light-duty vehicle standards. SwRI took two LDVs (a 1997 3.8L Buick
LeSabre and a 1997 4.6L Ford Crown Vic) and one LDT2 (3.4L Toyota T100) certified to the
federal Tier 1 standards and replaced the original catalytic converters with more advanced
catalytic converters provided by MECA members. The catalysts were thermally aged to roughly
50,000 mile of in-use operation. SwRI then attempted to optimize the emission performance by
modifying the existing secondary air and exhaust gas recirculation (EGR) strategies. This was
accomplished by using a computer controlled intercept system (Emissions Reduction Intercept
and Control system or ERIC). This computer intercept methodology was used to recognize and
modify only driving modes associated with high tailpipe emission modes, thereby minimizing
the level of modifications to the base vehicle control system. The control tuning approach
developed for each vehicle was unique to the individual vehicle. The computer intercept
techniques used in this program were capable of modifying secondary air and EGR without
setting any on-board diagnostic codes. The modified control strategies also did not have any
measurable impact on fuel economy, nor were any detectable changes to vehicle driveability
observed during FTP evaluations.
After these modifications, all three vehicles met the Tier 2 NMOG usefull life design
targets. The LeSabre and T100 both met the NOx design target. The Crown Victoria, however
was a little short of the design target, but did meet the proposed Tier 2 standard with a headroom
IV-18
-------
Chapter IV: Technological Feasibility
of 23 percent. The actual test results are summarized in Table IV-4 below.
Table IV-4. MECA Test Program: Emissions with Catalysts Aged to 100,000 Miles (g/mi)
Tier 2 Design Targets
Crown Victoria (LDV)
Buick LeSabre (LDV)
ToyotaT100(LDT2)
NMOG
0.045-0.063
0.049
0.038
0.052
NOx
0.035-0.049
0.057
0.037
0.014
iv. ARE Test Program
ARB tested five different 1997-98 model year production LEV LDV models. Two of the
six models met the proposed Tier 2 design targets for NMOG and NOx. Each vehicle was tested
for baseline emissions at approximately IK miles before any modifications to the vehicle's
emission controls were made. Table IV-5 lists the average emissions from these FTP tests.
Table IV-5. ARB Production LEV LDV Passenger Car Emission Data
Test Vehicle
1997 Mercury Sable
1998 Mercury Grand Marquis
1 998 Nissan Altima
1998 Honda Accord EX
1998 Toy ota Aval on
NMHC (g/mi)
0.035
0.048
0.031
0.025
0.044
CO (g/mi)
0.9
0.6
0.7
0.3
0.4
NOx (g/mi)
0.072
0.014
0.040
0.066
0.111
After the baseline FTP results were complete, new advanced catalysts supplied by
various catalyst suppliers were installed on each test vehicle. In general, the advanced catalysts
were placed in the same position as the OEM catalysts. Two of the vehicles had small close-
coupled catalysts added to the OEM configuration. FTP tests were then conducted. If the
emission results were not below the proposed LEV II standards with a reasonable margin, engine
calibration modifications such as spark retard at engine start, O2 sensor biasing (typically rich),
or secondary air injection modifications were made to reduce tailpipe emission levels further. In
a couple of instances, approximately 4K miles were accumulated on the "green" catalysts before
FTP tests were conducted again. All of the vehicles, once modified, had emission levels well
below the proposed Tier 2 NMOG and NOx design targets. While these results are not with
catalysts aged tofull usefull life, we believe these results are still very promising, since in-use
IV-19
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
deterioration rates have been steadily declining. Even if these emissions were to double, they
would still be very close to or below the proposed Tier 2 standards. Table IV-6 lists the modified
passenger car emission results.
Table IV-6. ARB Modified Passenger Car Emission Data
(advanced catalysts with modifications to fuel and/or spark & secondary air)
Manufacturer
Mercury
Mercury
Nissan
Honda
Model
Sable
Grand
Marquis
Altima
Accord EX
Mileage
0
4000
0
0
NMOG
(g/mi)
0.029
0.033
0.028
0.026
CO
(g/mi)
1.0
0.5
0.7
0.4
NOx
(g/mi)
0.036
0.004
0.033
0.035
ARB also tested two identical 1998 Ford Expeditions (LDT4). Both vehicles were tested
in the baseline OEM configuration at 2,000 miles with promising results. Table IV-7 lists the
baseline emission results for the two Expeditions.
Table IV-7. ARB Ford Expedition Baseline Emission Test Results
Vehicle
#2
#3
No. of Tests
8
6
NMHC
(g/mi)
0.090
0.077
CO
(g/mi)
1.69
1.57
NOx
(g/mi)
0.030
0.031
ARB installed advanced Pd/Rh catalyst systems bench aged to 50,000 miles along with
50,000 mile bench aged oxygen sensors on both vehicles and were able to reduce NOx emissions
about 50 percent from the NOx certification level of 0.14 g/mi. ARB also added secondary air to
the vehicles and made some modifications to the spark timing (retarded) and oxygen sensor bias
(rich) and found that they were able to further reduce emissions. Table IV-8 lists the emission
results of the Expeditions with advanced catalyst systems.
IV-20
-------
Chapter IV: Technological Feasibility
Table IV-8. ARB Expedition Emission Results with Advanced Catalyst Systems
Vehicle
#2
#3
No. Of Tests
4
7
NMHC
(g/mi)
0.111
0.112
CO
(g/mi)
3.32
2.91
NOx
(g/mi)
0.048
0.052
• EPA Test Program
Our test program was aimed at lowering the emissions of a large 1999 LDT3 vehicle.
This vehicle has a high horsepower engine, four wheel drive, and a curb weight of 4,500 pounds
(GVW of 6,000 Ibs). The exhaust system of the pickup was modified to incorporate two close
coupled and two underfloor catalytic converters provided by MECA. The catalytic converters
used for these tests were nearly identical to the system used on "Vehicle 1" in the work by Webb
et al.s The only modifications from the specification used by Webb et al. were insulation of the
close coupled catalytic converters using a woven ceramic fiber mesh, and the use of a somewhat
larger exhaust pipe diameter (same as OEM) to limit exhaust backpressure. All of the catalyst
"bricks" were constructed of a 4-mil ceramic monolithic material. The two close coupled
catalytic converters each used two "bricks"and were mounted immediately downstream of each
exhaust manifold. The first "brick" in each close coupled catalytic converter had a diameter of
7.6 cm, was 7.6 cm long, with a cell density of 600 cpsi, and was coated with 250 g/ft3 of Pd.
The second "brick" in each close coupled catalytic converter had a diameter of 10.2 cm, was 8.5
cm long, with a cell density of 400 cpsi, and was coated with Pd/Rh at a 9:1 ratio to 150 g/ft3.
The two underfloor catalytic converters each contained a single "brick" with a diameter of 10.2
cm, was 11.1 cm long, with a cell density of 400 cpsi, and was coated with Pd/Rh at a 9:1 to 80
g/ft3.
SC.C Webb, B.B. Bykowski, P.A. Weber, D.L. McKinnon; "Using Advanced Emissions Control Systems to
Demonstrate LEV IIULEV on Light-Duty Gasoline Vehicles". SAE Technical Paper Series, Paper No. 1999-01-
0774, 1999.
IV-21
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
tfl
c
o
'«
«
E
uj
O
z
Q
I
S
z
n dfi
0.44 :
0.42 :
0.4 -
0.38 :
0.36 :
0.34 :
0.32 :
0.3 -
0.28 -
0.26 :
0.24 :
0.22 -
0.2 :
0.18 :
0.16 :
0.14 :
0.12 :
0.1 :
0.08 -
0.06 :
0.04 :
0.02 :
n
-^-^-u^
A
=t
5
1
a
i,,,,!,,,,!,,,,!,,,,!,,,,!,,,,!,,,,!,,,,!,,,,!.
/\ Initial Certification Level
9 Current Status
| Research Goal
(— i
fl?
|
Q)'
1 —
-
o
-^-^-U^-^-U-^-^-U^-^-U-r-^
^^4-———^-———^-——
0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95
NOx Emissions (g/mi)
Figure IV-2: Certification Emissions and Initial Research and Development Results for a
Large 1999 LDT3 Pickup, as Compared to Federal Tier 1 Emissions Standards, California
LEV-I Emissions Standards, and a Tier 2 Corporate Average Target.
Initially, no attempts were made to alter the calibration of the electronic engine controls.
Figure IV-2 shows emissions results from testing of the LDT3 vehicle. In this configuration, the
pickup achieved emissions levels of 0.060 ± 0.002 g/mi NOx and 0.09 ± 0.01 g/mi NMHC.
Based on initial modal emissions results, EPA staff engineers have indicated that the research
targets of 0.04 g/mi NOx and 0.06 g/mi NMHC will likely be reached through elimination of
fuel cut-offs during decelerations, slight increases in EGR, and a minor degree of air-injection
during cold-start.
v. Summary of Technical Feasibility Data
It is very apparent from the data presented above that it is technically feasible for LDVs
to meet our proposed Tier 2 emission standards in the proposed time frame. Although the bulk
of the data is for LDVs with smaller displacement engines, there are examples of vehicles with
larger displacement engines - the 1999 Ford Mustang Convertible and the Ford Crown Vic and
Buick LeSabre modified by MECA. Neither of the MECA or ARB test programs modified the
basic engine calibrations of the vehicles tested. In general, the engine calibration is designed
specifically to match the engine exhaust to the catalytic converter being used on the vehicle.
IV-22
-------
Chapter IV: Technological Feasibility
Since recalibrations optimize engine performance, durability, emissions and safety
simultaneously, neither MECA nor ARB modified the original engine calibration. However, it is
very likely that such recalibration could better match engine operation and the advanced catalytic
converters being tested and reduce emissions beyond the emission levels measured in the test
programs. Therefore, we are confident that by 2004, all LDVs should be capable of meeting
Tier 2 standards.
Fewer data are available addressing the ability of LDTs to meet the design targets
implied by the proposed Tier 2 NMOG and NOx standards. No current LDTs have been
certified at such low emission levels. However, this is partially due to the fact that their current
emission standards are generally well above those for LDVs. Also, the number of LDTs
required to comply with ARB's current LEV and ULEV standards is much lower than the
number of LDVs (and LDTls). Thus, manufacturers have focused their early LEV and ULEV
development efforts on LDVs.
The fact that a number of LDTs have been certified at emission levels near the proposed
Tier 2 standards indicates that control technology has developed to the point where the
differences between LDV and LDT emissions are shrinking.
As highlighted by MECA at the November 1998 annual meeting between EPA and
MECA to discuss recent emission control technology developments, there are several areas
where technology gaps exist between LDV and LDT technology. Table IV-9 lists these gaps.
Table IV-9. Emission Technology Gaps Between LDVs and LDTs
LEVLDVs
Tight A/F control
Close-coupled + underfloor catalyst
combination
Ratio of catalyst volume to engine
displacement Vcat/Veng = 1-1.5
Catalysts with advanced washcoats and
higher cell density substrates
Tier 1 LDTs
Relatively loose A/F control
Underfloor catalyst with long
pipe runs
Vcat/Veng = 0.5 or less
Less sophisticated catalysts on standard
substrates
These differences have been due to the fact that LDT standards were less stringent than
those applicable to LDVs. However, there are no technological reasons why LDTs cannot
employ the exact same technology, or even better technology, as LDVs.
The Toyota T100 pick-up (LDT2) modified by MECA and optimized by SwRI had very
impressive emission results at 50,000 miles - considerably lower than the Tier 2 design targets.
The ARB Ford Expedition had emission reductions of over 50 percent from certification levels,
IV-23
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
achieving results just at the LEV II and Tier 2 NOx standard (NMOG was below the design
target) with the addition of an advanced catalyst system and the introduction of secondary air at
start-up. Finally, our own LDT3 test program had emission results below Tier 2 standards. As
stated above, none of the above mentioned test programs modified the basic engine calibrations
of the vehicles tested. The modifications they made were minor compared to what vehicle
manufacturers are capable of doing. For example, the 1999 certification results for the Ford
Expedition show a slight increase in NMOG levels relative to the 1998 model, but a 50 percent
reduction in NOx levels at 50,000 miles. The full usefull life NOx certification level in 1998
was 0.14 g/mi but only 0.07 g/mi in 1999. The only difference between the two model years is
engine calibration - both model years have the same catalyst system. This highlights the
importance of engine calibration on emission control system performance.
Traditionally LDTs have had higher standards because they tend to generate more
emissions due to larger displacement engines and the fact that they have typically been operated
under high loads. The greatest concern with controlling emissions for LDTs has always been
that catalysts couldn't be placed close enough to the engine to reduce cold start emissions
because of concerns over thermal damage to the catalyst resulting from the high levels of heat
generated by the trucks when under load. But as discussed above, catalysts are now capable of
withstanding much higher temperatures, and this is no longer an issue.
Another reason why the emission standards for LDT3s and LDT4s were set so much
higher than those for LDVs and lighter LDTs was that the heavier LDTs were tested at adjusted
loaded vehicle weight, not loaded vehicle weight. Adjusted loaded vehicle weight is the
vehicle's curb weight plus half of its maximum payload capacity. Loaded vehicle weight is the
vehicle's curb weight plus 300 pounds. This was done in the past because the LDT3s and
LDT4s were believed to be used primarily as cargo carrying vehicles and should be regulated
under these conditions. While their weight during emission testing generally increases
emissions, the applicable emission standards were numerically increased to compensate for this.
As part of the Tier 2 proposal, LDT3s and LDT4s will be tested like LDVs and the
lighter LDTs, at the vehicle's curb weight plus 300 pounds. This change represents the recent
trend for these trucks to be used predominantly as passenger carrying vehicles. This change
would also reduce the certified emission level of any current LDT3 or LDT4 simply by reducing
the amount of fuel the vehicle consumes over the test cycle. Under the proposed test procedure,
all current LDT3s and LDT4s would be closer to the design targets for the proposed Tier 2
standards. Likewise, the difference between the current Tier 1 standards for these vehicles and
the proposed Tier 2 standards is actually much smaller than a comparison of the numerical
standards would indicate.
Overall, several certified and research LDVs have met the design targets for the proposed
Tier 2 standards in 1998 with, in many cases, considerable margin to spare. This indicates that
significant margin exists with which to accommodate the greater weight and aerodynamic drag
of LDT3s and LDT4s when meeting the same design targets. In addition, significant LDT
emission reductions were achieved from vehicles emitting well below the applicable LEV design
targets by primarily just changing catalysts (in the case of MECA and ARB) or just changing
IV-24
-------
Chapter IV: Technological Feasibility
engine calibration (in the case of Ford). This confirms indicates that the benefits of advanced
emission control technology on LDTs has not been exhausted in meeting current LEV emission
standards. Overall, these findings indicate that the proposed Tier 2 standards for LDTs should
be feasible with the same basic types of emission controls as required by LDVs. The heavier
LDTs will likely require somewhat larger catalysts than LDVs and the lighter LDTs and possibly
also incorporate a greater number of supporting technologies, such as a low-thermal capacity
manifold, in order to meet the same numerical emission standard.
c. Lean-Burn Technology
The above discussion focused on advancements in emission control technology. New
gasoline engines designs are also being developed to reduce fuel consumption. In particular,
gasoline direct-injection (GDI) engines have been developed (and are being sold in Japan and
Europe) which operate on 10-20 percent less fuel than today's gasoline engines.
One of the reasons that these engines use less fuel is that they use much more air than is
needed just to burn the fuel. In this respect, they operate similar to a diesel engine. While this is
advantageous for fuel efficiency, it makes it more difficult to eliminate NOx emission using
aftertreatment technology. Highly efficient 3-way catalysts require that there be little excess
oxygen in the exhaust stream in order to convert NOx emissions to nitrogen and oxygen.
Unfortunately, if a GDI engine is operated in this way, nearly all of its fuel efficiency benefits
are lost.
A number of potential techniques are being developed to control NOx emissions when
excess air are present. These techniques are discussed in more detail in Section 5. below. The
most promising of these techniques for GDI engines are the lean NOx catalyst and the NOx
adsorber. A GDI engine may be able to meet the highest NOx bin of 0.2 g/mi with a lean NOx
catalyst and 30 ppm sulfur gasoline. However, this is dependent on achieving engine out
emissions below 0.4 g/mi, and this is likely to be a very difficult challenge. NOx adsorbers are
potentially more efficient than lean NOx catalysts. Thus, the required engine out NOx emission
levels are likely to be well within the capability of GDI engine technology. However, the NOx
adsorber is in an earlier stage of development relative to the lean NOx catalyst. Much
development work is still necessary in order for this technology to be applied commercially.
Also, even 30 ppm sulfur levels degrade long-term performance of NOx adsorbers. Thus, either
methods to regenerate the NOx adsorber onboard the vehicle must be developed or even lower
sulfur levels will be required.
2. CO Emissions from Gasoline Fueled Vehicles
EPA is only proposing tighter CO emission standards for LDT2s, LDT3s and LDT4s.
Basically, CO emissions from these vehicles must be reduced to the levels now required for
LDVs and LDTls under the NLEV program. Also, LDVs and LDTs must comply with the
NLEV CO standards over a slightly longer useful life of 120,000 miles instead of the current
IV-25
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
useful life of 100,000 miles.
Compliance with the proposed Tier 2 CO emission standards should not be difficult given
compliance with the proposed Tier 2 NMOG standards. The control of both pollutants utilizes
much of the same technology and the proposed Tier 2 NMOG standards are the more stringent of
the two sets of standards. In addition, the above mentioned change in test weight should make it
even more easy to meet the proposed Tier 2 CO emission standards. The following table IV-10
summarizes CO emissions from vehicles certified to the LEV standards in California.
Table IV-10. CO Emissions from California LEVs (g/mi)
LDV/LDT
0.8
LDT2
1.13
LDT3
2.35
LDT4
2.95
As can be seen, the CO emissions from all of these vehicles are well below the proposed
Tier 2 CO standard of 4.2 g/mi. While CO emissions from LDT3s and LDT4s are more than half
the proposed 4.2 g/mi standard, the current LEV standards for these vehicles is more than twice
the proposed Tier 2 NMOG standard of 0.09 g/mi. As NMOG emissions are reduced to meet the
0.09 g/mi standard, CO emissions will decrease further, as well. CO emission control is also not
a problem for GDI engines. Thus, compliance with the proposed Tier 2 CO standard should not
add any additional burden to manufacturers relative to compliance with the proposed NMOG and
NOx standards.
3. Formaldehyde Emissions from Gasoline Fueled Vehicles
EPA is only proposing tighter formaldehyde emission standards for LDT2s, LDT3s and
LDT4s. Basically, formaldehyde emissions from these vehicles must be reduced to the levels
now required for LDVs and LDTls under the NLEV program. Also, LDVs and LDTls would
have to comply with the NLEV formaldehyde standards over a slightly longer useful life of
120,000 miles versus the current 100,000 mile useful life.
Again, as with CO emissions, compliance with the proposed Tier 2 formaldehyde
emission standards should not be difficult given compliance with the proposed Tier 2 NMOG
standards. The control of both pollutants utilizes much of the same technology and the proposed
Tier 2 NMOG standards are the more stringent of the two sets of standards. Table IV-11, below,
summarizes formaldehyde emissions from vehicles certified to the LEV standards in California.
IV-26
-------
Chapter IV: Technological Feasibility
Table IV-11. Formaldehyde Emissions from California LEVs (g/mi)
LDV
0.0012
LDV/LDT1
0.0016
LDT2
0.0013
LDT3
0.002
LDT4
0.002
As can be seen, formaldehyde emissions from current California vehicles are roughly a
factor of 10 below the proposed Tier 2 formaldehyde standard of 0.018 g/mi. Thus, compliance
with the proposed Tier 2 formaldehyde standard should not add any additional burden to
manufacturers relative to compliance with the proposed NMOG and NOx standards.
4. Evaporative Emissions
The standards we are proposing today for evaporative emissions are technologically
feasible now. Many designs have been certified by a wide variety of manufacturers that already
meet these standards. A review of the 1999 model year certification results indicates that the
average family is certified at less than 1.0 grams per test on the 3 day diurnal plus hot soak test,
i.e. at less than half the current 2.0 g/test standard.
The standards we are proposing will not require the development of new materials or, in
many cases, even the new application of existing materials. Low permeability materials and low
loss connections and seals are already used to varying degrees on current vehicles. Today's
standards will likely ensure their consistent use and discourage manufacturers from switching to
cheaper materials or designs to take advantage of the large safety margins they have under
current standards ("backsliding").
There are two approaches to reducing evaporative emissions for a given fuel. One is to
minimize the potential for permeation and leakage by reducing the number of hoses, fittings and
connections. The second is to use less permeable hoses and lower loss fittings and connections.
Manufacturers are already employing both approaches.
Most manufacturers are moving to "returnless" fuel injection systems, and at least one
major manufacturer utilizes returnless systems on all of their vehicles. Through more precise
fuel pumping and metering, these systems eliminate the return line in the fuel injection system
which carries unneeded fuel from the fuel injectors, which has been heated from its close
proximity to the hot engine, back to the fuel tank. Returned fuel is a significant source of fuel
tank heat and vapor generation. The elimination of return lines also reduces the total length of
hose on the vehicle and also reduces the number of fittings and connections which can leak.
Low permeability hoses and seals as well as low loss fittings are available and are
already in use on many vehicles. Fluoropolymer materials can be added as liners to hose and
component materials to yield large reductions in permeability over such conventional materials
as monowall nylon. In addition, fluoropolymer materials can greatly reduce the adverse impact
IV-27
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
of alcohols in gasoline on permeability of evaporative components, hoses and seals.
5. Diesel Vehicles
Very few LDVs and LDTs are currently sold with diesel engines. Out of annual sales of
more than 15 million vehicles, roughly only 30,000 are equipped with diesel engines. This is in
part due to the low price of gasoline over the past decade. Recently, a number of vehicle
manufacturers have announced aggressive programs to increase the sales of diesel LDVs and
LDTs. These programs are scheduled to begin between model year 2000 and 2002 and appear to
be aimed at easing manufacturers' compliance with the corporate fuel economy standards,
particularly for LDTs.
NOx and PM are the two biggest emission-related challenges for diesel engines. Diesels
have inherently low emissions of CO and NMOG and should have no problem meeting the
proposed Tier 2 standards for these pollutants. Therefore, the remainder of this discussion will
focus on NOx and PM.
Considerable progress has been made over the past 10 years in reducing engine-out
emissions from diesel engines. In heavy trucks and buses, PM emission standards which were
projected to require the use of exhaust aftertreatment devices were met with only engine
modifications. NOx emissions from heavy trucks and buses sold starting in 2002 will also
reflect deep reductions from emission levels typical of engines produced in the mid-1980's.
However, the benefits of improved diesel engine design appear to be reaching their limits. EPA
projects that diesel LDVs and LDTs could only meet NOx emissions standards of roughly 0.5-
0.6 g/mi and PM emission standards of roughly 0.05-0.08 g/mi without aftertreatment. These
levels are well above the highest allowable certification emission levels included in today's
proposal.
In order to comply with the proposed Tier 2 emission standards for NOx and PM, diesels
will require the use of effective aftertreatment devices. For NOx emissions, these devices
include lean NOx catalysts, selective catalytic reduction (SCR) and NOx adsorbers. Lean NOx
catalysts are still under development and appear capable of reducing NOx emissions by 15-30
percent. Therefore, they are unlikely to be sufficiently effective to enable compliance with the
proposed Tier 2 standards.
SCR has been demonstrated commercially on stationary diesel engines and can reduce
NOx emissions by 80-90 percent. However, SCR requires that the chemical urea be injected into
the exhaust before the catalyst. This means that vehicle owners would have to periodically
refuel their vehicle with urea, as well as diesel fuel. Ammonia emissions also occasionally occur
with use of SCR, which has a very objectionable odor.
Finally, NOx adsorbers can be up to 90 percent efficient at removing NOx from the
exhaust. However, these adsorbers are quickly poisoned by sulfur in the fuel and would require
reductions in diesel fuel sulfur content. Their use also requires that the engine be run with
IV-28
-------
Chapter IV: Technological Feasibility
excess fuel occasionally, so that the stored NOx can be converted to nitrogen and oxygen. These
adsorbers, coupled with techniques for introducing fuel into the exhaust periodically are still at
the research stage.
Overall, use of either SCR or a NOx adsorber should be able to enable compliance with
the proposed Tier 2 standards. The issue of reducing diesel fuel sulfur levels to enable NOx
adsorbers and other technologies is discussed in a separate ANPRM.
Regarding PM, applicable aftertreatment devices tend to fall into two categories:
oxidation catalysts and traps. Diesel oxidation catalysts look very similar to the 3-way catalysts
used on gasoline vehicles. Diesel catalysts convert the hydrocarbon compounds in the exhaust to
water and carbon dioxide. This reduces exhaust NMOG emissions and heavier HC compounds
which comprise about 30 percent of total PM mass emissions. The oxidation catalyst can be
from 50 percent to 90 percent effective at converting HC. Thus, an oxidation catalyst can reduce
total PM emissions by roughly 15-27 percent. By itself, the oxidation catalyst is not likely to be
sufficient to enable compliance with the proposed Tier 2 standards without further advancements
in engine technology.
Traps can eliminate up to 90 percent of diesel PM emissions. The trap first filters the
carbonaceous particles from the exhaust. Then, periodically, this trapped PM must be burned, or
the trap will fill up and cause problems in operating the engine. Diesel traps are currently being
used on buses in an number of U.S. cities. It appears that these traps can regenerate frequently
enough given the operating temperatures of bus engines and over-the-road trucks. However,
there is some question whether or not these traps could regenerate frequently enough with the
somewhat lower operating temperatures of diesel engines in LDVs and LDTs. Regeneration can
be enhanced at lower exhaust temperatures through the use of more active catalysts on the
surface of the trap. However, these catalytic materials convert sulfur dioxide in the exhaust to
sulfuric acid. Thus, their use requires the removal of most of the sulfur in the fuel. Research
indicates that low temperature regeneration may also be enhanced through the use of catalytic
fuel additives comprised of cerium or iron. However, particulate containing these chemicals can
be emitted from the tailpipe, raising some health concerns. Use of these catalytic fuel additives
does not require the removal of sulfur from diesel fuel. An efficient trap should enable
compliance with the proposed Tier 2 PM standards.
B. Feasibility of Removing Sulfur from Gasoline
1. Source of Gasoline Sulfur
Sulfur is in gasoline because it naturally occurs in crude oil. Crude oil contains anywhere
from fractions of a percent of sulfur, such as less than 0.05 weight percent (0.05 percent is the
same as 500 ppm) to as much as several percent.2 The average amount of sulfur in crude oil
refined in the U.S. is about one percent.3 Most of sulfur in crude oil is in the heaviest part, or in
the heaviest petroleum compounds, of the crude oil (outside of the gasoline boiling range). In
the process of refining crude oil into finished products, such as gasoline, some of the heavy
IV-29
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
compounds are broken up into smaller compounds and the embedded sulfur ends up in gasoline.
Thus, the refinery units which converts the heavy parts of crude oil into gasoline are the units
most responsible for putting sulfur into gasoline.
The fluidized catalytic cracker (FCC) unit is the refinery processing unit most responsible
for moving sulfur into gasoline. The FCC unit cracks large carbon molecules into smaller ones
and produces anywhere from 30 to 50 percent of the gasoline in most refineries. Because the
FCC unit makes gasoline out of the heavier, higher sulfur-containing compounds, more than 90
percent of sulfur in gasoline comes from that unit.4 Another refinery unit which is responsible
for a significant amount of sulfur in gasoline is the coker unit. These units produce coke from
the heavy part of the crude oil. In the process of producing coke, some gasoline blendstocks are
produced and some of these blendstocks are blended directly into gasoline. While the volume of
gasoline blendstock produced by the coker is small (normally less than one percent of the
gasoline pool), this stream usually contains more than 3000 ppm sulfur,5 so the contribution of
sulfur to gasoline is significant.
Another gasoline blendstock which contributes sulfur to gasoline is the straight run.
Straight run is the portion of the crude oil which falls in the gasoline boiling range which is
blended directly into gasoline. Usually only the light straight run is blended into gasoline which
has a small amount of sulfur (i.e., on the order of 100 ppm sulfur), although in trying to meet a
low sulfur standard, even this amount sulfur of becomes significant. The heaviest portion of
straight run, which would have more sulfur, is normally desulfurized and reformed in the
reformer (to improve its octane), so its contribution to the gasoline pool is virtually nil. Alkylate
is another stream which can have enough sulfur worth mentioning. Most refineries have less
than five ppm sulfur in this pool, however, some refineries which feed coker naphtha to the
alkylate plant can have much more. On average, alkylate probably has about 10 ppm sulfur.
Other gasoline blendstock streams with either very low or no sulfur are alkylate, hydrocrackate,
and isomerate. Oxygenates which are blended into gasoline usually have no sulfur, however,
during shipping through pipelines, they can pick up some sulfur. The implementation of a low
gasoline sulfur standard, though, would reduce much of the sulfur which oxygenates could pick
up in the pipeline.
Since FCC units and cokers contribute so much sulfur to gasoline, then a reasonable
conclusion which could be reached would be that refineries could simply shut down these units
to meet a low sulfur standard. This conclusion is not reasonable considering the quality of crude
oil which is used today and the products demanded of the oil industry. Much of volume of crude
oil is composed of heavy compounds which has no end use, and thus is not usable without
processing by these units. Thus, these units make marketable products from what would
otherwise be a waste oil stream.
2. Current Levels of Sulfur in Gasoline
It is important to know the average level of sulfur in gasoline because it helps determine
the most effective removal methods which should be used. The American Petroleum Institute
IV-30
-------
Chapter IV: Technological Feasibility
(API) and the National Petrochemical Refiners Association (NPRA) surveyed gasoline producers
to gather information concerning refining operations during the Summer of 1996.6 They
collected information on the qualities of gasoline for various regions called Petroleum
Administrative Districts for Defense (PADDs), as well as for the country as a whole. (These
PADDs are illustrated below in Figure IV-2). Their study showed that the gasoline sulfur,
outside of California, averaged 340 ppm during the Summer of 1996.
Petroleum Administration (or Defense {PAD} Districts
U -\ (MM) ^JM.f.^.^yMW*
-.. KJWAV <*kH1
I- „ ^j^fiSgr
I W»NS •! •:..'•• m > /WVA ' VvNnc
I ; MO vj /A D* '\\XDE"
•*! (xi*. p-^'TE^xfV ; h.c. ^
^ L*- /'
•
\
,FU;\
Figure IV-2. Map of U.S. Petroleum Administrative Districts for Defense
When looking closer at the information provided in the report, we discovered that some
PADD sulfur levels calculated from the API/NPRA data were not in agreement with some of the
average blendstock sulfur levels presented within the same report, nor was it consistent with data
reported to EPA for the RFG program in 1995 and 1996. One possible reason for the
disagreement between the API/NPRA gasoline pool sulfur level and that reported to the RFG
data base is that API and NPRA only surveyed refiners for their summertime gasoline qualities.
Other possible reasons for the discrepancy are, that some refiners did not participate with the
API/NPRA survey (especially in PADDs 1 and 5), while data handling complications also
precluded the inclusion of gasoline sulfur data from some refiners from being reported in the
RFG data base. However, because the RFG data base contains year-round data and because it
often represents a larger portion of the gasoline sulfur pool, when the sulfur levels were
compared between the two data bases, or when the API/NPRA information was compared
internally, and there was disagreement, then the RFG data were used over the API/NPRA survey
information. The methodology for adjusting the average sulfur levels is described in more detail
below in the section on fuel costs.
IV-31
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
After adjusting the sulfur values for each PADD, the national average gasoline sulfur
level for domestically produced gasoline is 305 ppm. Table IV-12 below summarizes the U.S.
sulfur levels by PADD, and for the country as a whole. Because California has it own low sulfur
gasoline program, gasoline produced there was excluded from consideration in this analysis.
Table IV-12. Average Sulfur Levels by PADD and for the Nation.
Estimated Average
Sulfur Levels
PADD 1
215
PADD 2
338
PADD 3
308
PADD 4
265
PADD 5
OC*
506
U.S.
Avg. *
305
* Outside of California
It is important to note that the gasoline sulfur values reported in Table IV-12 are an
attempt to estimate the average gasoline sulfur level for estimating the cost of desulfurization. In
actuality, each sulfur value represents the volumetric average of a range of sulfur values with
each refinery representing a single data point. This range can vary from the tens of ppm to
almost 1000 ppm. The 1000 ppm sulfur level is the upper limit of the amount of sulfur permitted
to be shipped in pipelines in accordance with the American Society for Testing Materiels
(ASTM) consensus standards.7
3. Feasibility of Meeting the Proposed Gasoline Sulfur Standards
The feasibility of meeting the proposed standards for low sulfur gasoline can be
demonstrated in two distinct ways. The first way is to assess whether there is technology
available, or that can reasonably be expected to be available in the lead time provided, to the
refining industry to meet the proposed standard. The second way is to determine if refiners are
already demonstrating that they can meet a low sulfur gasoline standard similar to this proposed
rule. Evidence that refiners are already meeting a stringent gasoline sulfur program is a more
compelling example of feasibility since the technology must already be available if low sulfur
gasoline is already being produced.
It is indeed the case that there are low sulfur gasoline programs already in place. The
State of California requires gasoline sold in the State to meet a 30 ppm gasoline sulfur standard
on average and a 80 ppm cap, among a number of other fuel standards.8 Furthermore, refiners
can produce gasoline which varies in composition, provided that the California Predictive
Emissions Model (which, like EPA's Complex Model, estimates vehicle emissions from fuels of
varying composition) confirms that the proposed fuel formulation meets or exceeds the
emissions reduction that would occur based on the default fuel requirements. California
refineries are using the flexibility provided by the Predictive Model to surpass the prescriptive
standards for gasoline sulfur and are producing gasoline which contains 20 ppm sulfur on
average.9 They are making this very low sulfur gasoline despite using Californian and Alaskan
crude oils which are poorer quality than most other crude oils being used today. Thus, the
IV-32
-------
Chapter IV: Technological Feasibility
experience in California demonstrates that commercial technologies already exist to permit
refiners to produce low sulfur gasoline.
In addition to the California experience here in the U.S., a low sulfur requirement in
Japan provides additional evidence that reducing gasoline sulfur levels to low levels is feasible.
Japanese refineries must meet a 100 ppm per-gallon cap. Based on the gasoline sulfur cap
established there, gasoline in Japan averages about 20 ppm sulfur.
4. Meeting a Low Sulfur Gasoline Standard
The methodology that would be used refiners to lower their sulfur level depends on a
number of factors specific to their refinery. These factors include:
• The gasoline sulfur level prior to the start of the gasoline sulfur program
The refinery configuration (A typical complex refinery is illustrated in Figure IV-3,
below.)
The amount of excess refinery desulfurization equipment on hand
• The quality of feedstocks available, especially crude oil
• The quality and types of products produced
Any plans to change the feedstocks or products of the refinery
The desulfurization technologies available and their cost
• Other regulatory programs affecting refinery operations in the same time frame
IV-3 3
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Natural
Gas
Vacuum Tower
Coker
*• Fuel Gas
LPG
Gasoline
Aromatics
Kerosine
Jet Fuel
Off-roadDiesel
*• Fuel Oil
Resid
Coke
Figure IV-3. Diagram of a Typical Complex Refinery
IV-34
-------
Chapter IV: Technological Feasibility
A refinery's average gasoline sulfur level is the most important factor determining
whether a refiner would need to make a substantial capital investment to meet a sulfur standard.
After numerous discussions with refiners, we believe that those refiners with low gasoline sulfur
levels to begin with (i.e., gasoline sulfur levels lower than, perhaps, 60 ppm) will probably not
need to invest in expensive capital. These refineries have such low sulfur levels due to one or
more of a number of possible reasons. For example, some of these refiners may not have certain
refining units, such as a fluidized catalytic cracker (FCC) unit, or a coker, which convert heavy
boiling stocks to gasoline (Figure IV-3 shows these where these units are placed in a refinery).
As stated above, these units push more sulfur into gasoline and their absence means less sulfur in
gasoline. Alternatively, refiners may use a low sulfur (sweet) crude oil, which results in lower
sulfur gasoline. Or, these refiners may have already installed a hydroprocessing unit, such as
FCC feed hydrotreating, to improve the operations of their refinery which uses a heavier, higher
sulfur (more sour) crude oil. This unit removes much of the sulfur from the heaviest portion of
the heavy gas oil before it is converted into gasoline.
Of the refiners in this first category, the refineries with sulfur levels below the proposed
sulfur standard would not have to do anything. On the other hand, those refineries with sulfur
levels above the proposed standard, but below some level, such as 60 ppm, could probably meet
the standard employing operational changes only, which means avoiding capital investments if
this is desirable. For example, these refiners may be able to meet the proposed sulfur standard
by running existing desulfurization units harder, at capacity if the unit has headroom, or by
debottlenecking the unit, perhaps by using more effective catalyst. Alternatively, these refineries
may be able to meet the proposed sulfur standard by using a slightly sweeter crude oil. Refiners
also have FCC additives available to them which could allow them to reduce their FCC gasoline
sulfur level by 15 to 35 percent.10 Another strategy that these refiners could use to meet the
proposed gasoline sulfur standard would be to undercut the FCC gasoline. By cutting out the
heaviest 10 volume percent of FCC gasoline and sending that part to on-road diesel before
hydrotreating, to off-road diesel or to heating oil, refiners would be able to cut out 40 percent or
even more of the sulfur from this high sulfur blendstock. However, the refiners which choose to
undercut their FCC gasoline would then produce less volume of gasoline which generally
provides the highest profit margin. For this reason undercutting the FCC gasoline would likely
be a short term strategy. Finally, these refinery may be able to meet the sulfur standard by
blending in low sulfur oxygenates. Refiners may be able to employ several of these
desulfurization strategies together in varying degrees. These refiners with very low gasoline
sulfur levels to begin with produce only a small portion of the gasoline consumed in this country,
on the order of five percent.11
The vast majority of gasoline is produced by refineries with higher sulfur levels, and
refiners are expected to install capital equipment in these refineries to meet the proposed
gasoline sulfur standard. As stated above, the FCC unit is responsible for most of the sulfur in
gasoline. Thus, investments for desulfurizing gasoline would likely involve the FCC unit to
maximize the sulfur reduction, and to minimize the cost. This desulfurization capital investment
can be installed to treat the gas oil feed to the FCC unit, or treat the gasoline blendstock which is
produced by the FCC unit. Each method has advantages and disadvantages.
IV-3 5
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
FCC feed hydrotreating can be accomplished by a hydrotreater or a mild hydrocracker.
These units are designed to operate at high pressures and temperatures to treat a number of
contaminants in gas oil. Besides sulfur, FCC feed hydrotreating also reduces nitrogen and
certain metals such as vanadium and nickel. These nonsulfur contaminants adversely affect the
FCC catalyst, so the addition of this unit would improve the yield of higher profit-making
products such as gasoline and diesel. While FCC feed hydrotreating provides these benefits
which partially offsets the costs of adding this type of desulfurization, the costs are still high
enough that many refiners would have a hard time justifying the installation of this sort of unit.
For a medium to large refinery, the capital costs may exceed $100 million. Because of the
higher temperatures and pressures involved, operating costs are expensive relative to other forms
of hydrotreating explained below. Refiners may be better able to justify this approach if they
switch to a heavier, more sour crude oil. These crude oils are less expensive per barrel and can
offset the increased cost of the FCC desulfurization unit, providing that the combiniation of
reduced crude oil costs and higher product revenues justify the switch. Another benefit for using
FCC feed hydrotreating is that the portion of the distillate pool which comes from the FCC unit
would be hydrotreated as well. While this distillate blendstock, termed light cycle oil, comprises
a relatively small portion of the total distillate produced in the refinery (about 20 percent of on-
road diesel comes from light cycle oil), like gasoline, it contributes a larger portion of the total
sulfur which ends up in distillate. Thus, FCC hydrotreating would allow a refiner to produce
more low sulfur onroad diesel or meet a lower sulfur standard for onroad diesel, which could
apply in the future.
A less expensive alternative to hydrotreating the FCC feed for gasoline desulfurization is
FCC gasoline hydrotreating. FCC gasoline hydrotreating only treats the gasoline produced by
the FCC unit. Understandably, this unit is much smaller because only about 60 percent of the
feed to the FCC unit ends up as gasoline produced by the FCC unit. The unit is often smaller
than that as refiners typically only treat the heavier, higher sulfur portion of that stream. FCC
gasoline hydrotreaters operate at lower temperatures and pressures as well which further reduces
the capital and operating costs associated with this type of desulfurization equipment. For a
medium to large refinery, the capital costs would be on the order of $50 million for a
conventional hydrotreater. One drawback of this process is that octane value and some of the
gasoline yield is lost during desulfurization. The loss of this octane must be made up by
additional octane production by other units in the refinery or by oxygenate addition, and the
volume loss can be made up by additional throughput to gasoline producing units, or also by
oxygenate addition.
The loss of octane and gasoline yield caused by FCC gasoline hydrotreating is much
lower with gasoline desulfurization technologies which were recently introduced. CDTECH and
Mobil Oil each have developed new or improved technologies for desulfurizing gasoline.
CDTECH calls its two hydrotreating units for treating FCC full range gasoline CDHydro and
CDHDS. Mobil Oil calls its process OCTGAIN 220. These processes preserve much of the
octane and gasoline yield because they were designed for treating gasoline blendstocks. One
reason why these processes preserve octane and yield is that they operate at lower temperature
and pressure compared to conventional hydrotreating processes. The less severe conditions also
IV-36
-------
Chapter IV: Technological Feasibility
lowers the capital and operating costs for this process. The capital cost for these improved
processes ranges from $20 to $40 million for a medium to large sized refinery. While the capital
costs are somewhat less than conventional hydrotreaters, much more cost savings arise out of the
reduced utility and ancillary costs. For example, because these processes are less severe, there is
much less saturation of olefins, which means that there is much less hydrogen used. Less olefin
saturation also translates into less octane loss which would otherwise have to be made up by
octane boosting processing units in the refinery.
While these improved gasoline desulfurization technologies have been in the limelight
for over a year now, we have also learned that other desulfurization processes are under
development. In conversations with several refiners, they shared with us that they are in the
process of developing their own desulfurization technology. We would expect them to use their
technologies if their development work is completed in time. Furthermore, other refining
process licensing firms shared with us that they are developing their own desulfurization
technology. We recently became aware that biodesufurization, which is the removal of sulfur
from petroleum using biological means, is on the verge of commercialization for the
desulfurization of diesel.12 The vendor of this technology informed us that the process will be set
up and running in about two years in a Petrostar refinery in Alaska. The vendor also shared that
gasoline desulfurization is a little further behind, and they believe that they could be ready to
market that process for desulfurizing gasoline within the next two years. Two important
advantages of this technology are that hydrogen is not needed and that the feed to the unit does
not need to be heated to high temperature and compressed to high pressures. The cost savings of
these benefits are offset, though, by the need for extensive mixing to maintain an oxygen-rich
environment for the bacteria to work effectively. We anticipate that this processes will receive
much more attention in the near future as the vendor completes its initial pilot plant work. Our
initial assessment is that this process, given time, may be cost competitive with any of the
improved gasoline desulfurization technologies. However, at this time, we are only focusing on
the improved gasoline desulfurization technologies for which we were given cost information
that allowed us to estimate the cost of gasoline desulfurization.
5. Improved Gasoline Desulfurization Technology
We will briefly describe the improved gasoline desulfurization technologies below, but
first we'll describe conventional desulfurizing technology to establish a point of reference.
Conventional desulfurization occurs in a fixed bed reactor.13 The reactor is termed fixed bed
because a catalyst, which helps to improve the reaction rate and specificity, is contained in a
stationary bed within the reactor. The high sulfur gasoline blendstock is heated to a high
temperature (on the order of 600 degrees Fahrenheit) and pressure, to maintain the stream as a
liquid, and is combined with hydrogen before it enters the reactor. The reactions occur over the
bed of the catalyst. While the petroleum is in contact with the catalyst in the reaction vessel, the
sulfur is removed from the petroleum compounds and is converted to hydrogen sulfide. Also,
most or all of the olefin compounds which are present in the cracked stream are saturated. Since
olefin compounds are much higher in octane than their saturated counterparts, their saturation
causes a significant octane loss in any stream with a high olefmic content. Conventional
IV-37
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
hydrotreating has generally been used for streams with little or no olefins, such as for virgin gas
oil which is treated by the FCC unit, or for feed streams to the reformer and isomerizer. Thus,
octane loss upon hydrotreating these streams is not a problem. However, FCC gasoline is very
rich in olefins such that hydrotreating this stream normally causes a large reduction in octane (up
to 8 octane numbers units can be lost if the entire FCC gasoline stream is treated). The
saturation of olefins also substantially increases the need for hydrogen.
The catalyst also tends to cause some of the petroleum compounds to "crack, "
converting them from gasoline boiling range compounds to compounds too light for keeping in
gasoline, which is termed yield loss. After the reactor, the gaseous compounds, which include
unreacted hydrogen, hydrogen sulfide, and any light end petroleum compounds which may have
been produced in the reactor, are separated from the liquid compounds. The hydrogen sulfide
must be stripped out from the other compounds and then converted to elemental sulfur which is
then sold off. If there is enough hydrogen and it can be economically recovered, it is separated
from the remaining hydrocarbon stream and recycled. Otherwise it is burned with the light
hydrocarbons as fuel gas.
Mobil Octgain will be discussed next since the process is similar to a conventional
hydrotreater. This gasoline desulfurization process is the third generation of this process for
Mobil, and it is called Octgain 220. Like a conventional hydrotreater, Mobil uses a fixed bed of
catalyst for its Octgain process.14 One primary difference between Octgain and conventional
hydrotreating is that Mobil Oil has developed its own catalyst for the reactor. The catalyst not
only causes the desulfurization of petroleum, it also causes isomerization reactions to occur
(straight chain petroleum compounds are changed to branched chain compounds) which
increases the octane of the resultant stream.15 The octane improvement caused by the catalyst
compensates for octane loss resulting from olefin saturation. Mobil designed this generation of
Octgain process to operate over a range in severity. If less desulfurization is needed, then the
process temperature and pressure are reduced. The less severe operating conditions reduces the
saturation of olefins, which in turn reduces less hydrogen consumption and utility use, in
addition to causing less octane loss. Under the less severe conditions, much or all of the octane
loss can be made up by the isomerization reactions caused by the catalyst. The less severe
operating conditions also causes less yield loss, as the conditions are less favorable for causing
cracking of the larger petroleum compounds to smaller compounds. Most of these benefits
would be negated if extensive desulfurization, on the order of 99 percent, is necessary. The steps
of the process following the reactor are the same for conventional FCC gasoline hydrotreaters.
One advantage of the Octgain process is that the reactor vessel for the process is the same as
vessels used for conventional hydrotreating. Thus, refiners can save on capital costs by using a
spare hydrotreating unit which it may have on hand in the refinery.
The CDTECH process is significantly different from either conventional hydrotreating or
Octgain, which makes it a little more complex to describe. The CDTECH process utilizes
catalytic distillation.161718 Catalytic distillation is a technology which has been applied for a
number of different purposes. CDTECH is currently licensing the technology to produce MTBE
and selective hydrogenation processes. Based on their experience and success with that process,
they applied the same technology to desulfurizing gasoline. As the name implies, distillation and
IV-38
-------
Chapter IV: Technological Feasibility
desulfurization, via a catalyst, take place in the same vessel. This design feature may save the
need to add a separate distillation column in some refineries. All refineries have a distillation
column after the FCC unit (called the main fractionation column) which separates the gasoline
from the most volatile components (such as liquid petroleum gases), the distillate or diesel (light
cycle oil), and the heavy ends or residual oil. However, if a refiner only wishes to treat a portion
of the FCC gasoline, then he may have to add a second distillation column to be able to separate
off the portion of the FCC gasoline which he wishes not to treat. With the CDTech process, the
refiner can choose to treat the entire pool or a portion of the pool, but choosing to treat a part of
the pool can be an option in how the CDTech hardware is applied, thus negating any need for an
additional distillation column. The choice is realized by setting up the distillation and associated
catalyst to treat only the portion of the pool which needs to be treated.
The most important portion of the CDTech desulfurization process is two distillation
columns loaded with desulfurization catalyst in a packed structure. The first vessel, called
CDHydro, treats the lighter compounds of FCC gasoline, while the second column, called
CDHDS, treats the heavier compounds of FCC gasoline. All of the FCC gasoline is fed to the
CDHydro column. The five- and six-carbon petroleum compounds boil off and head up through
the catalyst mounted in the column, along with hydrogen which is also injected in the bottom of
the column. The reactions in this column are unique in that the sulfur in the column are not
hydrotreated to hydrogen sulfide, but they instead are reacted with dienes in the feed to form
thioethers. Their higher boiling temperature causes the thioethers to fall to the bottom of the
column. They join the heavier petroleum compounds at the bottom of the column and sent to the
CDHDS column. Because the pressure and temperature of the first column is much lower than
conventional hydrotreating, saturation of olefms is reduced to very low levels (according to
CDTECH, the saturation which does occur is desirable to eliminate diolefms). Thus, little
excess hydrogen is consumed. An option for the refiner is to put in an additional catalyst section
in the CDHydro column to increase octane. This octane enhancing catalyst isomerizes the some
of the olefms which increases the octane of this stream by about three octane numbers.
The seven-carbon and heavier petroleum compounds leave the bottom of the CDHydro
unit and are fed into the CDHDS column. There, the heavier compounds head down the column,
and the lighter compounds head up. Both sections of the CDHDS column have catalyst loaded
into them which serve as hydrotreating reaction zones. Similar to how hydrogen is fed to the
CDHydro column, hydrogen is fed to the bottom of the CDHDS column.
The temperature and pressure of the CDTech process columns are lower than Mobil Oil's
Octgain process. These operating conditions minimize yield and octane loss. While the CDTech
process is very different from conventional hydrotreating, the catalyst used for removing the
sulfur compounds is the same. Thus, if concerned about the reliability of the process, refiners
can look at the track record of the catalyst in conventional hydrotreating to get an indication of
its expected life, and then adjust that expectation based on the milder conditions involved. One
important different between the CDTech process and conventional hydrotreating is that CDTech
mounts its catalyst in a unique support system, while conventional catalyst is simply dumped
into the fixed bed reactor. Although the CDTech desulfurization process is different from
conventional hydrotreating processes, the use of a distillation column as the basis for the process
IV-39
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
is very familiar to refiners. Every refinery has distillation in its refinery, thus, refiners are very
skilled in its application.
6. Expected Desulfurization Technology to be Used by Refiners
If the proposed gasoline sulfur standard is finalized, refiners which produce gasoline
would have to meet the standard to be able to continue participating in the U.S. gasoline market.
As stated above, most refiners will have to invest in desulfurization technology. Arguably,
refiners would try to minimize the cost to their business. As stated above, the improved gasoline
desulfurization technology costs of CDTECH and Mobil Oil Octgain provide refiners a lower
cost option for meeting a gasoline sulfur standard. However, many refiners have shared with
EPA that they may be hesitant to use these improved, but recently developed technologies for
gasoline desulfurization. They claim that until the technologies have been installed in one or
more refineries and operated for a while, that there will continue to be a significant measure of
uncertainty. This uncertainty could tip the balance away from using these improved
desulfurization technologies, to applying more expensive conventional desulfurization
technology.
While there is a concern now on the part of some refiners about using these improved,
but not commercially tested desulfurization technologies, we believe that much of this concern
will dissipate shortly. Both processes are expected to be installed and operated in refineries later
on this year. Mobil Oil has an Octgain hydrotreater installed at its Joliet, Illinois refinery. Up to
now, though, only the second generation Octgain catalyst (Octgain 125) was demonstrated at
that site. However, Mobil Oil plans to load its new Octgain 220 catalyst in the hydrotreater at
Joliet to determine how it runs in a full scale hydrotreater. This experience should start to occur
sometime before midyear of 1999.19 CDTech has years of accumulated service with its
CDHydro unit. In its present service, certain olefms called dienes are reduced and octane is
improved. While this service is not intended for desulfurization, these units do in fact combine
sulfur compounds together much like how the process will operate in a desulfurization service.
Thus, this unit has extensive operational experience already. The CDHDS unit, on the other
hand, has not been installed in any refinery. It is to be installed in a Motiva refinery (which was
Star) in Port Arthur, Texas, with a start-up date sometime around the third quarter of 1999.
Another refinery, which is Transamerican in Louisiana, is also planning to install the CDHDS
unit, followed by the CDHydro unit for desulfurization, sometime in the first half of the year
2000.
7. Feasibility for a Low Gasoline Sulfur Standard in 2004
We believe that sufficient evidence exists which supports the conclusion that it is feasible
for the U.S. refining industry to meet the 30 ppm average standard in four years or less. We
discussed the possibility of meeting a stringent gasoline sulfur program with refiners. The
American Petroleum Institute (API) communicated to us that a minimum of four years is needed
between when a fuel regulatory requirement is promulgated and when the requirement must be
IV-40
-------
Chapter IV: Technological Feasibility
implemented.20 In API's comments on this issue, it provided a schematic of a typical refinery
project development timeline. This schematic showed that a "best estimate" of installation time
would be 4.5 years. However, if contingencies are considered, then 5 years would be required.
In more detail and without contingencies, API's project timing schedule without contingencies is
summarized here relative to date of the final rule (FR). The table also shows EPA's estimates
for the same process steps.
Table IV-13. Leadtime Required Between Promulgation of the Final Rule and
Implementation of the Gasoline Sulfur Standard (years)
Scoping Studies
Process Design
Permitting
Detailed Engineering
Field Construction
Start-up/Shakedown
EPA
Time for
Individual Step
0.5-1.0*
0.75-1.5
0.5-1.0
0.5-1.0
0.75-1.25
0.25-0.5
Cumulative
Time
0.5
1.25-2.0
1.75-2.5
2.0-3.0
2.25-3.5
2.5-4.0
API (without contingencies)
Time for
Individual Step
1.5*
1.5
1.5
1.25
1.5
0.75
Cumulative
Time
1.0
2.0
2.5
2.75
4
4.5
* Can begin before FRM
Scoping and screeening studies refer to the process whereby refiners investigate various
approaches to sulfur control. These studies involve discussions with firms which supply
gasoline desulfurization and other refining technology, as well as studies by the refiner to assess
the economic impacts of various approaches to meeting the sulfur standard. In the case of
gasoline desulfurization, a refiner would likely send samples of their FCC gasoline to the firms
marketing gasoline desulfurization technology to determine how well each technology removed
the sulfur from that particular type of FCC gasoline (e.g., sulfur removal efficiency, octane and
yield loss, hydrogen consumption, etc.).
API projects that these studies can start 6 months prior to the final rule and would
continue for a year after the final rule. Based on discussions with both refiners and technology
providers, it is clear that many refiners have already been conducting these studies for at least a
year. We believe that by the time of the final rule, refiners will already have a very good idea of
the performance and economics of the various gasoline desulfurization technologies at the pilot
plant level. Some time will be required to process the details of the final rule. More
importantly, however, is that many of the new gasoline desulfurization technologies will be
demonstrated in actual refinery applications this year. Refiners naturally desire as much
demonstrated experience with any new technology as possible prior to investing significant
amounts of capital in these technologies. As discussed above, we believe that these new
IV-41
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
technologies are sufficiently similar to existing gasoline desulfurization technology that long-
term performance will not be an issue if the process operates as designed initially. Thus, we
believe that sufficient performance data will be available within 6 months after the final rule to
allow refiners to initiate process design. This is the same time period estimated by API, though
API estimates that the scoping studies will extend 6 months into the process design period.. As
this does not affect the total leadtime needed, the ability to modify the design during the process
design period simply gives a refiner more flexibility to optimize its design.
API then estimates that process design, permitting and detailed engineering will require
almost two and a half years. While not shown in the table, API estimates that the major
appropriation decision will be made two and a half years into the process, or just after receiving
the final permit approval and just before detailed engineering is completed.
The time required for process design will depend on the extent of the refinery
modifications planned. We expect that the great majority of refiners will hydrotreat their FCC
gasoline. If no existing equipment is used, this primarily involves building the hydrotreater and
its associated equipment (distillation columns, furnaces, pumps, compressors). The refiner
would also require a source of a hydrogen for the desulfurization unit. This could come from
hydrogen already being generated in the refinery, or from an outside source. In the extreme, the
refiner would have to build its own hydrogen plant. Finally, the refiner will have to ensure that
the hydrogen sulfide being generated from the desulfurization equipment can be processed in the
refinery's existing sulfur recovery plant. Given the small amount of sulfur being removed from
gasoline compared to the amount of sulfur already being processed in the refinery, this is likely
to be possible with little change to the sulfur recovery plant. However, some expansion could be
required in the extreme.
All of this equipment is already common to refineries. Gasoline desulfurzation units are
either very similar to existing distillation columns or gasoline and diesel fuel hydrotreaters
already being used in essentially every refinery. Hydrogen plants are widely used throughout the
refining and chemical industries and can be purchased from vendors as basically stand alone
units. The same is true for sulfur recovery plants. Also, design and construction time has been
reduced by up to 40 percent between 1991 and 1996 alone by computerized design and
improving construction scheduling using state of the art methods.21 For example, CDTECH
estimates that 10-12 weeks are needed for the basic process design of their equipment. While
this estimate for basic process design may not represent all the technologies which may be used,
even if the process design for these other technologies is 2 or even three times longer, this time
would still be significantly less than API's estimate. API does not mention any recent reductions
in required project schedules associated with computerization and standardization in design and
construction improvements. They do make comparisons to past leadtimes provided for RFG, but
only indicate that the leadtime required has increased, not decreased. Thus, it is not clear why
API is estimating the need for so much more time for process design and engineering.
It is possible that some refiners might decide to implement more major changes to the
refinery, such as adding a FCC feed hydrotreater. This equipment is more unique to each
refinery and could require some additional time to design. However, this equipment would
IV-42
-------
Chapter IV: Technological Feasibility
significantly reduce a variety of emissions from the FCC unit. This should ease permitting and
compensating for any emission increases elsewhere in the refinery. The upper limits of our
estimates shown in Table IV-13 apply to the limited number of these more major modifications
which might occur.
API also indicates that the time necessary for permitting has increased even since the
mid-1990's when refineries made modifications for RFG production. EPA recognizes that
permitting is a necessary step in the process and that it is often outside of the refiner's control.
EPA has committed to working with states and local agencies to streamline this process as much
as possible. We believe that the permitting can be reduced to as short as six months and to at
most a year.
According to a general estimate for construction time for about the 1980 time frame, 18
to 42 months should be allowed for large construction projects over $10 million, which could
range up to the installation of an entire refinery or plant.22 Smaller projects are projected to
require substantially less time. The projected capital cost of gasoline desulfurization is near the
lower end of the cost range cited, considering inflation. Thus, the 18 month estimate should be
considered applicable, which is the same as API's estimate. However, this estimate was well
prior to the 35-40% reductions cited above. Also, the construction time for those refiners
planning to install demonstration units of the advanced desulfurization technology are well
below 18 months. Thus, we estimate the time of construction to be a year, plus or minus three
months.
Overall, shortening the permitting and basic process design has a dramatic impact on
API's estimated four year time for complying. Based on this analysis, we estimate that refiners
should be able to complete the process of designing and installing gasoline desulfurization
hardware, and make other refinery changes, in about 2!/2 to four years. Again, the upper end of
this range should only apply to a small number of refiners which will be making major changes
to their refinery configuration as a result of gasoline desulfurization.
API estimates that up to an additional year is needed for contingencies. Those refiners
which would need that extra time will have added flexibility to meet the proposed program
implementation date by participating in the averaging, banking and trading program. Those
which can meet the proposed implementation date sooner will be able to generate credits which
can be traded to those which need more time to comply. For those refiners which would like to
install their own gasoline desulfurization processing units, but are not yet very far along, the
trading program would potentially allow them more time to develop their processes.
Several different fuel programs already in place suggest that a stringent gasoline
desulfurization program can be phased-in sooner than what API claims. The California sulfur
control program which was promulgated in June of 1975, started to phase in only six months
after promulgation, and was fully phased in 41/2 years later.23 Similarly, the Phase II California
Reformulated Gasoline Program was promulgated in November 1991 and took effect about 41/2
years after promulgation.24 However, in addition to a stringent sulfur control standard, refiners
also had to meet stringent controls for aromatics, olefms, Reid vapor pressure, and distillation
IV-43
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
index. Thus, if California refiners began to meet the low sulfur program within six months, and
the very stringent RFG requirements in 41/2 years, then, this argues that if only a stringent
gasoline sulfur standard were to be finalized, that most refiners should be able to meet that
requirement in less than four years. Also, because the refining industry already has extensive
experience with meeting the California low sulfur requirement, it likely could meet a similar
standard sooner.
The On-Road diesel sulfur rulemaking provides an example of refiners meeting a much
shorter compliance period. Refiners nationwide met the on-highway low sulfur diesel standards
in three years time; since the rulemaking was promulgated August 1990 and took effect October
1993.25 That rulemaking required refiners to reduce diesel sulfur levels from over 2000 ppm
down to under 500 ppm. Diesel hydrotreaters are fixed bed hydrotreaters which, as described
above, are essentially the same design as Octgain units
For the Reformulated Gasoline Program, EPA proposed to give refiners 4 years to meet
the Complex Model requirements of the Reformulated Gasoline program. We felt that 4 years
was necessary so that refines could take time to understand how to most cost-effectively use the
Complex Model, and to install whatever capital which needed to be installed. However, this
rulemaking specifies a single specification and not require the use of a complex emissions
model.
Refiners will have a minimum of four years leadtime to comply with the gasoline sulfur
standard if the final rule is promulgated by the end of 1999. However, the sulfur averaging
banking and trading program provides up to an extra two years for those refiners participating in
the program. Thus, the leadtime provided in this proposal should be sufficient, with one
exception.
Small refiners may need more time to comply with a sulfur control program. Small
refiners generally have a more difficult time obtaining funding for capital projects, and must plan
further in advance of when the funds are needed. We contracted a study of the refining industry
which included assessing the time required for small refiners to obtain loans for capital
investments. The simple survey revealed that small refiners would need two to three months
longer than large refiners to obtain funding. If small refiners are forced to or prefer to seek
funding through public means, such as through bond sales, then the time to obtain funding could
be longer yet, by up to one third longer.26 In addition, because of the more limited engineering
expertise of many small refiners, the design and construction process for these refineries is
relatively more difficult and time consuming. We also think that the contractors which design
and install refinery processing units will likely focus first on completing the more expensive
upgrade projects for large refiners. Thus the design and construction of desulfurization hardware
in the refinery would take longer as well. For this and other economic reasons, we are proposing
to delay the implementation of the low sulfur program for small refiners.
8. Phase In of Compliance with the Proposed Sulfur Standards
IV-44
-------
Chapter IV: Technological Feasibility
In the previous section, we estimated that it would take 2.5-4 years (30-48 months) to
design and construct sulfur desulfurization equipment. The lower end of this range was more
relevant to those refiners choosing to desulfurize FCC gasoline, while the upper end of the range
would more likely apply to more significant changes, possibly involving the construction of a
FCC feed hydrotreater and changes to other units in the refinery to accommodate a change in
crude oil which the FCC feed hydrotreater enabled.
The proposed gasoline sulfur requirements begin on October 1, 2003. On October 1,
2003, the per gallon cap of 300 ppm takes effect, followed by similar caps of 180 ppm and 80
ppm which take effect on January 1, 2005 and January 1, 2006, respectively. Also, refiners'
actual gasoline production must contain less than 120, 90 and 30 ppm sulfur on average during
calendar years 2004, 2005 and 2006 (and beyond). Finally, refiners must also meet a refinery
average sulfur standard of 30 ppm starting on January 1, 2004.
Between January 1, 2004 and December 31, 2005, the 30 ppm refinery average standard
can be met using credits generated or purchased. Between January 1, 2000 and December 31,
2003, credits can be generated by selling gasoline (conventional or reformulated) that contains
no more than 150 ppm sulfur. The credit is the difference between the sulfur content of the fuel
produced and the refinery's 1997-1998 sulfur baseline in the case of conventional gasoline or
wintertime RFG, multiplied by the volume of gasoline produced. In the case of summertime
RFG, the credit is the difference between the sulfur content of the fuel produced and 150 ppm,
again multiplied by the volume of gasoline produced. Starting in 2004, credits (and debits) are
the product of the difference between the sulfur content of the fuel produced (whether
conventional or reformulated) and 30 ppm.
The phase in of both the per gallon sulfur caps and the actual average sulfur standards
between 2004 and 2006 and the ability to generate sulfur credits starting in 2000 is intended to
allow refiners to stagger their investment in desulfurization technology while protecting Tier 2
vehicles from unreasonable sulfur levels and getting substantial emission reductions from both
Tier 2 and existing vehicles. In order to estimate the rate at which individual refineries must
install desulfurization equipment, we estimated each refinery's current sulfur level and compared
this to the proposed requirements. We based our estimates of refinery's current sulfur levels on
these refinery's official 1990 baselines, which were developed in support of the EPA's RFG and
antidumping programs. These 1990 baselines were updated using 1997 data which refiners
submit, again as part of the RFG and antidumping programs. When the 1997 data were available
for an individual refinery, it was used in lieu of the 1990 baseline. However, in some cases,
1997 data were only available for all of a corporation's refineries on an aggregate basis. In this
case, each refinery's 1990 sulfur level was multiplied by the ratio of the refiner's 1997 corporate
average sulfur level to its 1990 corporate average sulfur level. This procedure assumes that any
change in sulfur between 1990 and 1997 occurred proportionately at each refinery. In a few
cases, 1997 data were not yet available for a specific refiner, so the 1990 baseline was used
unchanged.
We used these sulfur estimates along with estimates of each refinery's gasoline
production to develop a profile of the industry. We grouped refineries by current sulfur level
IV-45
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
and determined the amount of gasoline produced by non-California refineries in each sulfur
grouping. Refiners meeting the definition of a small business were also excluded, as they would
have more time to comply with the proposed sulfur requirements than other refiners. The results
of this analysis are shown in Table IV-14.
Table IV-14. Distribution of Refineries by Current Gasoline Sulfur Level
Range of Gasoline Sulfur (ppm)
Average Baseline Sulfur (ppm)
Percentage of total fuel produced
Number of Refineries
Refinery Groupings
I
0-40
17
2%
10
2
41-180
105
26%
33
3
181-300
246
26%
25
4
>300
476
47%
39
Starting with this information, we projected the actions which refiners could take in
meeting the proposed gasoline sulfur requirements. This analysis revealed that the per gallon
caps of 300 and 180 ppm applicable nominally in 2004 and 2005 were the most constraining
features of the proposal in these early years of the program. Thus, this discussion will begin
with compliance with those caps, followed by a discussion of compliance with the absolute
corporate average standards of 120 and 90 ppm in 2004 and 2005 and compliance with the 30
ppm standard, which can be met with credits.
A refiner can take two basic types of actions to meet the 2004 and 2005 sulfur caps. It
can make operational changes without adding or substantially modifying its current refining
equipment or, it can add new capital equipment. Operational modifications include switching to
a low sulfur crude oil, operating existing FCC feed and naptha hydrotreaters more severely,
debottle-necking these units to process more volume, and shifting the heaviest portion of FCC
gasoline (which contains a disproportionate share of FCC gasoline sulfur) to the distillate pool.
As an alternative to the last approach, a refiner could hydrotreat this heavy FCC gasoline in its
distillate hydrotreater, desulfurize it, separate it from the hydrotreated distillate and then reblend
the desulfurized FCC gasoline back into the gasoline pool. We believe that these techniques
would enable refiners to reduce sulfur levels in the near term. However, the degree of this
reduction is difficult to estimate. Also, refiners have to meet the Phase 2 requirement of the RFG
program in 2000, which is generally expected to require RFG to contain roughly 150 ppm sulfur
during the summer months.
Very few refiners have built new gasoline desulfurization equipment in order to meet this
requirement. Thus, they appear to be planning to meet the 150 ppm level with operational
changes or by reblending. Reblending would involve the production of two distinct gasolines
within the refinery, one with lower sulfur levels for the RFG market and another gasoline with
higher sulfur levels for the conventional gasoline market. The antidumping requirements which
apply to non-RFG gasoline implicitly limit the degree to which this can occur, but it still possible
that many refiners will use blending strategies to at least partially meet the Phase 2 RFG
requirements. In any event, the Phase 2 RFG requirements are likely to utilize some of refiner's
IV-46
-------
Chapter IV: Technological Feasibility
existing ability to reduce gasoline sulfur without new capital equipment. This, plus the
uncertainty in the degree that operational changes can reduce sulfur levels, led us to believe that
refiners would not rely on operational changes as the primary means of meeting the 2004 and
2005 per gallon sulfur caps of 300 and 180 ppm. As a result of this, for the purpose of this
analysis, we assumed that any major reductions in gasoline sulfur needed to meet the per gallon
caps, as well as the eventual 30 ppm standard, would be met through the construction of new
desulfurization equipment.
This uncertainty in the way in which refiners will meet the Phase 2 RFG requirements
also led us to assume that the pool average sulfur level will not decrease in 2000 due to these
requirements. In fact, some reduction is likely. However, this reduction would only serve to
delay the timing of investments in desulfurization equipment. In this respect, the projections
made below probably overestimate the number of refineries which must invest in the near term
to some degree.
In most cases, the overall costs of desulfurization equipment is minimized when its
construction is conducted in one step (enabling the production of 30 ppm gasoline), as opposed
to a two step process involving some intermediate sulfur level. Thus, we also assumed that if a
refinery had to install new equipment to meet a particular sulfur standard, it built equipment
sufficient to comply with the 30 ppm standard and 80 ppm cap. There are some situations,
particularly those involving higher baseline sulfur levels, where the equipment needed to meet
the final 30 ppm standard involves more than one desulfurizing unit. In this case, it may be
possible to install the major unit prior to October 1, 2003 and meet the 2004 requirements and
delay the other unit(s) until 2005 or 2006. This possibility was not accounted for here.
Sulfur group #4 in Table IV-14 shows that 39 refineries have current sulfur levels above
300 ppm. These refineries must reduce their gasoline sulfur levels to 300 ppm or less by
October 1, 2003, due to the per gallon cap taking effect at that time. In accordance with the
above premises, the 39 refineries in this group would require fully operational desulfurization
equipment by this date. In fact, this is likely an overestimation. Some of the refineries in this
group are within 10-15 ppm of the 300 ppm cap. It is likely that these refineries could reduce
their sulfur levels by a few ppm without installing new equipment. However, regular grade
gasoline tends to contain more sulfur than premium grade. Meeting a per gallon cap with regular
grade gasoline may require the refinery average sulfur level to be somewhat below 300 ppm.
Also, refineries do not produce identical gasoline each day; there is some variation in the quality
of gasoline produced throughout the year. Since the 300 ppm cap applies to each and every
batch of gasoline produced, this variation also requires that the average sulfur level be below 300
ppm.
This implies that some of the refineries in Group #3 would also have install
desulfurization units to meet the 300 ppm cap. However, this ignores the possibility of sulfur
reductions from operational refining modifications and reducing historic variation in gasoline
quality given an economic incentive to do so. We project that these more modest modifications,
along with improved blending techniques, would be sufficient for those refineries currently near
the 300 ppm level to produce regular and premium grades of gasoline under the 300 ppm cap.
IV-47
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Thus, we project that 39 desulfurization units would have to be installed and operating
prior to October 1, 2003. We also project that six of these units would be installed by mid-2002,
12 would be installed by January 1, 2003 and the remaining 21 units would be installed by
October 1, 2003. The projection of the 6 early units reflects the fact that technology vendors
wishing to license their technology to refiners will need to demonstrate this technology early in
the process. EPA projects that at least six technologies will be competing in this timeframe. We
are already aware of at least three plants which will be in operation during 2000 or early 2001,
two CDTech units and one Mobil Octgain unit. Assuming that EPA issues its final Tier 2/Sulfur
rule by the end of 1999, mid-2002 is exactly 30 months later, which is generally the minimum
time EPA projects for installing such equipment. Of course, given the commercial purposes for
constructing and operating this equipment, plans could begin well before the final rule is signed,
as indicated by the two projects mentioned above.
The 12/21 split between the 2003 and 2004 units was based on the premise that some
refiners would want to generate credits for use in their other refineries in 2004 and 2005 or for
sale to other refiners. We also expect that refiners, process design firms, and construction firms
would want to spread out the design and construction of this new equipment as much as possible.
Installing this new equipment within 36 months of the final rule (i.e., by January 1, 2003) is
certainly feasible, given the projections made in the previous section. Thus, EPA believes that
there would be an economic incentive to begin construction at some refineries sooner than at
others.
Using the same assumptions, those refineries with current sulfur levels between 180 and
300 ppm would have to install equipment for use by January 1, 2005. As shown in Table IV-14,
this group includes 25 refineries. Finally, the 30 ppm standard effective in 2006 would require
another 33 refineries to install desulfurization equipment. Ten refineries out of the total of 107
refineries located in the U.S. outside of California already have sufficiently low sulfur levels to
avoid the need for new equipment.
Table IV-15 shows the effect of the phase-in of this new equipment on average sulfur
levels in the non-California U.S. gasoline pool. As can be seen, pool average sulfur levels in
2004 and 2005 are 105 and 49 ppm, respectively, well below the proposed 120 and 90 ppm
corporate average standards. Thus, no additional controls should be required to meet these
average standards if sufficient credits are available either from within the corporation or on the
open market. This conclusion presumes that refineries with current sulfur levels in between the
corporate average standard and the per gallon cap (e.g., 120-300 ppm in 2004) are a part of a
corporation which has another refinery at 30 ppm, so that its actual corporate average sulfur level
is below 120 or 90 ppm, as applicable, since these standards must be met without the use of
credits.
IV-48
-------
Chapter IV: Technological Feasibility
Table IV-15. Effect of Phase-In of Sulfur Control to Meet Sulfur Caps in 2004 and 2005
Pool Average
Sulfur Level
(ppm)
Number of
Refineries
Building Sulfur
Units
Percentage
Controlled to 30
ppm
2000
312
0
0
2001
312
0
0
2002
296
6*
7%
2003
216
12
22%
2004
105
21
47%
2005
49
28
73%
2006
30
30
100% **
Credits (ppm over the entire U.S. gasoline pool for an entire year)
Credits Generated
from Winter RFG
Credits Generated
from Sulfur Units
Credits Used to
Allow Sale of >30
ppm Gasoline
Cumulative Credit
Balance
20
0
0
+20
20
0
0
+40
20
16
0
+76
20
96
0
+192
0
0
-75
+117
0
0
-19
+98
0
0
0
+98
**
Units only operate during second half of 2002.
Includes 2% of fuel which is currently below 40 ppm sulfur.
This table also shows the sulfur credits generated and used by refineries between 2002
and 2006. As can be seen, 112 ppm of credits were generated by the 18 units operating before
January 1, 2004. This amount is greater than the 94 ppm of credits which are needed by the 58
refineries which did not begin operating sulfur units until 2005 or 2006.
Table IV-15 also shows a series of 20 ppm credits under the heading of wintertime RFG.
The proposed credit provisions allow winter RFG to generate sulfur credits if its sulfur level is
below the individual refinery's baseline sulfur level, which on average is 312 ppm. Summer
RFG is expected to contain 150 ppm sulfur or less. EPA projects that the cost of extending this
control (or reblending) to the winter months would be minimal and that refiners would likely
choose to generate credits in this way. The 20 ppm per year of credits represents the difference
between the U.S. pool average sulfur level (312 ppm) and the 150 ppm level (difference of 162
ppm), multiplied by a factor of one-half, since only winter fuel would receive the credit and by a
IV-49
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
factor of 0.25, since RFG is roughly 25 percent of the non-California gasoline market.
Given this expectation of significant RFG-related credits, EPA projects that a significant
excess of credits would be available. This analysis assumes that all credits which are generated
are either used within a corporation or are put up for sale on the open market. This may not
occur in all cases. However, this analysis indicates that only about half of the credits expected to
be generated are actually needed to allow refineries with current sulfur levels below either 300
ppm or 180 ppm to delay installation of desulfurization units until 2005 or 2006, respectively. It
is also possible that many more credits would be available, if many refineries use operational
techniques to reduce sulfur to levels below the 150 ppm trigger level before the time that they
install desulfurization units. On the other hand, fewer credits would be available if refineries
adopt an alternative schedule to phase-in desulfurization capacity (because of capital constraints
or site specific factors affecting individual control decisions) than the schedule posited by this
analysis. As noted above, for example, some refineries may be close enough to the 300 ppm cap
to be able to reduce their sulfur levels to meet the cap without installing new equipment. There
may also be cases where refineries have such high baseline sulfur levels that more than one
desulfurization unit would be required to produce 30 ppm sulfur gasoline on average. In these
cases a decision to phase-in these units rather than installing 30 ppm technology initially would
reduce the number of credits generated early. However, even if fewer refineries install
desulfurization equipment to generate early credits than our analysis assumed, we believe there
will be sufficient credits generated from winter RFG and from early installation of some 30 ppm
technology to fully cover the credits needed by the industry.
IV-50
-------
Chapter IV: Technological Feasibility
Chapter IV. References
1. SAE paper 1999-01-0774, "Using Advanced Emission Control Systems to Demonstrate
LEV IIULEV on Light-Duty Gasoline Vehicles," Webb, et.al.
2. Baseline Submissions for the Reformulated Gasoline Program.
3. Swain, Edward J., Gravity, Sulfur Content of U.S. Crude Slate Holding Steady, Oil and
Gas Journal, January 13, 1997.
4. Upson, Lawrence L, Schnaith, Mark W., Low-sulfur Specifications Cause Refiners to
Look at Hydrotreating Options, Oil and Gas Journal, December 8, 1997.
5. Final Report, 1996 American Petroleum Institute/National Petroleum Refiners
Association Survey of Refining Operations and Product Quality, July 1997.
6. Final Report, 1996 API/NPRA Survey of Refining Operations and Product Quality.
7. Standard Specification for Spark-Ignition Fuel, D 4814-92c; American Society for
Testing Materials.
8. California Code of Regulations, Title 13, §2262.2.
9. Final Report, 1996 API/NPRA Survey of Refining Operations and Product Quality.
10. Davey, Steven W., Haley, John T., FCC Additive Technology Update, 1996 O&G J
International Catalyst Conference and Exhibition, Houston Texas, February 1996.
11. Final Report, 1996 API/NPRA Survey of Refining Operations and Product Quality.
12. Pacheco, Michael; Symposium for Diesel Desulfurization, Houston, Texas, February
1999.
13. Petroleum Refinery Process Economics, Maples, Robert E., PennWell Books, Tulsa,
Oklahoma, 1993.
14. Podar, Syamal K., Hilbert, Timothy L., Octgain, Evaluation for the Manufacture of
Reformulated Gasoline via LP Modeling, NPRA 1995 Annual Meeting.
15. Mobil Technology Company, FCC Gasoline Desulfurization Reaches a New
Performance Level, Sulfur 2000, A Special Edition, Hart's Fuel Technology and
Management, Summer 1998.
16. CDTECH, FCC Gasoline Sulfur Reduction, CDTECH, Sulfur 2000, Hart's Fuel and
Technology Management, Summer 1998.
IV-51
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
17. Rock, Kerry J., Putman, Hugh, Global Gasoline Reformulation Requires New
Technologies, Presented at Hart's World Fuels Conference, San Francisco, March 1998.
18. Rock, Kerry L., et al, Improvements in FCC Gasoline Desulfurization via Catalytic
Distillation, Presented at the 1998 NPRA Annual Meeting, March 1998.
19. Personal conversation with Trig Tryjankowski, Manager, Licensing, Mobil Technology
Company, January 1999.
20. Letter from Bill O' Keefe, American Petroleum Institute, to Bruce Jordan, Office of Air
Quality Planning and Standards, EPA, February 12, 1999.
21. Shanley, A., Benchmarking for the Next Century, Chemical Engineering, April 1996.
22. Chemical Engineers Handbook, Perry, Robert H., McGraw Hill Book Company, New
York, 1984
23. California Code of Regulations, Title 13 §2252.
24. California Code of Regulations, Title 13 §2260 - §2272.
25. 55 FR 34138, August 21, 1990.
26. Refining Industry Profile Study; EPA contract 68-C5-0010, Work Assignment #2-15,
ICF Resources, September 30, 1998.
IV-52
-------
Chapter V: Economic Impact
Chapter V: Economic Impact
A. Impact of Tier 2 Standards on Vehicle Costs
This section presents a detailed analysis of the vehicle-related costs we estimate would be
incurred by manufacturers and consumers as a result of the Tier 2 standards. Section B. of this
Chapter presents cost estimates for fuels changes. For manufacturers, the economic impact of
the Tier 2 standards would include incremental costs for various vehicle hardware components,
as well as up-front costs for research and development (R&D), certification, and facilities
upgrades. Impacts on consumers would include increases in vehicle purchase price and changes
in vehicle operating costs. Finally, this section provides estimates of the annual nationwide
aggregate costs for Tier 2 vehicles.
1. Manufacturer Costs for Tier 2 Vehicles
a. Methodology
This section A.I. discusses EPA's estimates of costs to manufacturers for Tier 2 vehicles,
including both hardware and developmental costs. The estimates are based on projections of
technology changes we consider most likely to be used by manufacturers to comply with the Tier
2 standards. To estimate costs, we have analyzed two sets of technologies for each vehicle class
and engine type, a baseline technology package and a Tier 2 technology package. We used as a
baseline, projected NLEV technologies for LDVs, LDTls, and LDT2s, and Tier 1 technologies
for LDT3s and LDT4s. These are the standards that vehicles will be meeting in 2003.' We have
estimated the baseline technology packages based primarily on California Air Resources Board
technology analyses done in support of the California LEV program,1 with adjustments based on
discussions with manufacturers about trends in technology.
The following analysis projects a relatively uniform emission control strategy for various
LDV and LDT models. However, this should not suggest that a single combination of
technologies would be used by all manufacturers. Selecting technology packages requires
extensive engineering judgement and EPA does not know future technology mixes and costs
with certainty. New technological developments could significantly change the approach
manufacturers would take to meet the standards. In addition, there are several emissions control
technologies and several manufacturers of each. The Technological Feasibility portion of this
RIA details many of the available technologies. Each manufacturer will choose the mix of
technologies best suited for their vehicles. Manufacturers would have as many as eight years for
R&D for some vehicles due to the phase-in schedule. We expect a large R&D effort involving
extensive systems optimization to find the most cost effective mix of technologies for particular
1 Even though the NLEV program ends in the Tier 2 time frame, we have not included the NLEV program
in our Tier 2 analysis, since we have analyzed and adopted NLEV previously.
V-l
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
vehicle lines.
Nevertheless, we believe that the projections presented here provide a cost estimate
representative of the different approaches manufacturers may ultimately take. Clearly, there are
key technologies that manufacturers will likely use to meet the standards in most cases. We
expect Tier 2 standards would be met through refinements of current emissions control
components and systems rather than through the widespread use of new technologies. Current
LDV and LDT certification levels also suggest this approach makes sense. We have made a best
estimate of the combination of technologies that any manufacturer might use to meet the
proposed standards at an acceptable cost and these technologies form the basis of the cost
estimates. Since California, in their LEVII program, has adopted essentially the same standards
and time-line that EPA is proposing, we used California's technology and cost analyses as a
source of information.2 We also had several conversations with equipment and vehicle
manufacturers whose input we also used for these analyses. Most manufacturer input is
considered confidential business information and therefore is not described in detail.
We have not specifically analyzed smaller incremental changes in technologies which
might occur due to interim standards between the baseline and the Tier 2 standards. For LDVs
and LDT Is, the interim standards are a continuation of NLEV and therefore are equivalent to the
baseline standards. For LDT2s, given the state of technology on current vehicles, we expect
only minor changes in response to the interim standards. Many engine families are already
certified at levels meeting the interim standards. In addition, broad averaging would be available
which manufacturers could use in the early years of the phase-in when significant numbers of
LDVs and LDT Is are also in the averaging program for the interim standards.
In 2006, when LDT2s may make up the large majority of vehicles remaining in the
interim program manufacturers could use credits from model years 2004/2005 to comply with
the interim standards. If this is not an option, we expect manufacturers could make a few minor
modifications which would result in needed reductions. Most likely, the standards could be met
through calibration changes which entail changes to software. These changes would not involve
hardware or tooling changes. The R&D costs associated with these changes are already included
in the relatively large R&D costs included for the program as a whole. In addition there are
likely to be incremental improvements in the standard catalyst system for these vehicles due to
progress made by catalyst manufacturers. These incremental improvements in washcoat
technology are part of the normal progression of technology and would not likely result in an
increase in the catalyst cost due to the competitiveness of the catalyst industry.
For LDT3s and LDT4s, there is a phase in to an interim fleet average NOx standard of
0.20 g/mile with an accompanying NMHC average of 0.156 g/mile. Vehicles have their
emissions capped at 0.60 g/mile NOx and 0.23 g/mile NMHC. Most engine families currently
meet the caps. EPA expects that manufacturers could apply calibration changes and incremental
catalyst improvements, as noted above for LDT2s, where necessary to ensure compliance with
the caps. In addition, much of the R&D will have already taken place due to the California
program which includes the same standards (MDV2 standards) for pre-2004 model year LDT3s.
We do not expect these changes to result in increases to the cost of the program.
V-2
-------
Chapter V: Economic Impact
For the interim fleet average NOx standard, (average standard of 0.2 g/mile NOx with a
NMHC standard of 0.156 g/mile), the approaches noted above may not be adequate in some
cases. For vehicles well above the standard, manufacturers could redesign the vehicles to meet
the interim standards. However, we believe it is more likely that manufacturers would phase
these vehicles into the interim standards later in the phase-in period and use the program
averaging flexibility to meet the interim standard. Therefore, rather than project a cost for
vehicles to meet the interim standards, we have projected sales of Tier 2 vehicles prior to 2008 to
average with and off-set those exceeding the interim standards. We believe this approach is
reasonable considering manufacturers are likely to avoid significant R&D efforts to meet a
standard that is in effect for only a few model years. Essentially, a few such vehicle models
would have to be immediately redesigned to meet Tier 2 levels. Due to timing considerations,
manufacturers are more likely to focus their resources on meeting the Tier 2 standards.
Vehicle phase-in estimates are needed to project annual aggregate costs during the phase-
in period. For both phase-in periods (for LDVs, LDTls, LDT2s, and for LDT3s, LDT4s), EPA
has modeled that manufacturers will start the phase-in of Tier 2 standards with lighter vehicles
and work their way to heavier vehicles until all vehicles up through LDT4s meet the Tier 2
standard in 2009. The phase-in projections described in further detail in section A.3., below.
Costs to the manufacturer are broken into variable costs (for hardware and assembly
time) and fixed costs (for R&D, retooling, and certification). EPA projected costs separately for
LDVs, the different LDT classes, and for different engine sizes (4, 6, 8-cylinder) within each
class. Cost estimates based on the projected technology packages represent expected
incremental variable and fixed costs for vehicles in the near-term, or during the first years of
implementation.. For the long term, we have identified factors that would cause cost impacts to
decrease over time. The analysis incorporates the expectation that manufacturers and suppliers
will apply ongoing research and manufacturing innovation to making emission controls more
effective and less costly over time. Also, we project that fixed costs would be recovered over the
first five years of production, after which these costs would be recovered. These factors are
discussed in further detail below.
b. Hardware Costs for Exhaust Emissions Control
The following section briefly describes each of the technologies EPA has included in the
cost analysis and their costs incremental to the baseline use of the technology. Tables V-l
through V-5 at the end of this section provide the complete detailed projection of hardware
changes and costs for each vehicle and engine type. A breakdown of the hardware costs for the
evaporative system follow in section A. I.e. The Technological Feasibility portion of this RIA
provides further detail on the technologies included in the cost analysis, as well as others that are
less likely to be used to meet Tier 2 standards. The costs presented in this section are near-term
costs, during the first few years of production. Long-term hardware costs are discussed in a
following section.
Manufacturers are likely to use a systems approach to meeting the Tier 2 standards and
V-3
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
much of the effort will be in optimizing how the various components and subsystems (engine,
catalyst, fuel system, etc.) interact to achieve peak emissions performance. Some of these items
are included as part of the technology discussions below. However, there are no hardware costs
associated with these changes. The costs of optimization and calibration are part of a significant
R&D effort EPA anticipates will be necessary to meet the Tier 2 standards.
/'. Catalytic Converter System
The catalytic converter system is central to meeting current standards and improvements
to the systems will be critical in meeting Tier 2 emissions standards. EPA projects that all Tier 2
LDVs and LDTs will be equipped with advanced catalysts. Catalyst manufacturers are currently
working with engine manufacturers on new catalyst systems. To determine the cost increases
due to improved catalyst systems, we first analyzed current Tier 1 and NLEV systems for the
baseline and then projected what changes may be necessary to meet Tier 2 standards.
EPA first determined an average catalyst system for the baseline vehicles. Catalyst
systems vary in size and configuration due to factors such as engine size and emissions levels,
vehicle packaging constraints, cost, and manufacturer preference. Catalyst systems typically
consist of single or dual units (main or underfloor catalysts) and may also include one or two
smaller catalysts placed close to the engine (close coupled). For the baseline, we examined the
total volume, precious metal loading, and architecture of the main, or underfloor catalysts to
derive an average baseline catalyst for the various vehicle types and engine sizes. We also noted
whether or not vehicles were also equipped with additional close coupled catalysts.
After establishing baseline catalyst systems, we then projected changes to the catalyst
system for the Tier 2 analysis. In general, manufacturers could meet the standards by using very
large catalysts with relatively high precious metal loading. Many of the test programs that have
been conducted to demonstrate the feasibility of very low standards have featured vehicles with
such catalyst systems. However, based on uniform input from catalyst manufacturers, this is not
the approach we expect manufacturers to take in meeting the Tier 2 standards. Catalyst
manufacturers anticipate that improvements to the catalyst systems design, structure, and
formulation will also play a critical role in reducing emissions. These improvements are aimed
at decreasing emissions while minimizing the increase in catalyst volume and precious metal
loading. Manufacturers are working on these catalyst systems today.
We do expect some increase in average catalyst size (volume) and precious metal
loading. We believe that it is reasonable to expect catalyst systems to be sized such that the
underfloor catalyst volume will be equal to engine displacement and that loading will increase by
about 10 percent. Perhaps of equal importance will be the R&D efforts on the vehicle
manufacturers part to optimize engine performance and control systems so that the catalyst can
function at peak efficiency. Additional information on catalyst test programs and catalyst
changes is available in the Technical Feasibility Section of this RIA.
For the main or underfloor catalysts, EPA projects that improvements to the catalyst
V-4
-------
Chapter V: Economic Impact
architecture and formulation will increase catalyst costs by $2.44 to $6.59, depending on the
vehicle and engine type. These improvements include double layer washcoats and increasing the
cell density of the catalyst substrate to 600 cells per inch (cpi). We estimate that increases in the
catalyst volume and precious metal loading will account for the largest portion of the catalyst
cost increase due to the high cost of precious metals. We anticipate the change in catalyst
volume to cost between $10.00 and $55.00 per vehicle. We derived the increased volume cost
by taking the baseline cost of the catalyst per liter ($50/liter) and multiplying by the increase in
catalyst volume. Larger catalyst volume increases are projected for 6-cylinder engines in LDT
applications than for 8-cylinder engines due to relatively low baseline catalyst volumes for 6-
cylinder engines. We projected an increase in precious metal loading, in addition to the
increased volume, at a total cost of between $1.84 and $11.26 per vehicle. The details of the
underfloor catalyst cost estimates are provided in Table V-l.
V-5
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-l. Main or Underfloor Catalyst Cost Breakdown
Vehicle
TVDC
LDV
LOT
Engine
TVDC
4-cylinder
6-cylinder
8-cylinder
4-cylinder
6-cylinder
8-cvlinder
Sales wtd.
Engine
Displacement
(liter)
2.0
3.2
4.5
2.3
3.7
5.4
Projected
Baseline Cat.
Volume
(liter)
1.8
2.8
4.0
2.3
2.6
4.7
Projected
Tier 2 Cat.
Volume
(liter)
2.0
3.2
4.5
2.3
3.7
5.4
Increased
Volume
Cost (a)
(dollars)
10.00
20.00
25.00
0.00
55.00
35.00
Increased
Platinum
(Pt)
(grams)
0.000
0.000
0.000
0.000
0.035
0.082
Increased
Palladium
(Pd)
(grams)
0.000
0.000
0.000
0.000
0.540
0.550
Increased
Rhodium (b)
(Rh)
(grams)
0.085
0.138
0.194
0.097
0.157
0.229
Added
Pt cost
(dollars)
0
0
0
0
0.43
1.01
Added
Pd cost
(dollars)
0
0
0
0
5.17
5.28
Added
Rh cost (b)
(dollars)
1.84
2.95
4.14
2.10
3.41
4.97
Precious Metal Costs
$/troy ounce
Pt 384
Pd 300
Rh 675
$/gram
12.35
9.64
21.70
(a) Catalyst cost is $50/liter. Increased catalyst volume costs are the increase in catalyst volume multiplied by $50/liter.
(b) Increase in Rh of 1.2 g/cu ft
V-6
-------
Chapter V: Economic Impact
Close coupled catalysts are typically small relative to the main catalysts, under one-half
liter in volume. Their small size is due to packaging constraints associated with their location
close to the engine and their purpose, to warm-up quickly and reduce cold-start emissions. They
also typically have relatively high precious metal loading. Due to these factors, EPA is not
projecting changes to the close coupled catalysts, only changes in their usage. For NLEV
vehicles (LDV, LDT1 and LDT2), the percentage of baseline vehicles equipped with close
coupled catalysts is high, between 60 and 100 percent, depending on the vehicle and engine type.
We believe that the use of close coupled catalysts has likely peaked in these classes and we have
not projected increases in usage for Tier 2. For LDT3s and LDT4s, the use of close coupled
catalysts is currently low relative to the other classes. For Tier 2 LDT3s and LDT4s, we have
projected the use of close coupled catalysts to increase to be equivalent to the other vehicle
categories. The cost of dual close coupled catalysts are projected to be between $90 and $110,
for six and eight liter engines, respectively.
/'/'. Improved Fuel Control and Delivery
Precise fuel metering is critical to keeping the catalyst at peak operating efficiency.
Much of the effort for improved fuel control is in calibration and system optimization. For some
vehicles, EPA has included costs for hardware changes including improved exhaust gas oxygen
sensors and air-assisted fuel injection. There are two types of improved oxygen sensors that
EPA believes will be used increasingly for Tier 2 vehicles, universal exhaust gas oxygen sensors
(UEGO) and fast light-off or planar sensors. UEGO sensors are the most expensive type of
sensor and offer the most precise fuel control. However, only some manufacturers believe the
additional control is worth their higher incremental cost of 10 dollars. We believe more
manufacturers will opt for planar sensors, which offer a key advantage of quick warm-up,
allowing for precise fuel control sooner during cold starts. Many baseline vehicles also will
likely be equipped with planar sensors. The incremental cost of planar sensors is estimated to be
four dollars per sensor. We expect that the improved sensors would be used only before the
catalyst in the exhaust system for fuel control, with conventional heated exhaust gas oxygen
sensors used post catalyst for catalyst monitoring and additional fuel control.
Air assisted fuel injection is used to provide a better air fuel mixture to the engine, which
can be especially critical during engine warm-up. The technology can offer other advantages in
terms of engine performance which also makes it an attractive technology. For air assisted fuel
injection, the injectors must be redesigned to include a new adapter. We have projected that 50
percent of Tier 2 vehicles will be equipped with air assisted fuel injection at a cost of two
dollars for each improved injector.
As indicated above, much of the improvements in fuel control are likely to be
accomplished through system calibration. As such, they include software upgrade costs, rather
than hardware costs. EPA has included such costs in the R&D cost. These improvements may
include individual cylinder fuel control and adaptive learning.
V-7
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
/'/'/'. Secondary Air Injection
Manufacturers sometimes use a rich air/fuel mix during cold start to improve engine
performance and driveability. Secondary injection of air into exhaust ports after cold start when
the engine is operating rich can be used to promote combustion of unburned HC and CO which
results from the rich air/fuel mix. Air injection can also be used in conjunction with spark retard
to provide additional heat to the catalyst for quicker catalyst warm-up. EPA projects increased
use of electric air injection strategies for Tier 2 vehicles equipped with 6- and 8- cylinder
engines. The air injection systems consist of an electric air pump with integrated filter and relay,
wiring, an air shut-off valve with integrated solenoid, a check valve, tubing, and brackets. We
estimate the system cost to be 50 and 65 dollars for six- and eight- cylinder engines, respectively.
iv. Exhaust System Improvements
Manufacturers can insulate the exhaust system so the exhaust heat does not escape, but is
instead maintained within the system to promote catalyst warm-up. Improved materials include
laminated thin-walled exhaust pipes and double walled low thermal capacity manifolds (the two
walls have a small air gap between them that acts as an insulator). EPA estimates that improved
exhaust pipe costs one dollar per foot, with total system costs of between one and six dollars,
depending on engine size. Low thermal capacity manifolds are estimated to cost 20 to 40 dollars
depending on engine size. Due to the relatively high cost of these improvements, we have
projected manufacturers would use them only on LDTs, where it may be more difficult to meet
the Tier 2 standards. In some cases, manufacturers may be able to use exhaust system
improvements in lieu of adding close-coupled catalysts.
In addition, exhaust systems can be made leak-free which improves fuel control and
catalyst efficiency. As noted in the previous section, precise fuel control is critical to catalyst
performance and the oxygen sensor is a key element of fuel control. Air leaking into the exhaust
system can influence the oxygen sensor causing an improper fuel adjustment. Also, additional
air in the exhaust stream can lead to an oxidizing environment in the catalyst, diminishing the
catalyst's ability to reduce NOx. Leak-free systems include corrosion-free flexible couplings,
corrosion-free steel, and improved welding of catalyst assemblies. We estimate that many
baseline vehicles and all Tier 2 vehicles will be equipped with leak-free exhaust systems at an
incremental cost of 10 to 20 dollars depending on engine size.
v. Engine Combustion Chamber Improvements
Manufacturers may make a number of improvements to their engines as they are
redesigned, including adding a second spark plug to each cylinder, adding a swirl control valve
to improve mixing of air and fuel, or other changes needed to improve cold start combustion.
Engine changes are not likely to be uniform throughout the industry. EPA believes that
significant engine improvements for LDVs, LDTls and LDT2s are likely to have been made as
part of the effort to meet NLEV standards. The Tier 2 standards are not likely to drive a second
V-8
-------
Chapter V: Economic Impact
set of major changes to these engines. Therefore, EPA has not included an engine modification
cost for these vehicles. For LDT3s and LDT4s, which would be changing from Tier 1 to Tier 2
technology, we have included a hardware cost for engine modifications of $10 and $15 for six
and eight cylinder engines, respectively.
vi. Exhaust Gas Recirculation (EGR)
One of the most effective means of reducing engine-out NOx emissions is exhaust gas
recirculation. By recirculating spent exhaust gases into the combustion chamber, the overall air-
fuel mixture is diluted, lowering peak combustion temperatures and reducing NOx. Many EGR
systems in today's vehicles utilize a control valve that requires vacuum from the intake manifold
to regulate EGR flow. Some vehicles are being equipped with electronic EGR in place of
mechanical back-pressure designs. By using electronic solenoids to open and close the EGR
valve, the flow of EGR can be more precisely controlled. EPA projects that the use of full
electronic EGR systems will increase due to Tier 2 standards. We estimate that about 50 percent
of Tier 2 LDVs and LDTs will be equipped with electronic EGR at an incremental cost often
dollars per vehicle.
vii. Total Hardware Costs for Exhaust Emissions Control
Table V-3 provides a summary of the total hardware costs for each vehicle and engine
type. Tables V-3 through V-7 present detailed estimated manufacturer costs itemized for each
vehicle and engine type. The tables indicate EPA's estimate of the percentage of use of the
technologies for both the baseline and the Tier 2 vehicles. Some of the technologies listed, such
as individual cylinder fuel control and retarded spark timing, involve calibration changes only
and have no hardware costs associated with them.
Table V-2. Total Estimated Per Vehicle Manufacturer
Incremental Hardware Costs for the Tier 2 Standards
4-cylinder
6-cylinder
8-cylinder
sales weighted
LDV
($)
23.78
62.85
71.63
42.85
LDT1
($)
15.15
85.45
N/A
39.13
LDT2
($)
15.15
94.97
80.98
89.78
LDTS
($)
N/A
235.32
194.45
198.58
LDT4
($)
N/A
N/A
194.45
194.45
V-9
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-3. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDV Compared to NLEV LDV
Emission Control Technology
Universal Exhaust Gas Oxygen Sensor (UEGO)
Air-assisted fuel injection (a)
Individual cylinder fuel control (b)
Retarded spark timing at start-up (b)
Improved precision fuel control (c)
Faster microprocessor
Fast light-off exhaust gas oxygen sensor (planar)
Heat optimized exhaust pipe (d)
Leak-free exhaust system (e)
Engine modifications (f)
Full electronic EGR
Close-coupled catalyst
Underbody or main catalyst
Dual close-coupled catalyst
Dual underbody or main catalyst
Increased catalyst volume
Increased catalyst loading (Rh)
Improved double layer washcoat + 600 cpsi cell density
Secondary air injection (g)
Total Incremental Cost
4-Cylinder (53%)
Tech.
cost est.
(in dollars)
10.00
8.00
0.00
0.00
0.00
3.00
4.00
10.00
0.00
10.00
55.00
80.00
10.00
1.84
2.44
50.00
% of NLEV
vehs. that
use tech.
0
50
0
100
100
0
100
0
100
0
0
60
70
0
0
0
0
0
0
% Tier 2
that will
req. tech.
15
50
10
100
100
100
100
0
100
0
50
60
70
0
0
100
100
100
0
Inc. cost
over Tier 1
(in dollars)
1.50
0.00
0.00
0.00
0.00
3.00
0.00
0.00
0.00
0.00
5.00
0.00
0.00
0.00
0.00
10.00
1.84
2.44
0.00
23.78
6-Cylinder (39%)
Tech.
cost est.
(in dollars)
20.00
12.00
0.00
0.00
0.00
3.00
8.00
20.00
10.00
10.00
55.00
80.00
90.00
160.00
20.00
2.95
3.90
50.00
% of NLEV
vehs. that
use tech.
0
50
10
100
100
0
100
0
100
100
0
0
100
100
0
0
0
0
0
% Tier 2
that will
req. tech.
15
50
10
100
100
100
100
0
100
100
50
0
100
100
0
100
100
100
50
Inc. cost
over Tier 1
(in dollars)
3.00
0.00
0.00
0.00
0.00
3.00
0.00
0.00
0.00
0.00
5.00
0.00
0.00
0.00
0.00
20.00
2.95
3.90
25.00
62.85
8-Cylinder (8%)
Tech.
cost est.
(in dollars)
20.00
16.00
0.00
0.00
0.00
3.00
8.00
20.00
15.00
10.00
55.00
80.00
110.00
160.00
25.00
4.14
5.49
65.00
% of NLEV
vehs. that
use tech.
0
50
10
100
100
0
100
0
100
100
0
0
60
80
40
0
0
0
10
% Tier 2
that will
req. tech.
15
50
10
100
100
100
100
0
100
100
50
0
60
80
40
100
100
100
50
Inc. cost
over Tier 1
(in dollars)
3.00
0.00
0.00
0.00
0.00
3.00
0.00
0.00
0.00
0.00
5.00
0.00
0.00
0.00
0.00
25.00
4.14
5.49
26.00
71.63
(a) Air assisted fuel injection requires minor redesign of the idle air control valve at no additional cost and addition of an adapter to each injector at a cost of $2 each.
(b) Improved precision fuel control envisioned here and retarded spark-timing at start-up constitute software changes only, at no additional hardware cost.
(c) Improved precision fuel control constitute software changes only, at no additional hardware cost.
(d) Length of heat optimized exhaust pipe required is estimated to be one foot for 4-cylinder engines, four feet for 6-cylinder engines, and six feet for eight-cylinder engines, at a cost of $1 per foot incremental.
(e) Leak-free exhaust system includes corrosion free flexible coupling, plus improved welding of catalyst assemblies.
(f) Types of engine modifications may be less uniform throughout the industry and may include items such as an additional spark plug per cylinder, addition of a swirl control valve or other hardware needed to
achieve cold combustion stability, improved fuel economy
(g) Cost of air injection includes an electric air pump with integrated filter and relay, wiring, air-shut-off valve with integral solenoid, check valve, tubing and brackets.
V-10
-------
Chapter V: Economic Impact
Table V-4. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT1 Compared to NLEV LDT1
^mission Control Technology
Universal Exhaust Gas Oxygen Sensor (UEGO)
Air-assisted fuel injection (a)
ndividual cylinder fuel control (b)
Retarded spark timing at start-up (b)
mproved precision fuel control (c)
=aster microprocessor
=ast light-off exhaust gas oxygen sensor (planar)
Heat optimized exhaust pipe (d)
_eak-free exhaust system (e)
Engine modifications (f)
Full electronic EGR
Close-coupled catalyst
Underbody or main catalyst
Dual close-coupled catalyst
Dual underbody or main catalyst
ncreased catalyst volume
ncreased catalyst loading
mproved double layer washcoat + 600 cpsi cell density
Secondary air injection (g)
Total Incremental Cost
4-Cylinder (65.9%)
Tech.
cost est.
(in dollars)
10.00
8.00
0.00
0.00
0.00
3.00
4.00
1.00
10.00
0.00
10.00
55.00
80.00
0.00
0.00
0.00
1.84
2.81
50.00
% of NLEV
vehs. that
use tech.
0
50
10
100
100
0
100
0
100
0
0
60
70
0
0
100
0
0
50
% Tier 2
that will
req. tech.
15
50
10
100
100
100
100
100
100
0
50
60
70
0
0
100
100
100
50
nc. cost
over Tier 1
(in dollars)
1.50
0.00
0.00
0.00
0.00
3.00
0.00
1.00
0.00
0.00
5.00
0.00
0.00
0.00
0.00
0.00
1.84
2.81
0.00
15.15
6-Cylinder(34.1%)
Tech.
cost est.
(in dollars)
20.00
12.00
0.00
0.00
0.00
3.00
8.00
4.00
20.00
10.00
10.00
55.00
80.00
90.00
160.00
55.00
2.95
4.52
50.00
% of NLEV
vehs. that
use tech.
0
50
10
100
100
0
100
0
100
100
0
0
100
100
0
0
0
0
50
% Tier 2
that will
req. tech.
15
50
10
100
100
100
100
100
100
100
50
0
100
100
0
100
100
0
75
nc. cost
over Tier 1
(in dollars)
3.00
0.00
0.00
0.00
0.00
3.00
0.00
4.00
0.00
0.00
5.00
0.00
0.00
0.00
0.00
55.00
2.95
0.00
12.50
85.45
(a) Air assisted fuel injection requires minor redesign of the idle air control valve at no additional cost and addition of an adapter to each injector at a cost of $2 each.
(b) Improved precision fuel control envisioned here and retarded spark-timing at start-up constitute software changes only, at no additional hardware cost.
(c) Improved precision fuel control constitute software changes only, at no additional hardware cost.
(d) Length of heat optimized exhaust pipe required is estimated to be one foot for 4-cylinder engines, four feet for 6-cylinder engines, and six feet for eight-cylinder engines, at a cost of $1 per foot incremental.
(e) Leak-free exhaust system includes corrosion free flexible coupling, plus improved welding of catalyst assemblies.
(f) Types of engine modifications may be less uniform throughout the industry and may include items such as an additional spark plug per cylinder, addition of a swirl control valve or other hardware needed to
achieve cold combustion stability, improved fuel economy
(g) Cost of air injection includes an electric air pump with integrated filter and relay, wiring, air-shut-off valve with integral solenoid, check valve, tubing and brackets.
V-ll
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-5. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT2 Compared to NLEV LDT2
Emission Control Technology
Universal Exhaust Gas Oxygen Sensor (UEGO)
Air-assisted fuel injection (a)
Individual cylinder fuel control (b)
Retarded spark timing at start-up (b)
Improved precision fuel control (c)
Faster microprocessor
Fast light-off exhaust gas oxygen sensor (planar)
Heat optimized exhaust pipe (d)
Low thermal capacity manifold
Leak-free exhaust system (e)
Engine modifications (f)
Full electronic EGR
Close-coupled catalyst
Underbody or main catalyst
Dual close-coupled catalyst
Dual underbody or main catalyst
Increased catalyst volume
Increased catalyst loading (Pt)
Increased catalyst loading (Pd)
Increased catalyst loading (Rh)
Improved double layer washcoat + 600 cpsi cell density
Secondary air injection (g)
Total Incremental Cost
4-Cylinder (2.3%)
Tech.
cost est.
(in dollars)
10.00
8.00
0.00
0.00
0.00
3.00
4.00
1.00
20.00
10.00
0.00
10.00
55.00
80.00
0.00
0.00
0.00
1.84
2.81
50.00
% of NLEV
vehs. that
use tech.
0
50
10
100
100
0
100
0
25
100
0
50
60
70
0
0
0
0
0
0
0
0
% Tier 2
that will
req. tech.
15
50
10
100
100
100
100
100
50
100
0
50
60
70
0
0
0
0
0
100
100
0
Inc. cost
over Tier 1
(in dollars)
1.50
0.00
0.00
0.00
0.00
3.00
0.00
1.00
5.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1.84
2.81
0.00
15.15
6-Cylinder (73.7%)
Tech.
cost est.
(in dollars)
20.00
12.00
0.00
0.00
0.00
3.00
8.00
4.00
40.00
20.00
10.00
10.00
55.00
80.00
90.00
160.00
55.00
4.32
51.67
2.95
4.52
50.00
% of NLEV
vehs. that
use tech.
0
50
10
100
100
0
100
0
25
100
100
50
0
100
100
0
0
0
0
0
0
50
% Tier 2
that will
req. tech.
15
50
10
100
100
100
100
100
50
100
100
50
0
100
100
0
100
0
0
100
100
75
Inc. cost
over Tier 1
(in dollars)
3.00
0.00
0.00
0.00
0.00
3.00
0.00
4.00
10.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
55.00
0.00
0.00
2.95
4.52
12.50
94.97
8-Cylinder (24%)
Tech.
cost est.
(in dollars)
20.00
16.00
0.00
0.00
0.00
3.00
8.00
6.00
40.00
20.00
15.00
10.00
55.00
80.00
110.00
160.00
35.00
10.13
52.83
4.14
6.59
65.00
% of NLEV
vehs. that
use tech.
0
50
10
100
100
100
100
0
25
100
100
50
0
60
80
40
0
0
0
0
0
50
% Tier 2
that will
req. tech.
15
50
10
100
100
100
100
100
50
100
100
50
0
60
80
40
100
0
0
100
100
75
Inc. cost
over Tier 1
(in dollars)
3.00
0.00
0.00
0.00
0.00
0.00
0.00
6.00
10.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
35.00
0.00
0.00
4.14
6.59
16.25
80.98
(a) Air assisted fuel injection requires minor redesign of the idle air control valve at no additional cost and addition of an adapter to each injector at a cost of $2 each.
(b) Improved precision fuel control envisioned here and retarded spark-timing at start-up constitute software changes only, at no additional hardware cost.
(c) Improved precision fuel control constitute software changes only, at no additional hardware cost.
(d) Length of heat optimized exhaust pipe required is estimated to be one foot for 4-cylinder engines, four feet for 6-cylinder engines, and six feet for eight-cylinder engines, at a cost of $1 per foot incremental.
(e) Leak-free exhaust system includes corrosion free flexible coupling, plus improved welding of catalyst assemblies.
(f) Types of engine modifications may be less uniform throughout the industry and may include items such as an additional spark plug per cylinder, addition of a swirl control valve or other hardware needed to
achieve cold combustion stability, improve fuel economy
(g) Cost of air injection includes an electric air pump with integrated filter and relay, wiring, air-shut-off valve with integral solenoid, check valve, tubing and brackets.
V-12
-------
Chapter V: Economic Impact
Table V-6. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT3 Compared to Tier 1 LDT3
Emission Control Technology
Universal Exhaust Gas Oxygen Sensor (UEGO)
Air-assisted fuel injection (a)
Individual cylinder fuel control (b)
Retarded spark timing at start-up (b)
Improved precision fuel control (c)
Faster microprocessor
Fast light-off exhaust gas oxygen sensor (planar)
Heat optimized exhaust pipe (d)
Leak-free exhaust system (e)
Low thermal capacity manifold
Engine modifications (f)
Full electronic EGR
Close-coupled catalyst
Underbody or main catalyst
Dual close-coupled catalyst
Dual underbody or main catalyst
Increased catalyst volume
Increased catalyst loading (Pt)
Increased catalyst loading (Pd)
Increased catalyst loading (Rh)
Improved double layer washcoat + 600 cpsi cell density
Secondary air injection (g)
Total Incremental Cost
6-Cylinder(10.1%)
Tech.
cost est.
(in dollars)
20.00
12.00
0.00
0.00
0.00
3.00
8.00
4.00
20.00
40.00
10.00
10.00
55.00
80.00
90.00
160.00
55.00
0.43
5.17
3.40
4.52
50.00
% of Tier 1
vehs. that
use tech.
0
0
0
25
50
0
80
0
50
25
0
0
0
100
12
0
0
0
0
0
0
0
% Tier 2
that will
req. tech.
15
50
10
100
100
100
100
100
100
75
100
50
0
100
100
0
100
100
100
100
100
50
Inc. cost
over Tier 1
(in dollars)
3.00
6.00
0.00
0.00
0.00
3.00
1.60
4.00
10.00
20.00
10.00
5.00
0.00
0.00
79.20
0.00
55.00
0.43
5.17
3.40
4.52
25.00
235.32
8-Cylinder (89.9%)
Tech.
cost est.
(in dollars)
20.00
16.00
0.00
0.00
0.00
3.00
8.00
6.00
20.00
40.00
15.00
10.00
55.00
80.00
110.00
160.00
35.00
1.01
5.28
4.97
6.59
65.00
% of Tier 1
vehs. that
use tech.
0
0
0
25
50
0
80
0
50
25
0
0
0
60
55
40
0
0
0
0
0
0
% Tier 2
that will
req. tech.
15
50
10
100
100
100
100
100
100
100
100
50
0
60
80
40
100
100
100
100
100
50
Inc. cost
over Tier 1
(in dollars)
3.00
8.00
0.00
0.00
0.00
3.00
1.60
6.00
10.00
30.00
15.00
5.00
0.00
0.00
27.50
0.00
35.00
1.01
5.28
4.97
6.59
32.50
194.45
(b) Improved precision fuel control envisioned here and retarded spark-timing at start-up constitute software changes only, at no additional hardware cost.
(c) Improved precision fuel control constitute software changes only, at no additional hardware cost.
(d) Length of heat optimized exhaust pipe required is estimated to be one foot for 4-cylinder engines, four feet for 6-cylinder engines, and six feet for eight-cylinder engines, at a cost of $1 per foot incremental.
(e) Leak-free exhaust system includes corrosion free flexible coupling, plus improved welding of catalyst assemblies.
(f) Types of engine modifications may be less uniform throughout the industry and may include items such as an additional spark plug per cylinder, addition of a swirl control valve or other hardware needed to
achieve cold combustion stability, improved fuel economy
V-13
(g) Cost of air injection includes an electric air pump with integrated filter and relay, wiring, air-shut-off valve with integral solenoid, check valve, tubing and brackets.
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-7. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT4 Compared to Tier 1 LDT4
Emission Control Technology
Universal Exhaust Gas Oxygen Sensor (UEGO)
Air-assisted fuel injection (a)
Individual cylinder fuel control (b)
Retarded spark timing at start-up (b)
Improved precision fuel control (c)
Faster microprocessor
Fast light-off exhaust gas oxygen sensor (planar)
Heat optimized exhaust pipe (d)
Leak-free exhaust system (e)
Low thermal capacity manifold
Engine modifications (f)
Full electronic EGR
Close-coupled catalyst
Underbody or main catalyst
Dual close-coupled catalyst
Dual underbody or main catalyst
Increased catalyst volume
Increased catalyst loading (Pt)
Increased catalyst loading (Pd)
Increased catalyst loading (Rh)
Improved double layer washcoat + 600 cpsi cell density
Secondary air injection (g)
Total Incremental Cost
8-Cylinder(100%)
Tech.
cost est.
(in dollars)
20.00
16.00
0.00
0.00
0.00
3.00
8.00
6.00
20.00
40.00
15.00
10.00
55.00
80.00
110.00
160.00
35.00
1.01
5.28
4.97
6.59
65.00
% of Tier 1
vehs. that
use tech.
0
0
0
25
50
0
80
0
50
25
0
0
0
60
55
40
0
0
0
0
0
0
% Tier 2
that will
req. tech.
15
50
10
100
100
100
100
100
100
100
100
50
0
60
80
40
100
100
100
100
100
50
Inc. cost
over Tier 1
(in dollars)
3.00
8.00
0.00
0.00
0.00
3.00
1.60
6.00
10.00
30.00
15.00
5.00
0.00
0.00
27.50
0.00
35.00
1.01
5.28
4.97
6.59
32.50
194.45
(a) Air assisted fuel inje
(b) Improved precision fuel control envisioned here and retarded spark-timing at start-up constitute software changes only, at no additional hardware cost.
(c) Improved precision fuel control constitute software changes only, at no additional hardware cost.
(d) Length of heat optimized exhaust pipe required is estimated to be one foot for 4-cylinder engines, four feet for 6-cylinder engines, and six feet for eight-cylinder engines, at a cost of $1 per foot incremental.
(e) Leak-free exhaust system includes corrosion free flexible coupling, plus improved welding of catalyst assemblies.
(f) Types of engine modifications may be less uniform throughout the industry and may include items such as an additional spark plug per cylinder, addition of a swirl control valve or other hardware needed to
achieve cold combustion stability, improved fuel economy
(g) Cost of air injection includes an electric air pump with integrated filter and relay, wiring, air-shut-off valve with integral solenoid, check valve, tubing and brackets.
V-14
-------
Chapter V: Economic Impact
c. Hardware Costs for Evaporative Emissions Control
The standards proposed for evaporative emissions are technologically feasible now.
Many designs have been certified by a wide variety of manufacturers that already meet these
standards. A review of the 1999 model year certification results indicates that the average
family is certified at slightly less than 1.0 grams per test (gpt) on the three day diurnal plus hot
soak test, i.e. at less than half the current 2.0 gpt standard. Many families are certified at levels
considerably below 1.0 gpt, including a few families that are certified below 0.5 gpt.
The proposed standards will not require the development of new materials or even the
new application of existing materials. Low permeability materials and low loss connections and
seals are already used to varying degrees on current vehicles. The standards will likely ensure
their consistent use and discourage switching to cheaper materials or designs to take advantage
of the large safety margins manufacturers have under current standards ("backsliding").
Complex (and perhaps somewhat more expensive) approaches have been proposed which
involve pressurized fuel systems or fuel bladders. Such systems have not been implemented in
production, nor do we believe they are necessary for the standards we are proposing. We believe
manufacturers will follow more traditional paths in reducing their evaporative emissions.
There are two traditional approaches to reducing evaporative emissions. The first is to
minimize the potential for permeation and leakage by reducing the number of hoses, fittings and
connections. However, some joints and connections are necessary for vehicle assembly and
service and no known joint has zero emissions.
The second traditional approach is to use less permeable hoses and lower loss fittings
and connections. Low permeability hoses and seals as well as low loss fittings are currently
available. Fluoropolymer materials can be added as liners to hose and component materials to
yield large reductions in permeability over such conventional materials as monowall nylon. In
addition, fluoropolymer materials can greatly reduce the impact of alcohols on hydrocarbon
permeability of evaporative components, hoses and seals. Alcohols, present in about 10% of
gasoline sold in the U.S., cause swelling of conventional materials which leads to increases in
permeability and can also lead to tearing and leakage in situations where the materials are
constrained in place, such as with gaskets and O-rings. Due to the common presence of alcohols
such as ethanol in the gasoline pool and its adverse affect on materials and emissions durability,
we believe material upgrades such as those discussed above are necessary to ensure that the
benefits are captured in-use.
Steel fuel tanks and steel fuel lines have essentially zero losses due to permeation, but are
vulnerable to leakage at joints and interfaces. Manufacturers are moving toward plastic fuel
tanks for their lighter weight and greater ability to be molded to odd shapes. However, plastic
tanks are permeable and are also susceptible to seepage and higher permeability at areas where
connections and welds are made. Materials and manufacturing techniques exist to reduce these
losses.
V-15
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
To estimate the per vehicle cost of an improved evaporative system, we looked at the
incremental cost for an average current model year vehicle with a steel fuel tank (certified at ~
1.0 g) to go from a certification level of 1.0 grams per test to a level of about 0.5 grams per test
on the three day test cycle. The emission levels of 1.0 and 0.5 gpt were chosen because 1.0
represents the current average certification level and 0.5 gpt represents a certification target that
leaves a compliance margin of about 100 percent between the certification level and the
applicable standard (0.95 gpt for our proposed LDV/LLDT standard). The reductions and costs
of the individual items are shown in Table V-8 below, and reflect the incremental cost of moving
to low permeability materials, improved designs or low loss connectors. The items in the chart
are ranked in order of decreasing cost effectiveness. Since the evaporative test procedure
measures evaporative emissions each day over a three day period and then uses the highest day,
gram per day numbers in the table are a reasonable proxy for grams per test data.
Table V-8. Potential Evaporative Improvements and Their Costs to Manufacturers 3
(grams per day)
Emission Source
Fuel cap seal
Fuel pump assembly seal
Fuel and vapor line
Fuel rail/manifold connectors
Canister improvements
Fill tube clamps
Fuel and vapor line connectors
Fill tube/fill neck connector
emissions
Baseline
Vehicle
(a)
0.10
0.10
0.23
0.06
0.12
0.06
0.18
0.20
0 70
Improved
Vehicle
(b)
0.01
0.01
0.01
0.02
0.04
0.02
0.06
0.10
0 70
Change
(a-b)
0.09
0.09
0.22
0.04
0.08
0.04
0.12
0.10
n
Cost
($)
(d)
0.20
0.40
1.25
0.40
1.00
0.60
2.20
5.00
Cost
Effectiveness
Ranking
(d)/(a-b)
1
2
3
4
5
6
7
8
Table V-8 shows that a manufacturer can choose from a range of improvements, and
attain significant reductions in evaporative emissions. By selecting the first five items from the
table, the manufacturer can achieve a reduction in evaporative emissions of about 0.5 g/day for a
total cost of about three dollars per vehicle. The cost-effectiveness of these five items taken
together is approximately $2400 per ton of VOCs removed. While these figures were based on a
passenger car, we believe it is reasonable to assume the same costs here for light duty trucks
since the same basic components are used on trucks and cars. Non fuel emissions may be higher
V-16
-------
Chapter V: Economic Impact
for larger vehicles, but our proposed evaporative standard for HLDTs (1.2 gpt) is higher to
include a larger allowance for non-fuel losses.
Lastly, we note that most manufacturers are moving to "returnless" injection systems,
and at least one major manufacturer's current products are 100 percent returnless. Through more
precise fuel pumping and metering, these systems eliminate the return line in the fuel injection
system which carries unneeded fuel from the fuel injectors back to the fuel tank. Returned fuel
is a significant source of fuel tank heat and vapor generation, and therefore of evaporative
emissions. The elimination of return lines reduces the total length of hose on the vehicle and
also reduces the number of fittings and connections which can leak. We believe that most
vehicles will move to returnless injection systems either before or in conjunction with the phase-
in of the Tier 2 standards.
Our analysis is conservative in that it did not include the impact of these returnless
systems. We believe that changing to a returnless injection system may provide a 0.15 g/day
evaporative emissions benefit. If the example vehicle described above were equipped with a
returnless injection system, then, we would expect evaporative emissions of about 0.85 gpt.
Such a vehicle would require a smaller emission reduction (0.35 gpt) to hit the certification
target of 0.5 gpt.
Returnless vehicles have about one third less vapor and fuel line footage and
proportionately fewer connections and joints, accounting for most of the reduction attributable to
returnless systems. We would expect an emission improvement and cost about one third less
than those shown in the table above for fuel and vapor lines and fuel and vapor line connectors.
Because the emission improvement and cost change by the same fraction, we would not expect a
change in the cost effectiveness or ranking of these items. While the 0.15 gpt is also due to small
reductions in losses from all but the last item in the table above, we believe that, in the end, the
cost effectiveness of the proposed standards will not be significantly different for vehicles with
return or returnless systems.
d. Assembly Costs
Another variable cost manufacturers may incur are increases in vehicle assembly costs.
EPA has not estimated increased assembly costs for Tier 2 vehicles because the vast majority of
changes to the vehicles are likely to be improvements to existing emissions control systems.
Therefore, we believe that assembly cost increases are likely to be negligible. Assembly costs
for components would be incurred by the component supplier and included in the component
price estimates shown above.
e. Development and Capital Costs
In addition to the hardware costs described in the previous section, vehicle manufacturers
would also incur developmental and capital costs due to the Tier 2 standards. These fixed costs
V-17
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
include costs for reasearch and development (R&D), tooling, and certification, which
manufacturers incur prior to the production of the vehicles.
The Tier 2 standards would be phased-in over four model years beginning in 2004 for
LDVs, LDTls, and LDT2s and a two year period beginning in 2008 for LDT3s and LDT4s.
This approach would provide lead-time for R&D for the various vehicle lines to proceed
systematically. EPA estimates R&D costs of about $5 million per vehicle line (100,000
vehicles). R&D primarily includes engineering staff time and development vehicles. A large
part of the research effort will be evaluating and selecting the appropriate mix of emission
control components and optimizing those components into a system capable of meeting the Tier
2 standards. It also includes engine modifications where necessary and air/fuel ratio calibration.
Manufacturers will take differing approaches in their research programs. We estimate that $5
million would cover about 25 engineering staff person years and about 20 development
vehicles." We have estimated this large R&D effort because calibration and system optimization
is likely to be a critical part of the effort to meet Tier 2 standards. However, we believe that the
R&D costs are likely overstated because the projection ignores the carryover of knowledge from
the first vehicle lines designed to meet the standard to others phased-in later.
Tooling costs include facilities modifications necessary to produce and assemble
components and vehicles meeting the new standards. EPA has included tooling costs due to the
Tier 2 standards of approximately $2 million per vehicle line (100,000 vehicles). We believe
that this is a reasonable estimate based on engineering judgement, after reviewing previous
estimates of tooling costs for emissions control components.4
EPA recently conducted a detailed cost analysis of its vehicle certification program as
part of the CAP 2000 rulemaking, which revised the certification program and is expected to
significantly reduced manufacturer certification costs.5 For CAP 2000, EPA estimated a total
annual certification cost to the industry of between $40 and $65 million. Manufacturers incur a
large portion of these costs annually as part of certification and compliance and would incur
those costs without any change to the standards. However, EPA does allow manufacturers to
carry-over some data generated for certification when vehicles are not significantly changed
from one model year to the next. This test data is generated to demonstrate vehicle emissions
levels and emissions durability. Due to the new standards, such data would have to be generated
for the new Tier 2 vehicles, rather than carried over from previous model years. Therefore, we
believe it is appropriate to include the cost of generating new emissions test and durability data
as part of the cost analysis for Tier 2. Based on the CAP 2000 rule, EPA estimates the cost of
this testing to be about $15 million industry-wide.
EPA estimated that the R&D costs would be incurred on average three years prior to
production and the tooling and certification costs would be incurred one year prior to production.
These fixed costs were then increased by seven percent for each year prior to the start of
u This estimate is based on staff cost of $60 per hour and development vehicle cost of $100,000 per
vehicle.
V-18
-------
Chapter V: Economic Impact
production to reflect the time value of money. We estimated total R&D and tooling costs per
vehicle class by multiplying the costs per vehicle line (100,000 vehicles) by sales estimates for
each vehicle class divided by 100,000 vehicles. Finally, for the cost analysis, the fixed costs
were recovered over the first five years of production at a rate of seven percent.
EPA estimates the average per vehicle fixed costs to be between $19 and $22, as shown
in Table V-9. We derived the per vehicle fixed cost by dividing the total fixed cost per vehicle
class over the five year recovery period by the estimated total sales per vehicle class over the
same period. Differences in fixed costs among vehicle classes occur because we have projected
a phase-in of Tier 2 LDVs and LDTs and changes in sales volumes over time for the vehicle
classes. The aggregate fixed costs, vehicle phase-ins, and sales projections are described in
sections., below.
Table V-9. Per Vehicle Fixed Costs
R&D
Tooling
Certification
Total
LDV
($)
16.10
5.63
0.30
22.03
LDT1
($)
14.23
4.97
0.27
19.47
LDT2
($)
14.08
4.92
0.26
19.26
LDTS
($)
14.34
5.01
0.26
19.61
LDT4
($)
15.48
5.41
0.29
21.18
f. Total Near-term and Long-term Manufacturer Costs
The previous section presented estimates of per vehicle variable and fixed costs to the
manufacturer for the first few model years of production. These near-term per vehicle costs are
shown in Table V-10. The costs in Table V-10 include the costs for the evaporative system.
Table V-10. Total Per Vehicle Manufacturer Costs - Near Term
Variable
Fixed
Total
LDV
($)
46.10
22.03
68.13
LDT1
($)
42.38
19.47
61.85
LDT2
($)
93.03
19.26
112.29
LDTS
($)
201.83
19.61
221.44
LDT4
($)
197.70
21.18
218.88
For the long-term, there are factors that EPA believes are likely to reduce the costs to
manufacturers. As noted above, we project fixed costs to be recovered by manufacturers during
V-19
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
the first five years of production, after which they would expire. For variable costs, research in
the costs of manufacturing has consistently shown that as manufacturers gain experience in
production, they are able to apply innovations to simplify machining and assembly operations,
use lower cost materials, and reduce the number or complexity of component parts. These
effects are often described as the manufacturing learning curve.6
The learning curve is a well documented and accepted phenomenon dating back to the
1930s. The general concept is that unit costs decrease as cumulative production increases.
Learning curves are often characterized in terms of a progress ratio, where each doubling in
cumulative production leads to a reduction in unit cost to a percentage "p" of its former value
(referred to as a "p cycle"). The organizational learning which brings about a reduction in total
cost is caused by improvements in several areas. Areas involving direct labor and material are
usually the source of the greatest savings. These include, but are not limited to, a reduction in
the number or complexity of component parts, improved component production, improved
assembly speed and processes, reduced error rates, and improved manufacturing process. These
all result in higher overall production, less scrappage of materials and products, and better
overall quality.
Companies and industry sectors learn differently. In a 1984 publication, Button and
Thomas reviewed the progress ratios for 108 manufactured items from 22 separate field studies
representing a variety of products and services.7 As shown in Figure V-l, of the 108 progress
ratios observed, eight were less than 70 percent, 39 were in the range of 71 to 80 percent, 54
were in the range of 81 to 90 percent, and seven were above 90 percent. The average progress
ratio for the whole data set falls between 81 and 82 percent. The lowest progress ratio of 55
percent shows the biggest improvement, representing a remarkable 45 percent reduction in costs
with every doubling of production volume. At the other extreme, except for one company that
saw increasing costs as production continued, every study showed cost savings of at least five
percent for every doubling of production volume. This data supports the commonly used p value
of 80 percent, i.e., each doubling of cumulative production reduces the former cost level by 20
percent. As each successive p cycle takes longer to complete, production proficiency generally
reaches a relatively stable plateau, beyond which increased production does not necessarily lead
to markedly decreased costs.
V-20
-------
Chapter V: Economic Impact
15
10
CD
3
o-
CD
0
Distribution of Progress Ratios
55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 101 103 105 107
Progress Ratio
From 22 field studies (n = 108).
Figure V-l. Distribution of Progress Ratios
V-21
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
EPA applied a p value of 20 percent beginning in the third year of production in this
analysis. That is, the variable costs were reduced by 20 percent for each doubling of cumulative
production. To avoid overly optimistic projections, we included several additional constraints.
Using one year as the base unit of production, the first doubling would occur at the start of the
third model year of production. To be conservative, we did not incorporate further cost
reductions due to the learning curve. We applied the learning curve reduction only once because
we anticipate that for the most part the Tier 2 standards would be met through improvements to
existing technologies rather than through the use of new technologies. With existing
technologies, there would be less opportunity for lowering production costs.
In addition, we did not apply the learning curve to the catalyst precious metal costs due to
the uncertainty of future precious metal prices. Although manufacturers may be able to reduce
the use of precious metals due to the learning curve, the future price of precious metals is highly
uncertain. Any savings due to a reduction in the amount of precious metals used for a catalyst
system could be overcome by increased precious metal unit costs. Finally, we did not apply the
learning curve to the evaporative system costs. Evaporative systems have been well developed
and the anticipated system improvements are available today and are likely to be employed by
manufacturers prior to 2004 on a large number of vehicles.
Table V-l 1 presents EPA's estimates of long-term per vehicle manufacturer costs. As
noted above, we have projected cost reductions due to the learning curve to occur in the third
year of production and the fixed costs to expire for the sixth year of production. Due to the
phase-in of standards, these cost reductions are not tied to particular model years. As shown in
Table V-l 1, we project manufacturer costs to decrease by 21 to 40 percent for the long-term.
The percentage decrease in costs varies largely due to the variation in projected costs for
precious metals, which are not subject to the learning curve cost reduction factor. We have
projected a larger increase in the use of precious metals for LDT3s and LDT4s than for LDVs.
Table V-ll. Long-term Total Incremental Per Vehicle Manufacturer Costs
Production Year
1st and 2nd year
3rd year: learning curve applied
6th year: fixed costs expire
LDV
($)
68.13
61.95
39.92
LDT1
($)
61.85
56.82
37.36
LDT2
($)
112.29
101.17
81.91
LDT3
($)
221.44
192.97
173.36
LDT4
($)
218.88
190.36
169.18
2.
Tier 2 Vehicle Consumer Costs
Costs to consumers consists of increases in vehicle purchase price and increases in
vehicle operating costs. EPA has not estimated an increase in vehicle operating costs due to the
Tier 2 standards. Manufacturers will most likely meet the standards through improvements to
V-22
-------
Chapter V: Economic Impact
existing technologies. We do not anticipate that the improvements to technologies will affect
fuel economy or in-use maintenance. The costs of fuel quality improvements are provided in
section B, below.
For the up-front cost or purchase price increase, EPA anticipates that manufacturers
would pass along their incremental costs for Tier 2 vehicles, including a markup for overhead
and profit, to vehicle purchasers. Thus, we expect consumers would experience purchase price
increases based on the manufacturer costs discussed in section A. 1. To account for manufacturer
overhead and profit, manufacturer incremental variable costs are multiplied be a Retail Price
Equivalent (RPE) factor. The RPE factor we used in this analysis, 1.26, is the same one EPA has
used in previous analyses for LDVs and LDTs. This methodology and the RPE mark-up factor
are based on contractor studies regarding hardware costs and RPEs.8'9 Table V-12 presents the
increases in vehicle costs to consumers EPA has estimated for Tier 2 vehicles. The costs shown
in Table V-12 include the costs of the evaporative system improvements, as well as the improved
exhaust emissions control system/ We expect decreases in manufacturing costs over time,
described in section l.f, above, to be passed along to consumers in the form of purchase price
decreases.
Table V-12. Incremental Per Vehicle Costs to Consumers for Tier 2 Vehicles
Production Year
1st and 2nd year
3rd year: learning curve applied
6th year: fixed costs expired
LDV
($)
80.12
72.34
50.31
LDT1
($)
72.88
66.55
47.08
LDT2
($)
136.48
122.47
103.21
LDTS
($)
273.92
238.05
218.44
LDT4
($)
270.28
234.35
213.17
3.
Annual Total Nationwide Costs for Tier 2 Vehicles
a.
Overview of Nationwide Vehicle Costs
The above analyses developed incremental per vehicle manufacturer and consumer cost
estimates for each class of Tier 2 LDVs and LDTs. With data for the current size and
characteristics of the vehicle fleet and projections for the future, we have translated these per
vehicle costs into estimated total annual costs to the nation for the Tier 2 standards. Table V-13
presents the results of this analysis. As shown in Table V-13, EPA projected total cost starting at
$256 million in 2004 and peaking at $1,587 million in 2009 when the phase-in of the standards is
complete. Per-vehicle costs savings over time reduce projected costs to a value of $1,346
million in 2014, after which the growth in vehicle population leads to increasing costs that reach
v EPA estimated costs to the manufacturer for evaporative system improvements to be $3.25. The RPE for
the evaporative system would therefore be $4.10.
V-23
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
$1,386 million in 2020. The calculated total costs represent a combined estimate of fixed costs,
as they are allocated over fleet sales during the first five years of sale, and variable costs
assessed at the point of sale. The aggregate costs include exhaust and improved evaporative
control systems. These estimates do not include costs due to improved fuel quality, which are
presented in section 2., below. The remainder of this section discusses the methodology we used
to derive the total annual cost estimates and provides total annual vehicle costs for calender years
2004 through 2020.
Table V-13. Estimated Annual Nationwide Costs
(thousands of dollars)
Category
LDV
LDT1
LDT2
LDT3
LDT4
Total
2004
246,026
0
0
10,507
0
256,533
2009
342,543
96,101
592,396
373,188
182,341
1,586,569
2014
285,556
70,113
512,604
329,438
147,904
1,345,614
2020
294,231
72,243
528,175
339,445
152,397
1,386,491
b. Methodology
To prepare these estimates, we projected sales for each vehicle class, the change in sales
over time, and the phase-in of Tier 2 vehicles for each class over the phase-in schedule. We
estimated current vehicle sales based on sales data submitted by vehicle manufacturers as part of
certification. These sales estimates correlated reasonably well with other available sales
information. We reduced the national sales numbers by 10 percent for LDVs and nine percent
for LDTs to account for sales in California.10 California sales were excluded from this analysis
because California emissions standards apply to those vehicles.
To account for the current trend in sales of fewer LDVs and more LDTs, we reduced the
LDV fraction of total sales and increased the LDT fraction of total sales by 1.6 percent per year
from 1998 through 2008. After 2008, sales were stabilized at a mix of 40 percent LDVs and 60
percent LDTs. We also applied this shift in sales in its analysis of emissions reductions. These
projections are based on the current trend toward increased sales of LDTs. We are aware of an
industry study that projects the sales split leveling off much sooner at half LDVs and half
LDTs.11 Using a higher percentage of LDT sales results in higher overall cost projections
because the per vehicle costs are higher for LDTs. In this way, EPA's cost analysis is more
conservative than if we assumed sales leveled off at one-half LDVs and one-half LDTs. Finally,
we have modeled overall vehicle sales to grow at 0.5 percent per annum on average over the
V-24
-------
Chapter V: Economic Impact
period of the analysis.12 Table V-14 provides EPA's estimates for vehicle sales for 1998 and
projections for select future years.
Table V-14. Estimated Annual 49-State Vehicle Sales
(thousands of vehicles)
Category
LDV
LDT1
LDT2
LDT3
LDT4
Total
1998
7,352
1,012
3,374
1,025
471
13,234
2004
6,266
1,268
4,228
1,284
591
13,636
2008
5,502
1,447
4,824
1,465
674
13,911
2072
5,620
1,475
4,917
1,493
687
14,192
2020
5,849
1,535
5,117
1,554
715
14,769
In addition to vehicle sales, EPA also projected a phase-in of Tier 2 vehicles (including
improved evaporative controls systems) for each vehicle class. Projecting the phase-in of Tier 2
vehicles is necessary to estimate aggregate costs of the standards during the phase-in period.
Rather than assume a phase-in of 25/50/75/100 percent for each vehicle class, LDV, LDT1, and
LDT2, we projected a phase-in based on cost and difficulty considerations. We projected that
manufacturers would begin the phase-in with LDVs and end with LDT2s. We believe
manufacturers will be able to meet Tier 2 standards more easily and at a lower cost for lighter
vehicles compared to heavier vehicles.
We have projected some sales of Tier 2 LDT3s and LDT4s prior to 2008, for reasons
described in section V. A. 1 .a. above. These early sales would off-set vehicles in higher bins in
the averaging program for the interim standards. To make these projections, we assessed the
current certification levels of LDT3s and LDT4s to determine how averaging could be used by
manufacturers to avoid redesigning vehicles to meet interim standards. We found that, currently,
about 25 percent of vehicles overall would fall into the highest bin (0.60 g/mile NOx), 30 percent
in the next highest bin (0.3 g/mile NOx) and the remaining 45 percent would meet the interim
standard (0.2 g/mile NOx). We conducted this analysis for each manufacturer and determined
how many vehicles meeting the Tier 2 standards would be needed to off-set vehicles in the
higher bins. In this analysis, the vehicles in the highest bin were phased-in last. This analysis
may overestimate the number of Tier 2 vehicles necessary because it does not account for the
manufacturers' ability to make minor adjustments to vehicles close to the interim standard (i.e.,
those in the 0.3 g/mile NOx bin) which may allow those vehicles to meet the interim standard.
Essentially, these analyses have resulted in projections of Tier 2 vehicle phase-ins which
start with the lighter vehicles within each of the two categories and progress through the heavier
vehicles until all vehicles meet the Tier 2 standards in 2009. Table V-15 presents EPA's
V-25
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
projected phase-in of Tier 2 vehicles we modeled for the aggregate cost analysis over the phase-
in period of 2004 through 2008. Manufacturers would select the appropriate phase-in for their
vehicle fleets. These modeling projections simply allow EPA to perform the aggregate cost
analysis, reasonably accounting for the standards phase-in and the manufacturer's ability to
average within the various programs.
Table V-15. Projected Overall Industry Phase-in of Tier 2 Vehicles and Improved
Evaporative Emissions Controls For Purposes of the Aggregate Cost Analysis
Model Year
2004
2005
2006
2007
2008
2009
LDV
(%)
50
100
100
100
100
100
LDT1
(%)
0
0
100
100
100
100
LDT2
(%)
0
0
30
100
100
100
LDT3*
(%)
2
7
22
55
100
100
LDT4*
(%)
0
0
0
0
35
100
*Improved evaporative systems have been projected to phase-in 50 percent in 2008 and 100
percent in 2009 for LDT3s and LDT4s, starting with LDT3s in 2008.
This is the phase-in schedule for Tier 2 vehicles EPA used in this analysis based on the
assumption that manufacturers would perceive a fleet-wide integrated strategy as the most
efficient and least-cost approach. Others are possible, but overall costs during the phase-in years
would not be significantly different.
c. Estimates of Total Nationwide Vehicle Costs by Vehicle Class
EPA used the above sales and phase-in projections along with per vehicle variable and
fixed costs to estimate total annual vehicle costs by vehicle class. We have summed the fixed
costs for the vehicle categories and have amortized them over the first five years of production at
a seven percent discount rate. We multiplied sales by per vehicle variable costs (with the RPE
mark-up applied) to calculate total annual variable costs. As discussed above, variable costs are
reduced after the second year of production due to the learning curve factor. Tables V-16
through V-20 present total annual nationwide costs by vehicle class for years 2004 through 2020.
Table V-21 presents these cost figures summed for all LDVs and LDTs.
V-26
-------
Chapter V: Economic Impact
Table V-16. Annual Nationwide Costs For Tier 2 LDVs
Calendar
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Fixed Cost
($)
64,020,172
128,040,345
128,040,345
128,040,345
128,040,345
64,020,172
0
0
0
0
0
0
0
0
0
0
0
Variable Cost
($)
182,006,058
353,094,621
319,146,974
286,574,509
276,809,911
278,523,090
279,915,706
281,315,284
282,721,861
284,135,470
285,556,147
286,983,928
288,418,848
289,860,942
291,310,247
292,766,798
294,230,632
Total Cost
($)
246,026,230
481,134,966
447,187,318
414,614,854
404,850,256
342,543,263
279,915,706
281,315,284
282,721,861
284,135,470
285,556,147
286,983,928
288,418,848
289,860,942
291,310,247
292,766,798
294,230,632
Table V-17. Annual Nationwide Costs For Tier 2 LDTls
Calendar
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Fixed Cost
($)
0
0
27,715,184
27,715,184
27,715,184
27,715,184
27,715,184
0
0
0
0
0
0
0
0
0
0
Variable Cost
($)
0
0
72,431,363
74,828,038
68,098,756
68,386,267
68,728,199
69,071,840
69,417,199
69,764,285
70,113,106
70,463,672
70,815,990
71,170,070
71,525,920
71,883,550
72,242,968
Total Cost
($)
0
0
100,146,547
102,543,222
95,813,941
96,101,452
96,443,383
69,071,840
69,417,199
69,764,285
70,113,106
70,463,672
70,815,990
71,170,070
71,525,920
71,883,550
72,242,968
V-27
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-18. Annual Nationwide Costs For Tier 2 LDT2s
Calendar
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Fixed Cost
($)
0
0
27,725,154
92,417,180
92,417,180
92,417,180
92,417,180
64,692,026
0
0
0
0
0
0
0
0
0
Variable Cost
($)
0
0
159,053,529
547,721,457
545,161,597
499,978,476
502,478,368
504,990,760
507,515,714
510,053,292
512,603,559
515,166,576
517,742,409
520,331,121
522,932,777
525,547,441
528,175,178
Total Cost
($)
0
0
186,778,683
640,138,637
637,578,777
592,395,655
594,895,548
569,682,786
507,515,714
510,053,292
512,603,559
515,166,576
517,742,409
520,331,121
522,932,777
525,547,441
528,175,178
Table V-19. Annual Nationwide Costs For Tier 2 LDTSs
Calendar
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Fixed Cost
($)
869,782
2,029,491
6,378,400
15,946,000
28,992,728
28,122,946
26,963,237
22,614,328
13,046,727
0
0
0
0
0
0
0
0
Variable Cost
($)
9,636,772
23,267,227
74,126,448
191,702,483
359,149,330
345,064,896
322,930,549
324,545,202
326,167,928
327,798,767
329,437,761
331,084,950
332,740,375
334,404,076
336,076,097
337,756,477
339,445,260
Total Cost
($)
10,506,553
25,296,718
80,504,848
207,648,483
388,142,058
373,187,842
349,893,785
347,159,529
339,214,655
327,798,767
329,437,761
331,084,950
332,740,375
334,404,076
336,076,097
337,756,477
339,445,260
V-28
-------
Chapter V: Economic Impact
Table V-20. Annual Nationwide Costs For Tier 2 LDT4s
Calendar
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Fixed Cost
($)
0
0
0
0
4,819,090
13,768,828
13,768,828
13,768,828
13,768,828
8,949,738
0
0
0
0
0
0
0
Variable Cost
($)
0
0
0
0
57,785,178
168,572,016
160,863,474
145,707,212
146,435,748
147,167,927
147,903,766
148,643,285
149,386,502
150,133,434
150,884,101
151,638,522
152,396,714
Total Cost
($)
0
0
0
0
62,604,267
182,340,843
174,632,302
159,476,040
160,204,576
156,117,665
147,903,766
148,643,285
149,386,502
150,133,434
150,884,101
151,638,522
152,396,714
Table V-21. Annual Nationwide Costs For Tier 2 LDVs and LDTs
Calendar
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Fixed Cost
($)
64,889,954
130,069,836
189,859,083
264,118,709
281,984,526
226,044,310
160,864,429
101,075,181
26,815,555
8,949,738
0
0
0
0
0
0
0
Variable Cost
($)
191,642,830
376,361,848
624,758,314
1,100,826,487
1,307,004,772
1,360,524,745
1,334,916,295
1,325,630,297
1,332,258,449
1,338,919,741
1,345,614,340
1,352,342,411
1,359,104,123
1,365,899,644
1,372,729,142
1,379,592,788
1,386,490,752
Total Cost
($)
256,532,784
506,431,684
814,617,397
1,364,945,196
1,588,989,298
1,586,569,055
1,495,780,724
1,426,705,478
1,359,074,004
1,347,869,479
1,345,614,340
1,352,342,411
1,359,104,123
1,365,899,644
1,372,729,142
1,379,592,788
1,386,490,752
B.
Gasoline Desulfurization Costs
V-29
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
In this section, we will first lay out the methodology for our analysis of gasoline
desulfurization costs. Then we will present the estimated cost of desulfurizing gasoline. Finally,
we will discuss other relevant issues concerning the desulfurization of gasoline.
1. Methodology
The approach to estimating gasoline desulfurization costs is different from how we
estimated costs in the Gasoline Sulfur Staff Paper. The costs presented in that report were
developed using a refinery model run by the Oak Ridge National Laboratory (ORNL). Due to
some improvement work which was being done with that refinery model, we were not able to
develop costs for this analysis using that model. For this reason, we developed our own gasoline
desulfurization model. For the Final Rule, we expect that the ORNL refinery model will be used
to develop gasoline desulfurization costs and we will consider those costs as well as other costs
from any other refinery modeling studies which may be performed by other studies.
The analysis was performed on a regional basis. The regions used are Petroleum
Administrative Districts for Defense (PADDs). The analysis was conducted this way to take
advantage of the PADD-level refinery information which is available for each PADD. This will
help improve the understanding of how the cost for desulfurization will differ between these
regions. Figure IV-2 above depicts the various PADDs of the country. As shown in the Figure,
PADD 1 comprises the Northeast states, PADD 2 comprises the Midwest states, PADD 3
comprises the Gulf Coast states, PADD 4 comprises the Rocky Mountain states, and PADD 5
comprises the West Coast states. One issue to note is that PADD 5 normally includes California.
However, since California already requires low sulfur gasoline as part of its California Phase II
Reformulated Gasoline program, it would be inappropriate to include California in any part of
this analysis. Thus, this analysis estimates the cost of desulfurizing gasoline in PADD 5 outside
of California, which will be indicated as PADD 5OC from this point on.
The cost analysis for each PADD is performed for a single refinery which represents the
average refinery characteristics for that PADD. Each PADD-average refinery is created by
taking all the refining capacity and throughput in that PADD and averaging it over the number of
refineries in that PADD.W The costs for the entire PADD can be calculated by simply
multiplying the individual refinery cost by the number of refineries in that PADD. This analysis
presumes that each refinery must install a desulfurization unit which slightly overestimates the
capital cost, as some refineries already produce gasoline with less than or close to 30 ppm sulfur
and they would meet the proposed sulfur standard without adding any gasoline desulfurization
units.
w This methodology of modeling an average refinery, or a representation of the PADD-average, is the
only practical method (since modeling every refinery would require proprietary information and substantial
complexity) and the industry-accepted method for estimating fuel costs. This method, however, comes with the
trade-off of only approximating the cost which would be developed if we were able to average and aggregate the
results of modeling each and every refinery.
V-30
-------
Chapter V: Economic Impact
Throughout most of this analysis, we presumed that the average refinery modeled will
make their capital investments to meet a 30 ppm sulfur standard starting January 1, 2004. This
investment scenario was assumed to simplify the analysis, however, it does not capture possible
investment scenarios which could arise in response to refiners taking advantage of the proposed
Averaging, Trading and Banking (ABT) program. Participating in the ABT program, refiners
would choose a timeframe which best fits their company's financial situation, credit generation
and banking capacity, and credit purchasing options, within the constraints of the per-gallon and
corporate average sulfur standards. The net effect is that capital investments would likely be
incurred over a six year period instead of all at once in 2003, which, if considered throughout
this analysis, would have a small, decreasing effect on the costs estimated in the analysis. At the
end of the analysis, we project the distribution of the capital costs over the years of the ABT
program.
Each PADD was calibrated so that the volumes and sulfur levels for the various streams
which contributed to the sulfur in the whole pool balanced with the sulfur level in the gasoline
pool. The streams which contribute any significant amount of sulfur in the gasoline pool include
the FCC gasoline, straight run (nonrefmed crude oil in the gasoline boiling range), alkylate, and
coker gasoline, if any was blended directly into gasoline (it may have been sent to other units
such as the reformer, the alkylate plant, and the isomerate unit, which all desulfurize their feeds
prior to processing). While alkylate sulfur levels can typically be equal to or less than five ppm,
refineries which make alkylate from coker naphtha can have high levels of sulfur in alkylate
(higher than 50 ppm). When these higher sulfur alkylate streams are averaged with the lower
sulfur streams, the level of the average alkylate sulfur level will probably be high enough that we
felt that it should be accounted for. The actual sulfur balance is described in detail further
below.
This analysis does not directly estimate gasoline desulfurization costs for a portion of the
industry affected. California refiners currently produce some non-California, low sulfur non-
reformulated gasoline which is shipped outside of the state, yet this analysis did not attempt to
estimate the cost to those refiners of desulfurizing that gasoline. Similarly, nondomestic refiners
import some gasoline to the U.S., and these costs are not estimated as well. However, after
estimating the average gasoline desulfurization cost for domestic refiners outside of California,
the aggregate desulfurization cost is calculated by multiplying the domestic cost outside of
California by a factor which accounts for the total number of gallons of gasoline sold in the U.S.
outside of California. Thus, this analysis uses the estimated cost increase of gasoline produced
by non-California U.S. refiners to represent the cost desulfurizing all gallons of gasoline
consumed in the U.S.
The first step in desulfurization was presumed to be complete use of at least some of the
existing desulfurization capacity available in the refinery. There are FCC feed hydrotreaters
already present in many refineries which apparently are not being operated at capacity. The
API/NPRA survey of 1997, which summarizes the operating characteristics and gasoline
qualities of most of the U.S. refining industry, summarized the capacity and utilization of that
capacity for FCC feed hydrotreaters for each PADD. Using that data, it was presumed that these
units could be run at 100 percent of stream day capacity minus five percent which would allow
V-31
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
for increased use which is expected to occur between now and when the gasoline sulfur program
is proposed to begin. Thus, before calculating the size and costs associated with a FCC gasoline
hydrotreater, the initial sulfur level in each PADD was adjusted to reflect the use of this capacity.
The operating cost for running FCC feed hydrotreaters at capacity was based on data from
operating costs data for a FCC hydrofmer. In some PADDs there was additional hydrocracker
capacity as well. However, running these units at capacity was found to cost more per unit
sulfur removed than putting in a Mobil Oil Octgain or a CDTECH unit. Thus, these units were
presumed to not be operated at capacity. It should be noted that there are other hydrotreaters in
refineries which perhaps could provide additional hydrotreating, either before additional
hydrotreating capacity is installed, or to meet a low sulfur gasoline target. Feeds to the reformer,
isomerate and alkylate units are almost always hydrotreated, and running these units at capacity
could provide additional desulfurization with some operating cost, and perhaps some additional
capital cost for debottlenecking. However, trying to estimate the cost and incremental
desulfurization available from these units was not possible with the information we had available
to us.
In most cases, cost estimation for desulfurization down to 30 ppm is made based on
CDTECH and Mobil Oil's desulfurization technologies, which are improved FCC gasoline
desulfurization technologies. For this analysis, we presume that half of the FCC gasoline
hydrotreaters which would be installed are CDTECH units and the other half would be Mobil Oil
Octgain units. Since past gasoline sulfur cost analyses were made based on Mobil Oil's Octgain
125 process (2nd generation Octgain), the cost of desulfurization of PADD 3 gasoline was also
estimated with this process as well. This will allow us to compare the estimated desulfurization
cost of newer desulfurization technologies with the technology previously relied upon in these
other studies. Coker gasoline (that part of the coker stream in the gasoline boiling range which is
blended directly into gasoline), if there was any, is assumed to be treated along with the FCC
gasoline. Because maximum hydrotreating with the improved FCC gasoline desulfurization
technologies did not reduce the gasoline sulfur levels in PADDs 4 and 5 down to 30 ppm, some
straight run was presumed to be desulfurized as well. The process presumed to be used for that
desulfurization is Merox. We obtained generic capital and operating cost data for adding a
Merox unit.
The CDTECH costs for achieving the target sulfur level are based on the combined units
of CDHydro and CDHDS. The minimum severity of these CDTECH units that would result in
treating the entire FCC gasoline stream down to 30 ppm was presumed to be used, in lieu of
more severely hydrotreating the heaviest fraction of the FCC gasoline stream. To allow us to
estimate desulfurization at different severities, CDTECH provided us cost and unit operations
data for a range in hydrotreating severities, from 50 percent to almost 98 percent, for their
process.
Mobil Oil provided OCTGAIN 3 desulfurization cost data at one desulfurization severity.
To estimate the desulfurization cost to reach 30 ppm, the fraction of FCC and coker gasoline to
be desulfurized at that severity was determined. The Octgain unit was then sized to process that
fraction. The sizing of the Octgain unit is consistent with how refiners are expected to use this
technology, which is treat only the heavy FCC gasoline if this achieves the sulfur target. If not,
V-32
-------
Chapter V: Economic Impact
then the heavy and medium FCC gasoline would be hydrotreated. Finally, only if this did not
achieve the desired sulfur target, the entire FCC gasoline stream would be treated. In all cases,
only one unit is used, but the unit is sized to treat the appropriate portion of the FCC gasoline
pool until the target sulfur level is met. To facilitate this selective desulfurization strategy, a
splitter (distillation column) may be needed between the FCC main fractionation tower and the
OCTGAIN unit (however, if the entire FCC pool is being treated then a splitter may not be
needed). We presume that half the refineries already have a splitter and are using it while the
other half will have to install this splitter. Thus each PADD-average refinery has half the capital
and operating cost of a full sized splitter. In cases of meeting a less stringent gasoline sulfur
target where less than all the heavy FCC naphtha is being treated, the splitter is sized according
to the volume of the FCC heavy naphtha being treated (if 60 percent of the FCC heavy naphtha
must be treated to reach a particular sulfur reduction target, then only 60 percent of the FCC
naphtha is routed to the splitter).
As stated above, cost estimation is based on vendor supplied operating and capital costs
for CDTECH, OCTGAIN, FCC feed hydrotreating, and Merox units; while the ORNL refinery
model is referenced for splitter operating and capital costs. Shell Oil (now Equilon) engineers
analyzed the cost of installing a CDTECH unit into a generic refinery and compared their costs
to those of CDTECH. This comparison is summarized below.
The cost of sulfur reduction was estimated for sulfur reductions down to PADD-average
levels of 150, 100, 80, 40 and 30 ppm. The national cost of desulfurization is calculated by
volume weighting the individual PADD costs. The costs are estimated for meeting the averaging
standard, which is the cost estimation methodology recommended to us by the oil industry.
Some additional cost may be incurred for meeting the cap standard, however, estimating these
costs is more uncertain, so only the issues associated with meeting the cap standard are
discussed. Therefore, no explicit costs of a cost over and above an averaging standard are
developed.
a. Cost Inputs
Vendors for various desulfurization technologies were contacted to obtain detailed
information on the raw material and utility needs and desulfurization capabilities for their
technologies. This information is summarized below in Table V-22:
V-33
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-22. Raw Material and Utility Needs and Desulfurization Capabilities for Several
Desulfurization Technologies
Feed (Bbls/Day)
Capital Cost
($MM) ISBL
Six Tenths Rule
Exponent
H2 Consumption
(SCF/Bbl)
Electricity
(KwH/Bbl)
HP Steam
(Lbs/Bbl)
Fuel Gas
(MBtu/Bbl)
Catalyst Cost
($/Bbl)
Cooling Water
(Gals/Bbl)
Octane Loss
(R+M)/2
Yield Loss
(vol.% gasoline)
(vol.% LPG)
(vol.% diesel)
(vol.% resid)
Operating Cost
($/Bbl)
FCC Feed
Hydrotreating
-
-
-
290
1.5
14
56
0.04
-
(6.5)
(3.4)
2.2
3.1
CDTECH*
(96%
desul)
40,000
22.5
0.65
82
0.5
-
55
0.19
1.2
-
Mobil
Octgain 220f
(95% desul)
25,000
25
0.65
125
1.1
40
12
0.25
220
0.8
0.7
Mobil
Octgain 125
(98% desul)
8000
14.5
0.65
420
2.3
-
51
0.43
45
1.6
14
Merox
(60%
desul)
10,000
3.5
0.6
-
-
-
-
-
-
0.06
CDTECH provided data for desulfurization from 50 percent to 98 percent; only the data for 96 percent
gasoline desulfurization is summarized here.
Data was presented separately for light cat naphtha, medium cat naphtha, and heavy cat naphtha; however,
in this table the data was volume-averaged together into one column for treating all three together.
V-34
-------
Chapter V: Economic Impact
/. Capital Cost
Capital costs are the one-time costs incurred by purchasing and installing new hardware
in refineries. A number of factors are accounted for when estimating capital costs. The cost is
calculated by first starting with the capital costs summarized in Table V-22 above. However,
those capital costs are for the throughputs indicated, and must be adjusted to reflect throughputs
different from that listed. The throughput normally listed for a new facility is the day to day
throughput used to calculate operating cost, which is normally expressed as barrels of feed per
calendar day. This throughput must be adjusted to estimate the capital cost based on stream
days.x A calendar day to stream day factor is used to account for the difference between the
throughput in calendar days versus the throughput in stream days. The ORNL refinery model
provides calendar day to stream day inflation factors and the factor for the FCC unit, which is
seven percent, is used here (the calendar day throughput is multiplied by 1.07 to estimate a
stream day throughput). The factor for the FCC unit is used because these improved
technologies are designed to operate with and be shutdown for maintenance on the same
schedule as the FCC unit.
Also, a 15 percent design safety factor was applied to the capital cost. This means that
facilities are sized 15 percent larger than what planned throughput would otherwise require.13
This design factor, also sometimes called a contingency factor, is normally applied to cost
estimates to account for uncertainties in the design. An additional five percent is added to the
safety design factor, for a total of 20 percent, to account for the newness of these technologies.14
Once the stream day throughput is estimated, and the design factor is applied, if the
recalculated throughput is different from the throughputs listed in the above table, the capital
cost is estimated at this other throughput using an exponential equation termed the "six-tenths
rule."15 The equation is as follows: (Sb/Sa)exCa=Cb, where Sa is the size of unit quoted by the
vendor, Sb is the size of the unit for which the cost is desired, e is the exponent, Ca is the cost of
the unit quoted by the vendor, and Cb is the desired cost for the different sized unit. The
exponential value "e" used in this equation is 0.9 for a splitter, 0.65 for OCTGAIN and
CD TECH units, and 0.6 for a merox unit.
The capital costs are adjusted further to capture other cost factors which affect the
ultimate cost of installing capital, and these factors vary by PADD. One of these factors adjusts
x The throughput in calendar days is simply the total throughput of a unit in a year divided by the
number days in a year. The throughput in stream days is the total throughput of a unit in a year divided by the
number of days which the unit is operating. The stream day daily throughput determines the necessary capacity of
the unit since a unit must be able to handle that throughput on the days which the unit is operating.
V-35
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
the onsite costs upward to account for offsite costs.y According to Gary and Handwerk, the
proportion of offsite cost to onsite cost varies depending on the crude oil throughput of the
refinery. These varying factors are summarized in Table V-23.16
Table V-23. Offsite Factors for Different Sized Refineries
Refinery Size in Crude Oil Feed (BPSD)
Less than 30,000
30,000 - 100,000
More than 100,000
Offsite Costs, Percent of Inside Battery
Limit Costs
50
30
20
Based on these offsite factors, because PADD 3 refineries average about 150,000 barrels per day,
the representative refinery used in this analysis for that PADD was assigned an offsite factor of
1.2 (or a 20 percent increase). PADD 1, PADD 2, and PADD 5OC, refineries all average about
100,000 barrels of crude oil per day, so the representative refinery used in this analysis was
assigned an offsite factor of 1.25. Finally, PADD 4 refineries average less than 30,000 barrels of
crude oil per day, so the representative refinery for that PADD was assigned an offsite cost
factor of 1.5. These factors are summarized in Table V-25 below.
Another factor which varies from PADD-to-PADD is the labor cost for installing the
capital. Gary and Handwerk provide estimates for labor costs for a number of different cities,
and these estimates are summarized in Table V-24, below.17
y When vendors normally report a capital cost, that cost includes what are called onsite costs for
complete installation of the hardware, but excludes other costs integral for the functioning of the unit, and these
other costs are called offsite costs. Onsite costs normally include the capital cost for the process unit, storage
facilities, cooling water facilities, and steam facilities. Offsite costs normally include electric power distribution,
any fuel gas facilities, water supply and treatment, plant air, fire protection, flare hookup, drain system, waste
containment, plant communication, roads and walks, railroads, roads and walkways, fences and buildings. Thus
offsite costs are other capital costs which will allow the facility to run as an integral unit within the refinery.
V-36
-------
Chapter V: Economic Impact
Table V-24. Labor Costs in Selected Cities
Location
U.S. Gulf Coast
Los Angeles
Portland, Seattle
Chicago
St. Louis
Detroit
New York
Philadelphia
Alaska, North Slope
Alaska, Anchorage
Relative Cost
1.0
1.4
1.2
1.3
1.4
1.3
1.7
1.5
3.0
2.0
Based on this information, each PADD was assigned a cost factor to account for the labor cost
for installing capital. PADD 1 was assigned a value of 1.5 which corresponds with the factor for
Philadelphia, where a number of PADD 1 refineries are located. PADD 2 was assigned a value
of 1.3 which corresponded with Chicago and Detroit. PADD 3 was assigned a value of 1.0,
which is accepted as the reference refinery-related labor cost for the country. PADD 4 was
assigned a value of 1.4, which corresponds to the value of St. Louis, the closest city to PADD 4,
and PADD 5 outside of California was assigned a value of 1.2, which corresponds with Portland
and Seattle. These location factors are summarized below in Table V-25.
Table V-25. Capital Cost Factors Which Vary by PADD
Offsite
Factor
Location
Factor
PADD 1
1.25
1.5
PADD 2
1.25
1.3
PADD 3
1.2
1.0
PADD 4
1.5
1.4
PADD 5
Outside CA
1.25
1.2
The capital costs which would be incurred by refiners in order to comply with the
proposed sulfur standards must be amortized in order to combine this cost with recurring
V-37
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
operating costs to produce a total per gallon cost in future years. In past analyses, EPA used a
cost amortization factor of 0.173, which was based on a 10 percent return on investment (ROI), a
34 percent income tax rate, a 13 year economic and project life, and a 13 year depreciation life.
Regarding the 10 percent ROI, in 1997 we received comments from the automobile
industry that the ROI for capital employed in the refining industry was less than 10 percent.
They recommended use of an eight percent ROI, which we accepted and used to develop sulfur
control costs in EPA's Staff Paper on Gasoline Sulfur Issues. The 1993 National Petroleum
Study presented the actual ROI for the refining industry during the 1980's and early 1990's. ROI
averaged close to eight percent during that timeframe. Since 1992, the refining industry has
experienced a much lower ROI, averaging roughly three percent. However, these levels are
clearly depressed and we do not believe that these low levels should be projected into the future.
Thus, eight percent appears to be a reasonable level of ROI for assessing the impact of these
regulations on the refining industry. However, in assessing the impact of these regulations on
society, OMB Circular A-94 suggests that EPA use a seven percent discount factor in
determining the net present value of both costs and benefits. Thus, a seven percent ROI will be
used in determining the cost of reducing gasoline sulfur content for use in the cost effectiveness
and cost benefit analyses. In assessing the impact of these regulations on the refining industry,
the eight percent ROI will be used. Since the ROI of individual refiners can vary, we will also
evaluate the impact of some variation around the average of eight percent ROI and use a range of
six to 10 percent.
The 1993 National Petroleum Study also used slightly different estimates for the
economic and depreciation life of capital and for the income tax rate than those cited above. In
particular, the 1993 National Petroleum Study used an economic life of 15 years, a depreciation
life of 10 years and an income tax rate of 39 percent.18 Since the NPC study received a
substantial amount of peer review, we decided to use these financial factors from that study,
coupled with the above-mentioned estimates of ROI. The one exception is the elimination of the
income tax from the assessment of societal costs. Since income taxes are simply transfer
payments between various sectors of society, they are not included in societal costs. These
factors and the resulting capital amortization factors are summarized in Table V-26 below.
Table V-26. Economic Cost Factors Used and the Resulting Capital Amortization Factor
Amortization
Scheme
Societal Cost
Capital
Payback
Depreciation
Life
10 Years
10 Years
Economic
and Project
Life
15 Years
15 Years
Federal and
State Tax
Rate
0%
39%
Return on
Investment
(ROI)
7%
6%
8%
10%
Resulting
Capital
Amortization
Factor
0.11
0.12
0.14
0.16
V-38
-------
Chapter V: Economic Impact
/'/'. Fixed Operating Costs
Operating costs which are based on the cost of capital are called fixed operating costs.
These are fixed because the cost is normally incurred even when the unit is temporarily
shutdown. These costs are incurred each and every year after the unit is installed and operating.
Maintenance must be performed on all operating hardware to keep it in an operable
condition, and when it is running, to keep the unit operating efficiently. Maintenance cost is
estimated to be four percent of capital cost after adjusting to include the outside battery limit
cost, and after adjusting the capital cost for the higher labor cost due to the location. This factor
is based on the maintenance factor used in the ORNL refinery model.
Other operating costs are accounted for as well in terms of generic cost factors which
were taken from the ORNL refinery model. These factors are three percent of capital costs for
buildings, 0.2 percent for land, one percent for supplies which must be inventoried such as
catalyst, and two percent for insurance. These factors sum to 6.2 percent which is applied to the
total capital cost (which includes offsites, and the adjustment for location) to generate a
perennial fixed operating cost.
Annual labor costs are estimated using the cost equation in the ORNL refinery model.
Labor cost is very small; on the order of one ten thousandth of a cent per gallon.
/'/'/'. Variable Operating Cost
Variable operating costs are those costs incurred to run the unit on a day to day basis, and
are based completely on the unit throughput. Thus, when the unit is not operating, variable
operating costs are not being incurred. These costs are summarized in Table # V-27 below.
V-39
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-27. Summary of Costs Taken From EIA and NPC Data Tables *
Electricity
(c/KwH)
LPG (c/Gal)
Gasoline
(c/Gal)
Diesel
(c/Gal)
Residual Oil
(c/Gal)
Octane Cost
(cents)
Octane
Spread
(R+M)/2
Fuel Gas
($/MMbtu)
Hydrogen
Cost
($/MSCF)
PADD 1
5.9
19.7
27.0
25.2
17.9
4.3
5.7
3.75
2.5
PADD 2
3.9
18.4
25.9
25.7
15.2
2.8
5.2
3.75
2.5
PADD 3
4.2
16.5
24.9
24.7
15.4
3.5
5.4
3
2.0
PADD 4
3.4
17.8
28.9
29.6
10.8
11.4
5.2
4.5
3.0
PADD 5OC
5.4
19.7
30
28.6
16.1
9.0
4.6
3.75
2.5
* c/KwH is cents per kilowatt-hour, c/Gal is cents per gallon, (R+M)/2 is octane number as
determined by Research and Motor octanes divided by two, c/Gal is cents per gallon,
$/MMbtu is dollars per million British Thermal Units (Btu), $/MSCF is dollars per
thousand standard cubic feet.
Electricity is consumed in running pumps, air coolers, and other refinery equipment
electrically powered. Electricity costs were taken from the EIA publication "Monthly Electric
Utility Sales and Revenue Report with State Distributions." The 1997 industrial electricity costs
for individual states which comprise a PADD are averaged together to form a single individual
PADD-wide cost.
Fuel gas is consumed in running furnaces for heating up streams including the reboilers
used in distillation. Fuel gas cost is based on an estimation factor which is three dollars per
million British thermal units (BTU) for PADD 3,19 one quarter higher than that for PADDs 1, 2
and 5OC, and half higher for PADD 4. Steam demand is converted to BTU demand on the basis
V-40
-------
Chapter V: Economic Impact
that it is 300 pound per square inch (psi) steam, and that demand is presumed to be met with fuel
gas. Producing steam is presumed to demand 809 BTU per pound of steam required.
Cooling water is used for cooling streams, especially the vapor which comes off the top
of a distillation column and must be condensed for recycling it back into the top of the
distillation column. Cooling water is estimated to cost seven cents per 1000 gallons for PADD
3, seven and a half cents per gallon for PADDs 1, 2, and 5OC, and eight cents per gallon for
PADD 4.20
Octane loss is caused by saturation of unsaturated compounds including olefins and
aromatics which normally present in FCC gasoline. For each PADD, the cost of this loss is
estimated by using the price differentials between premium and regular grades. The price
differentials were based on the cost of gasoline grades sales to resale from the Petroleum
Marketing Annual for 1997. Octane ((R+M)/2) spread, which is the octane difference between
premium and regular grades, is from 1993 refining study by NPC.21 According to DOE, octane
spread has been increasing in recent years, so the cost for making up lost octane may be
overestimated to some degree.
Yield loss is the loss of gasoline to lower boiling point petroleum compounds. It
sometime occurs as gasoline is processed and tends to occur with hydrotreating. The conversion
of gasoline to lower boiling point compounds incurs a cost because gasoline brings a higher
profit than the other compounds. For this analysis, yield loss is presumed to occur by gasoline
being converted to liquid petroleum gas (LPG).22 Thus, yield loss is the resale price of gasoline
minus the resale price of LPG. The costs of gasoline and LPG are from the Petroleum Supply
Annual for 1997.
Finally, hydrogen costs also vary by PADD. The cost of hydrogen supply was estimated
for PADD 3, and then increased for other PADDs that typically have higher costs. Hydrogen
cost for PADD 3 is based on an average of refiners putting in their own hydrogen plants, which
could cost as much as three dollars per thousand standard cubic foot (MSCF), and purchasing
hydrogen as a commodity from a large hydrogen plant at a little more than one dollar per
MSCF.23 Based on this range of possible cost, PADD 3 would be expected to have access to
hydrogen supplied at a cost of about two dollars per MSCF. PADD 4 is assumed to have to pay
the more conservative cost of three dollars per MSCF, and the other PADDs are assumed to
incur a cost between PADDs 3 and 4, which would be $2.5 per MSCF. This analysis does not
consider numerous other possibilities of providing hydrogen at a reduced cost by using hydrogen
recovery technology (which would recover hydrogen from plant gas), or by increasing hydrogen
production from the reformer by converting high pressure reformers to low or ultra low pressure
reformers.
b. Determination of Blendstock Sulfur Levels
A sulfur balance is performed for each PADD average refinery to establish the volumes
and sulfur levels of blendstocks which contribute significantly to the pool sulfur level (FCC
V-41
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
naphtha, alkylate, straight run, and coker). The sulfur levels for these streams were volume-
weighted and compared to the pool gasoline sulfur level. If the calculated pool sulfur level did
not agree with the pool sulfur level, then the FCC gasoline sulfur level or volume was adjusted,
under the presumption that the noncalculated value is more likely to be correct. This exact
process is explained in detail below in the discussion on how the calibration was carried out for
each PADD.
The volumes and sulfur levels of the various blendstocks are established based on
information from different sources. FCC gasoline volumes and sulfur levels were taken from the
1996 API/NPRA survey, or the RFG baseline data base. The RFG data base was used when the
API/NPRA data for a PADD was incomplete or internally inconsistent, as described further
below. The RFG data base was not used first because of the difficulty in gathering the data, and
because not all refiners reported their blendstock sulfur levels. Coker gasoline volumes and
sulfur levels were taken from the 1996 API/NPRA survey. Straight run sulfur levels and
volumes are from the 1989 NPRA survey.
Alkylate sulfur levels are set at 10 ppm. This value was arrived at through an analysis of
alkylate sulfur levels from the baselines submitted for the RFG program, and a review of alkylate
sulfur levels in various refining consultant refinery models. From the 1990 RFG baseline
database, alkylate sulfur levels from nine refineries were averaged together. Then, the averaged
value, which was determined to be 22 ppm, was compared to the alkylate sulfur levels used in
several refinery models. The refinery models alkylate sulfur levels averaged about 10 ppm (the
values ranged from 0 to 25 ppm). The difference between the average sulfur level seen in the
RFG data base and the average alkylate sulfur levels from the various refinery models was
reconciled by presuming that if the average alkylate sulfur level is indeed about 20 ppm, then
refiners could decrease alkylate sulfur levels by increasing the severity or better managing
existing desulfurization of the alkylate blendstock. Other blendstocks, such as isomerate,
reformate, raffinate, dimate, poly gasoline, hydrocrackate, aromatics, butane and any oxygenates
which may be blended into gasoline, are all assumed to make a negligible sulfur contribution to
the gasoline sulfur pool.
The gasoline pool sulfur levels (not calculated from blendstocks) were taken from either
the API/NPRA survey or the RFG data base and were compared to the values calculated from
the sulfur-containing blendstocks. For simplicity reasons, the API/NPRA data base was
consulted first, however, for reasons explained below, sometimes the RFG database was
preferred.
PADD 1 - The 1996 API/NPRA survey only collected data from refiners which comprise half of
the gasoline production in PADD 1 (nine reported gasoline quality, and only five reported FCC
sulfur level); thus, it did not seem viable to use that survey data. Instead, the RFG baseline data
was used exclusively (based on data from 11 refineries). The average gasoline pool sulfur values
for each refinery were obtained from the 1995/1996 data reported by refiners to EPA. When all
the refineries' average gasoline sulfur values were averaged together, the average ended up being
215 ppm. The FCC gasoline sulfur values for each refiner were used to estimate the average
sulfur level of FCC gasoline for the PADD, which was estimated to be about 460 ppm (although,
V-42
-------
Chapter V: Economic Impact
this value seems low compared to the straight run sulfur level from the 1989 NPRA survey,
which was reported to be 330 ppm). The FCC sulfur level of any refinery was adjusted if the
1995/1996 gasoline sulfur level was significantly different from the level reported in the 1990
baseline submission. Based on the RFG baseline submissions, the FCC volume was calculated
to comprise 46 percent of the gasoline pool. The blendstock calculated pool sulfur level was
higher than the calculated gasoline sulfur level, so the FCC volume was adjusted downward from
46 percent to 42 percent to result in a pool sulfur level of 215 ppm.
These figures may need to be adjusted to account for the implementation of Phase II RFG
in 2000. Phase II RFG plays an important role for PADD 1 refiners since those refiners produce
more than 60 percent of its gasoline as RFG. The average gasoline sulfur level was calculated
for RFG in 1995 and 1996 found to be about 150 ppm. Since we expect RFG to be about 150
ppm, no changes in sulfur level are expected to occur to produce Phase II RFG. The PADD 1
blendstock sulfur levels and relative volumes are summarized in Table V-28.
Table V-28. PADD 1 Blendstock Sulfur Levels and Gasoline Pool Fraction
Sulfur (ppm)
Percentage of
gasoline pool
Contribution to
pool (ppm)
FCC
442
42
185
Alky late
10
10
1
Straight Run
343
4
14
Coker
3289
0.44
14
PADD 2 - The API/NPRA survey data for the gasoline pool sulfur level and the FCC sulfur and
volume was used. According to the survey data, PADD 2 FCC gasoline has a sulfur level of 924
ppm and it comprises about 27 percent of the gasoline pool. However, based on that FCC sulfur
level and volume and other blendstock sulfur levels and volumes, the gasoline pool would have a
sulfur level of 260 ppm which is lower than the pool average of 338 ppm based on the
API/NPRA survey. To account for this discrepancy, the FCC contribution to the gasoline pool
was increased to 35 percent. Since PADD 2's RFG production is only 11 percent, Phase 2 RFG
is presumed to have no affect on the average sulfur level of PADD 2. The PADD 2 blendstock
sulfur levels and relative volumes are summarized in Table V-29.
V-43
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-29. PADD 2 Blendstock Sulfur Levels and Gasoline Pool Fraction
Sulfur (ppm)
Percentage of
gasoline pool
Contribution to
pool (ppm)
FCC
924
35
323
Alky late
10
13
1
Straight Run
397
3.4
14
Coker
0
0
0
PADD 3 - According to the 1996 API/NPRA survey FCC gasoline comprises 35 percent of the
gasoline pool and the sulfur level of that blendstock is 722 ppm. When considering all the
blendstocks together, they result in a pool sulfur level of 271 ppm. However, the 1996
API/NPRA survey has PADD 3 pool sulfur levels at 305 ppm. To make the blendstock agree
with the pool sulfur level, the PADD 3 FCC gasoline volume was increased from 35 percent of
the pool to 40 percent. The PADD 3 blendstock sulfur levels and relative volumes are
summarized in Table V-30.
Table V-30. PADD 3 Blendstock Sulfur Levels and Gasoline Pool Fraction
Sulfur (ppm)
Percentage of
gasoline pool
Contribution to
pool (ppm)
FCC
722
40
288
Alky late
10
14
1
Straight Run
139
2.8
4
Coker
3255
0.42
14
PADD 4 - According to the 1996 API/NPRA survey, 31 percent of the gasoline pool comes from
FCC gasoline blendstock, and the sulfur level of that blendstock is 1100 ppm. When considering
the sulfur contribution from the other blendstocks, the pool average sulfur level is calculated to
be about 350 ppm. However, according to the 1996 API/NPRA survey the pool sulfur level was
about 260 ppm, and this pool sulfur level is corroborated by 1995/1996 gasoline sulfur data
reported by refiners to EPA. The PADD 4 FCC gasoline sulfur level from refiner baseline
submissions, after adjusting for changes in gasoline sulfur levels from when the baseline were
submitted in 1995/1996 (based on simple ratioing), averaged 760 ppm. This FCC sulfur level
was used and, combined with other blendstocks, resulted in a pool sulfur level of 263 ppm. The
PADD 4 blendstock sulfur levels and relative volumes are summarized in Table V-31.
V-44
-------
Chapter V: Economic Impact
Table V-31. PADD 4 Blendstock Sulfur Levels and Gasoline Pool Fraction
Sulfur (ppm)
Percentage of
gasoline pool
Contribution to
pool (ppm)
FCC
762
31
236
Alky late
10
12
1
Straight Run
122
21
26
Coker
0
0
0
PADD 5 OC - Based on the 1996 API/NPRA survey data, the FCC gasoline sulfur level was 666
ppm (based on only four refineries), and the volume was 38 percent of the entire gasoline pool.
However, when all the blendstock sulfur levels and volumes were combined together, the
calculated gasoline pool sulfur level would only average 256 ppm which is much lower than the
pool sulfur levels from the API/NPRA gasoline parameter data, which averaged 480 ppm. Based
on the RFG data base, the pool sulfur level for PADD 5 was 510 ppm, and the FCC gasoline
sulfur level for the 6 refineries was about 1200 ppm. The RFG baseline FCC sulfur level was
much more consistent with the average gasoline sulfur level and thus was used for cost
estimation. To match the blendstock sulfur levels with the RFG data base average pool sulfur
level (510 ppm), the fraction of FCC gasoline to the rest of the gasoline pool was increased from
38 percent to 42 percent. The PADD 5 outside of California blendstock sulfur levels and relative
volumes are summarized in Table V-32.
Table V-32. PADD 5 Outside of California Blendstock
Sulfur Levels and Gasoline Pool Fraction
Sulfur (ppm)
Percentage of
gasoline pool
Contribution to
pool (ppm)
FCC
1197
42
503
Alky late
10
10
1
Straight Run
41
5.9
2
Coker
0
0
0
Gasoline Volume - To estimate the aggregate capital and operating cost of desulfurizing gasoline
by PADD, and for volume weighting the separate PADDs to calculate the national average cost,
the gasoline production volumes for each PADD and the production and consumption values for
the Nation as a whole are used. These values are summarized below in Table V-33.
V-45
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-33. Projected Volume of Gasoline Produced by an Average Refinery in each
PADD, by Each PADD of Refineries and for the U.S.* in 2004
Gasoline
Produced by
Avg. Refinery
(MBbl/day)
Total Gasoline
Produced
(MMBbl/yr)
Gasoline
Consumed
(MMBbl/yr)
PADD 1
77
404
PADD 2
66
764
PADD 3
76
1430
PADD 4
19
107
PADD
5OC
27
166
U.S. OC
_
2872
3192
California gasoline not included.
2. The Cost of Desulfurizing Gasoline
a. The Cost of the Averaging Standard
The refinery blendstocks sulfur levels, the vendor desulfurization technology
information, the various cost inputs, and various desulfurization assumptions were combined
together in a spreadsheet to estimate the cost of desulfurizing gasoline from the base sulfur level,
down to various gasoline sulfur levels. A parametric analysis was undertaken to understand how
the cost varies in each PADD as the sulfur standard is made more stringent. Costs are estimated
for average sulfur standards of 150 ppm, 100 ppm, 80 ppm, 40 ppm and 30 ppm. The costs for
desulfurizing gasoline to each of these levels is summarized below in Table V-34.
V-46
-------
Chapter V: Economic Impact
Table V-34. Per-Gallon Cost of Desulfurizing Gasoline
Average Sulfur
Level (ppm)
150
100
80
40
30
ROI
6%
8%
10%
PADD1
PADD2
PADD3
PADD4
PADD
5OC
U.S. Avg.
Societal Cost (7 percent ROI and no Income Taxes)
0.8
1.3
1.5
2.1
2.3
0.7
0.9
1.0
1.2
1.4
0.7
0.9
1.0
1.3
1.4
1.2
1.7
2.0
2.8
3.2
1.6
1.9
2.0
2.3
2.8
0.8
1.1
1.2
1.5
1.7
Cost to Refiners of a 30 ppm Average Sulfur Standard
2.4
2.5
2.6
1.5
1.5
1.6
1.4
1.5
1.5
3.3
3.5
3.6
2.8
2.9
3.1
1.7
1.8
1.9
As seen in the above table, our analysis shows that the per-gallon cost of desulfurizing
gasoline to 30 ppm varies from PADD to PADD. PADDs 2 and 3 would experience lower costs
than the other PADDs. Because of the smaller size of the refineries which increases the cost of
installing capital, and because of the higher expense of refining, PADD 4 is expected to be the
most expensive, and would be about twice as much to desulfurize gasoline as PADDs 2 and 3. A
national average cost is calculated by volume-weighting the various PADDs. The result is an
average national societal cost of about 1.8 cents per gallon to desulfurize gasoline down to 30
ppm. The societal cost represents capital amortized based on a seven percent rate of return on
investment (ROI), and no income taxes; and this cost was used to calculate cost-effectiveness.
We also show that the cost would be 1.8 cents per gallon based on a typical capital recovery
scenario for the refining industry, which is based on an eight percent ROI and taxes included.
As a sensitivity, we also show the cents per-gallon cost for six percent and 10 percent ROI. Both
the societal cost and typical refinery cost are intended to represent the average cost across an
entire PADD. Individual refiners within a PADD are expected to experience costs which are
either higher or lower than these costs. The societal costs are shown in graph form in Figures V-
2 - V-7 at the end of this Section.
To help the reader better understand the cost of the program for a typical refinery, the
per-refinery costs for 150 ppm and 30 ppm are summarized in Table V-35, below.
V-47
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-35. Estimated Average Per-Refinery Capital and Operating Cost of Desulfurizing
Gasoline to 150 ppm and 30 ppm
PADD 1
PADD 2
PADD 3
PADD 4
PADD
5OC
U.S. Avg.
150 ppm
Capital
Cost
($MM)
Operating
Cost
($MM/yr)
30
8
23
6
21
7
10
3
17
12
22
7
30 ppm
Capital
Cost
($MM)
Operating
Cost
($MM/yr)
73
25
40
13
40
15
23
9
25
21
43
15
Table V-34 shows that, on average, refiners would have to pay out $43 million in capital costs
for each refinery to lower gasoline sulfur to 30 ppm. In addition, each refinery would incur
about 15 million dollars per year in operating costs. While the smaller refiners in PADD 4 are
expected to pay out less than other refiners, their costs are higher on a per-gallon basis. To meet
a 150 ppm standard, the capital and operating costs are about half as much as having to meet a
30 ppm standard. Once again, since these figures are averages, larger refineries with high
gasoline sulfur levels will experience higher costs, while smaller refineries with lower sulfur
levels will experience lower costs.
The estimated yearly aggregate cost to the country, and to importers, of meeting 30 and
150 ppm sulfur standards is summarized in Table V-36, below.
V-48
-------
Chapter V: Economic Impact
Table V-36. Aggregate Cost of Desulfurizing Gasoline to 150 ppm and 30 ppm
PADD 1
PADD 2
PADD 3
PADD 4
PADD
5OC
U.S.
Total
150 ppm
Capital
Cost
($MM)
Operating
Cost
($MM/yr)
380
90
640
160
810
310
140
40
260
80
2230
680
30 ppm
Capital
Cost
($MM)
Operating
Cost
($MM/yr)
930
290
1100
340
1850
630
340
110
390
150
Year which Capital Dollars are Expended
2001
2002
2003
2004
2005
4650
1520
Amount
($MM)
145
727
1018
1358
1455
U.S.
Total w
Foreign
Cost
2450
750
5100
1670
Table V-35 shows that the aggregate capital cost to the U.S. refining industry for meeting
the proposed 30 ppm sulfur standard is expected to total about 4.7 billion dollars. With the
implementation of an averaging, trading and banking program, these capital investments are
expected to be spread out over several years. The bottom of Table V-35 summarizes our forecast
of the capital dollars expended during each year which the refining industry is expected to make
investments. This level of capital expenditure is less than previous capital expenditures made by
V-49
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
the refining industry for environmental programs. In 1996, the Energy Information
Administration studied and reported on the capital investment made by the major energy
producing companies in the U.S. during the early nineties.24 During this time, these companies
invested from one to two billion dollars per year in capital for environmental controls for their
refining operations; this cost represented about one third of the total capital expenditures made
by refiners for their refineries. Considering that these expenses were incurred by less than three
quarters of the refining industry, we believe that a program requiring the entire industry to spend
up to one and one half billion dollars of capital costs per year over several years is not
unreasonable. The aggregate operating cost to the U.S. refining industry is expected to be about
1500 million dollars per year. When considering the cost to foreign refiners, the capital and
operating costs of this program would increase to 5.1 billion dollars and 1670 million dollars per
year, respectively.
b. Verification of the Desulfurization Cost Based on the Improved Technologies
Shell Oil engineers (who now work for Equilon) provided EPA their estimate of the cost
of a 40,000 barrel per day CDTECH unit. The Shell cost estimate showed substantially higher
costs in certain areas compared to the CDTECH estimates (based on a May 1998 CDTECH cost
estimation booklet)25. Later (September and December 1998), CDTECH provided updates on
their costs, and the most updated cost table was integrated into our cost analysis spreadsheet.
We compared the new CDTECH costs to the Shell Oil costs and the previous CDTECH costs,
and summarized the comparison here. The sulfur reduction case which we costed out is
consistent with the past cost comparison between the early CDTECH costs and the Shell
engineers' calculated costs, which is 90 percent FCC gasoline desulfurization. We presumed
that the comparison was done for a Gulf Coast refinery, which would be in PADD 3, thus the
cost inputs we developed for PADD 3 above were used here. The cost comparison is
summarized in Table V-37 below.
V-50
-------
Chapter V: Economic Impact
Table V-37. Summary of the Cost of Desulfurization by the CDTECH Process Based on 90
percent Desulfurization Severity
Capital Onsite
$/BBL
H2 Reqd. SCF/BBL
Octane Value CPG
Hydrogen Cost
($/MSCF)
Previous
CDTECH
costs
375
60
0.6
3
Shell
600
200
1.0
O
Adjusted Shell
costs2
660
80
1.0
3
EPA (costs
revised by
CDTECH Dec.
'98)
690
68
0.7
2
$/BBL FCC
Gasoline
Maint(5%&4%Cap)
($/BBL)
Catalyst ($/BBL)
Utilities ($/BBL)
H2 ($/BBL)
Octane ($/BBL)
Net Misc
Downgrades
($/BBL)
Total ($/BBL)
c/gal Gasoline Pool
(FCC fract is 0.34)
0.10
(5%)
0.10
0.07
0.18
0.25
0
0.70
0.57
0.16
(5%)
0.10
0.10
0.60
0.42
0.10
1.48
1.20
0.16
(5%)
0.10
0.10
0.24
0.42
0.10
1.12
0.91
0.09
(4%)
0.15
0.19
0.13
0.27
0
0.83
0.67
As depicted in the above table, CDTECH's revised capital costs are substantially higher
than the initial costs, and, after adjusting the Shell costs to include the cost for a hydrogen
compressor, the revised CDTECH and the Shell capital costs are essentially the same. The
z Costs are adjusted to add a recycle compressor which, according to the Shell engineers, would
reduce hydrogen loss by over fifty percent, add to the capital cost by approximately 10 percent (about two million
dollars), and would, according to Shell, add a small cost to downgrades (not specified).
V-51
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
operating costs are closer now as well. Improved agreement in the operating costs occurs
because CDTECH's revised costs are somewhat higher in operating cost, and the adjustment in
Shell's hydrogen consumption to account for the addition of a hydrogen compressor dramatically
reduced the hydrogen cost. There are still cost differences between the Shell costs and the costs
we develop in our analysis. Most of the differences occur because of three cost factors.
The first cost factor which differs is the cost of hydrogen supply. Shell used three dollars
per MSCF compared to two dollars per MSCF in our analysis. In a conversation with one of the
Shell engineers, he stated that their hydrogen cost which they used was conservative. This
suggests the possibility that their cost estimating procedure may be conservative to provide a
safety factor for their cost analysis. Another cost factor which differs is the cost of maintenance.
Shell used a five percent cost rate while we used a four percent rate. Similarly, Shell later
informed us that the five percent maintenance cost factor could be conservative and that a four
percent factor is also a reasonable factor to use. For the octane cost factor, Shell used a one cent
per gallon factor while we used a 0.7 cents/gallon cost factor. We believe our estimate is
reasonable because it is based on the actual cost of making octane in PADD 3. In our analysis,
the cost of making octane is higher in other PADDs and considering those other costs would
increase our cost somewhat making our cost closer to the cost used by Shell. Finally, Shell
added a 0.1 dollar per barrel cost for downgrades, which provides for potential yield loss from
the CDTECH unit. Shell said that they add the factor to account for all losses from the unit. We
called CDTECH to ask them whether there are any losses from their desulfurization unit, they
stated that there is none. We are presuming that there is none. If the Shell costs were calculated
using these same cost factors which we used, their operating cost would decrease to 0.78 dollars
per barrel for the FCC gasoline and 0.63 cents per gallon for the entire gasoline pool, which is
essentially identical to our costs.
In summary, the revised CDTECH costs for their desulfurization unit brings their costs
much closer to the Shell costs. Since the Shell analysis of the CDTECH may have some
conservative cost factors involved, adjusting their analysis for these factors closes the remaining
gap between the two analyses. Shell engineers' review of the cost of using CDTECH CDHydro
and CDHDS corroborates the revised costs provided CDTECH, which corroborates that portion
of our analysis.
We have no third party verification of Mobil Oil's cost factors for their third generation
Octgain process. However, Mobil has monitored its own track record for estimating the
performance of a full scale unit based on pilot plant data. They went through this process two
times since they created two different generations of the Octgain process before this third
generation was created. Based on this experience, Mobil Oil feels confident that their process
will operate in a refinery as claimed.
c. Future Cost of Desulfurization
Like any refinery processing unit which was newly installed, the per-barrel cost will
normally decrease over time. We discussed how this change in cost would occur with several
V-52
-------
Chapter V: Economic Impact
different refining industry consultants who cited the following reasons.
Two of the consultants stated that the per-gallon costs could be expected to decrease
further through engineering improvements in the process. Normally, the vendor which licensed
the technology will discover engineering changes to the unit that would reduce its operating cost,
although the refinery engineering and operations staffs can also make such discoveries.
Engineering changes would be expected to occur in the catalyst technology which would lower
operating cost such as reduced hydrogen utilization, and reduced octane and yield loss.
Two of the consultants mentioned that a another type of cost reduction can occur
incrementally over time due to debottlenecking of the process throughput. The debottlenecking
of this unit would occur in step with the debottlenecking of other gasoline producing units, such
as the FCC unit, to help increase gasoline production to meet increased gasoline demand. Such
increases in throughput would result in decreased per-barrel fixed operating cost, such as
operating labor and maintenance costs, and insurance and other similar costs.
One consultant stated that refinery operations personnel will learn to operate the process
more efficiently. These improvements would most likely help to reduce operating cost, such as
improved energy utilization, reduced electricity demand and decreased operating labor and
maintenance costs. However, the other two consultants seemed to think that these sort of
improvements are less likely, as refiners have learned to already squeeze the most efficiency
from their refinery units.
Processing unit improvements can also reduce the capital cost of the improved
technologies. Capital improvements can primarily be taken advantage of when the unit is first
installed. However, units installed for 2004 will already be sunk investments if improvements to
the design of these technologies are discovered later. Thus, capital improvements would
probably not be taken advantage of until new investments are made in these desulfurization
processes.
For this analysis, we presumed that there would be reductions in costs in the ways stated
by at least two of the refinery consultants. First, there is a presumed reduction in operating cost
due to an improvement in catalyst technology. Similar to the estimate of future motor vehicle
costs, operating costs are presumed to decrease by 20 percent after two years. This improvement
is expected to occur in the catalyst cost, hydrogen cost, and decreased octane and yield losses.
This improvement in operating cost is presumed to only happen once, although the reduction
applies to additional throughput created through debottlenecking.
A second reduction in cost occurs in fixed operating cost because the unit is
debottlenecked to keep up with increased gasoline demand. Since there are no new refineries
being built, the increase in gasoline demand is presumed to occur by the existing refiners which
currently produce gasoline for the U.S. Gasoline consumption is presumed to increase at the
same rate as VMT is presumed to increase, and this growth rate is 2.05 percent per year. Since
capital is sized larger than necessary by a 15 percent margin, the first eight years of
debottlenecking are presumed to occur with no new capital expenditures. Then, as capital must
V-53
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
be invested to increase throughput, the newly invested capital is presumed to cost one-third of
the originally invested capital on a per-barrel basis. The ratio of one-third presumes that
debottlenecking would cost about half of the inside battery limit capital cost. These
debottlenecking capital costs are also amortized at a seven percent rate of return over a 15 year
period. The variable operating costs for the increased barrels desulfurized are presumed to be
the same on a per-barrel basis as the original throughput. However, the fixed operating costs of
the original equipment are presumed to stay the same, thus the same fixed costs are spread over a
larger volume of gasoline produced.
A third type of cost reduction occurs for future capital investments in desulfurization
units. This new investment is assumed to be made after 15 years, which is the assumed
economic life and project life of these units and the point at which they would have to be
replaced. At that point, the capital cost is presumed to be 20 percent lower than the cost in 2004.
Presuming that refiners would reinvest in capital after 15 years is probably conservative since
most refineries today are still using originally installed equipment which was erected 20, 30 and
sometimes even 50 years ago.
As expected, the implementation of these cost assumptions results in a decreasing cent-
per-gallon cost over time. The estimated future national cost of desulfurizing gasoline are
summarized below in Table V-38.
V-54
-------
Chapter V: Economic Impact
Table V-38. Projected Future Average Per-Gallon
National Cost of Desulfurizing Gasoline to 30 ppm
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2015
2020
2025
2030
Cost
(cents per gallon)
1.68
1.67
1.54
1.52
1.51
1.49
1.47
1.45
1.44
1.42
1.40
1.38
1.30
1.23
d. Comparison with Previous Cost Estimates
Over the last several years, EPA and other organizations have estimated the cost of sulfur
control. In our recent technical report entitled "EPA Staff Paper on Gasoline Sulfur Issues," we
provided a cost estimate for reducing sulfur in gasoline. That cost is summarized here in Table
V-39 along with our current costs.
V-55
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-39. Cost of Desulfurizing Gasoline by Refiners in PADDs 1 & 3 Reported in the
"EPA Staff Paper on Gasoline Sulfur Issues," and our Current Costs.
Previous Cost
(c/gal)*
Current Cost
(c/gal)**
Base to 150 ppm
1.1 - 1.8
0.7
Base to 100 ppm
1.9-3.0
1.1
Base to 40 ppm
5.1 -8.0
1.6
**
Based on Octgain 125 (second generation) technology, and calculated with the ORNL
refinery model with excessively high reformate sulfur levels, among other problems
which tended to overestimate costs.
Based on improved gasoline desulfurization technologies assuming a typical refinery
capital cost recovery.
Most of the difference in cost between these two cost estimates can be explained by a couple
factors. The most important factor is the type of desulfurization hardware which we presumed
would be used by the refining industry. For our previous cost study, we worked with the
Department of Energy to develop costs using refinery model run by the Oak Ridge National
Laboratory (ORNL). That refinery model chose Mobil Oil's Octgain 125, which is the second
generation of the Octgain process. Octgain 125 must be operated under very severe conditions
(higher temperature and pressure) to realize the octane recovery which the process is designed to
deliver. However, the more severe conditions also results in higher capital and operating costs
than that incurred by these improved gasoline desulfurization technologies. To quantify the cost
reduction of the improved technologies relative to what we were modeling with previously, we
put the inputs of the older Octgain process into our spreadsheet and developed a cost curve at the
different sulfur levels modeled. Since we only are trying to get a sense of the cost reduction for
the older Octgain process relative to the improved technologies, we only developed the cost
curve for Octgain 125 for PADD 3. A comparison of the cost of desulfurizing gasoline is
summarized below in Table V-40.
Table V-40. A Comparison of the Per-Gallon Gasoline Desulfurization Cost of Improved
Desulfurization Technologies to that of the Earlier Mobil Octgain Process for PADD 3
Mobil Octgain
125
Improved
Desulfurization
Technologies
750 ppm
1.1
0.7
100 ppm
1.7
1.0
80 ppm
2.0
1.1
40 ppm
2.6
1.3
30 ppm
2.9
1.5
V-56
-------
Chapter V: Economic Impact
This comparison shows that the difference between our current costs and our previous
costs for 150 and 100 ppm can mostly be explained by the difference in the technologies we used
in our modeling. To reach a low gasoline sulfur standard of 40 or 30 ppm, the improved
desulfurization technologies are nearly 50 percent less costly than the older technology. This
difference explains a part of the cost difference between these two studies, but not all the cost
difference. Probably the next most important factor is a problem which was discovered with the
refinery model which we were modeling with at the time. That model assigned reformate a
sulfur level of 35 ppm, which is normally 1 ppm or less. Achieving low sulfur levels with the
previous Octgain process caused a significant loss in octane and yield. Increasing either the
volume or octane of reformate would have been a likely source of the needed octane. However,
the refinery model did not select either of these options, due to the fact that the reformate sulfur
level was so high and could not be reduced. Thus, the refinery model had to find other more
costly ways to make up the octane losses caused by desulfurization, which likely increased cost.
The same refinery model was later run with reformate sulfur levels reset to low levels. This run
showed that this problem with the reformate sulfur level may have increased the cost of
desulfurization by as much as 1.5 cents per gallon. Thus, an unreasonably high reformate sulfur
level explains another large portion of the difference between the two studies.
Another possible reason for this difference in cost between the two studies is that we are
estimating costs using a simpler refinery model which focuses primarily on the sulfur content of
gasoline, instead of a more sophisticated refinery model, which attempts to optimize production
volumes and quality all at the same time. The advantage of the simpler model is that the source
of all costs is clear and can be easily evaluated. The disadvantage is that some aspects of
refinery operation, such as making up lost octane or gasoline yield, is handled in a fairly simple
fashion (e.g., by adding the current market cost of increasing octane or of producing gasoline).
The disadvantage of the more complex linear programming refinery models is that it is very
difficult for anyone except the operator of the model to understand why the model is making
certain decisions or the cause of many of the projected costs. The example of the over-estimated
reformate sulfur level is a case in point, as this problem and its impact on costs was not at all
obvious. On the other hand, the more complex models attempt to more realistically simulate the
actions which would be required for refiners to make up lost octane or gasoline volume. This
presumes that the model reflects all of the flexibilities and constraints facing refiners in
accomplishing these goals. Again, this is difficult to determine given the complexity of both
refining and these refinery models.
For example, some uncertainty exists concerning the cost of supplying hydrogen. Since
refinery cost data is not available for estimating hydrogen cost, we are using cost factors which
we believe are reasonable, but may slightly over or underestimate costs. A more complex
refinery model would include one or more hydrogen producing processes. However, the
fundamental operating and capital costs of these processes are usually not published and cannot
therefore be evaluated.
Differences in capital recovery used to develop the two different sets of costs shown in
Table V-39. provide a negligible impact on the cost difference.
V-57
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
An important point deserves to be made concerning the improvement which Mobil Oil
has made with its Octgain technology over the last several years. As seen in Table V-39, the
desulfurization cost to achieve a low sulfur gasoline with their later technology is about half as
much as with their earlier technology. This improvement over a several year period of
timeframe corroborates our assumptions above that desulfurization cost will decrease in future
years.
3. Other Effects of This Program
a. Effect of the Cap Standard
In addition to the 30 ppm averaging standard, we are proposing a 80 ppm per-gallon
standard. The per-gallon standard or cap on sulfur level provides an additional challenge to
refiners by preventing them from producing moderate and high sulfur batches of gasoline. While
the averaging standard would force refiners producing higher sulfur gasoline to produce lower
sulfur gasoline on average, which would be comprised mostly of batches of low sulfur gasoline
along with occasional batches at a moderate sulfur level, a sulfur cap would preclude refiners
from producing a single batch of gasoline with moderate sulfur levels that would exceed the cap.
High sulfur batches of gasoline would likely be produced when the refinery is
experiencing problems with the added desulfurization unit, or problems with other units within
the refinery responsible for, or associated with, desulfurizing gasoline blendstocks. However,
changes in other refinery operations or other factors can also result in varying amounts of sulfur
in gasoline. These include changes in feedstock qualities, changes in products produced,
changes in throughput, process fluctuations, and changes in hardware processing efficiency
caused by breakdown in equipment or catalyst inactivation.
During conversations some time ago with the economics committee of API, we discussed
how we could estimate the cost of a cap standard. The committee's response was that the cost of
the cap standard could be estimated by estimating the average sulfur level which would result
from the cap standard. Later on, API sent us a letter which stated that the relationship between a
cap standard and the resulting average sulfur level could be estimated from the variation in
current gasoline sulfur levels presuming that the cap would represent the 90th percentile of that
variation. The cost of meeting the cap standard could be estimated by estimating the cost of
reducing gasoline sulfur to meet the average sulfur level determined by this relationship. Based
on this advice, we analyzed this relationship using gasoline batch sulfur levels provided to EPA
for the Reformulated Gasoline Program. We also compared the proposed API methodology for
estimating the relationship between the cap and averaging standards to current capped sulfur
levels. We put the results of that study in the EPA Staff Paper on Gasoline Sulfur Issues.
The analysis showed that if a 80 ppm cap standard were established, the resulting
averaging standard would be in the range of 30 ppm to nearly 50 ppm. Because the averaging
standard is at the lower end of this range, this suggests that the cap standard would not control
the average gasoline sulfur level. Instead, the 30 ppm average standard would be the primary
V-58
-------
Chapter V: Economic Impact
controlling standard, and if the cap standard did not exist, then while meeting the 30 ppm
averaging standard, refiners would occasionally produce gasoline which exceeded the 80 ppm
sulfur level (our analysis shows probably about five percent of the time). The addition of the 80
ppm cap would require refiners to modify their refinery operations further to not produce batches
of gasoline with sulfur levels that would exceed the cap standard.
Since refiners are likely producing these high sulfur batches because they are not trying
to control gasoline sulfur now, stopping the production of these batches may not be a difficult
task. However, in our analysis of the relationship of the cap and averaging standards, the
refiners which have lower sulfur levels now are probably at that level because they are using
sweet crude oils, or don't have a FCC unit which elevates the sulfur level in gasoline. Sour
crude refiners, in the day-to-day variances in their refining operations, may have a more difficult
time preventing the occurrences of high sulfur gasoline batches. Their gasoline sulfur levels is
expected to be very high if their desulfurization units were to fall into disrepair, or would vary
more widely when any of the situations summarized above which cause variance in gasoline
sulfur levels were to occur. To manage this situation in those refineries, the refinery managers
would likely have to do a better job managing the entire refinery, not just the gasoline
desulfurization unit, to deliver low sulfur gasoline. This improved operations management
would likely involve changes in the computer systems which control the refinery operations.
There would likely have to be better management of the maintenance performed refining
processing units. Refiners would likely focus in improving the operations and maintenance of
critical units which divert sulfur into gasoline, or remove sulfur. However, after this is done, the
refinery would likely recoup at least some, or perhaps even all of the cost disbursed to make
these improvements in refinery operations from other improvements in refining operations.
Another change which refiners could make in their refining operations is to invest in a
gasoline sulfur analyzer. Such analyzers would enable them to meet the per-gallon cap at the
lowest possible cost. Refiners normally have to send a gasoline sample out to a lab to determine
the actual sulfur level. However, the lag time between when the sample was taken and when
they receive the results provides refiners with some uncertainty on whether the gasoline it is
producing is indeed meeting the cap standard. This uncertainty could cause refiners to produce
gasoline with lower sulfur levels than necessary to ensure that the cap standard would be met.
For this reason, refiners may choose to purchase a gasoline sulfur on-line analyzer. This
analyzer would keep the refinery manager up to speed on the exact sulfur level in their gasoline.
This information would empower refiners with confidence that they will consistently meet the
cap standard. According to an analyzer manufacturer which makes such a device, the cost for a
gasoline sulfur analyzer would be about 50,000 dollars, and to install it would cost another 5000
dollars. Compared to the capital and operating cost of desulfurizing gasoline, the cost for this
addition would be trivial.
If the gasoline desulfurization unit were to break down, or if a number of other problems
were to plague the refinery, the refinery would probably be producing gasoline which would
exceed the cap standard. Thus the refinery manager would have to take action preventing the
sale of off-specification gasoline. The most obvious near term solution would be to blend the
gasoline blendstocks together which it can to produce on spec gasoline. If the FCC gasoline
V-59
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
hydrotreater went down, the rest of the gasoline blendstocks could probably be blended together
to produce on-spec gasoline. However, a portion of the normal gasoline stream, which would be
the FCC gasoline, would have to be dealt with until the hydrotreating unit could be brought back
on line. There are several potential solutions to this problem. One would be to store the FCC
gasoline blendstock until the unit was back on line. Then the stored blendstock could be either
run through the desulfurization unit, or blended back with gasoline at a rate which would ensure
that the gasoline would still meet the cap standard. If the refinery does not already have a spare
tank in which it could store the high sulfur gasoline blendstock, then the refiner would have to
build one. Another possibility would be to sell off the blendstock to another refiner, who had the
excess desulfurization capacity to process it or blend it in with their gasoline. Since refiners
design the desulfurization units using a safety design factor, this excess capacity can be used to
process this excess feedstock.
b. Other Effects on the Refining Industry
If a gasoline sulfur program is finalized, oil companies are expected to take a number of
steps to maximize their profitability in the period after the program is implemented. First, and
foremost refiners will try to minimize their costs by investing in the most cost-effective refinery
changes. Despite frugal choices, almost every refiner will face capital and increased operating
costs, and the refiner will try to pay off those costs. The most obvious step to recover those costs
would be to increase the price of gasoline. However, in a competitive market, the effect of an
increase in refiners' cost on the price of gasoline depends on both the market supply and
demand. If market demand is "inelastic" (not sensitive to changes in price), then one would
expect the price of gasoline to rise by the full amount of the cost increase, and refiners would
recover all their operating cost and incrementally recover their capital costs. Since gasoline
demand is not perfectly inelastic, some reduction in demand would be expected due to the price
increase in gasoline. This would lead to a corresponding small decrease in the price of gasoline,
which would erode refiners' ability to recover their costs. In addition, changes in supply due to
imports from abroad would change the supply curve which would also affect refiners' cost
recovery; increased imports reduce domestic refinery cost recovery, while decreased imports
increase cost recovery.
Overall, the U.S. refining industry is currently producing gasoline and other refined
products at full capacity.26 This situation, coupled with ever increasing demand for gasoline,
would generally produce reasonable refining margins. However, increasing imports of gasoline
over the past few years appears to be keeping prices lower, as refining margins have been
relatively low on average over the past three to four years.
Both Canada and Europe are major exporters of gasoline and other refined products into
the U.S. market. Stringent sulfur requirements in Europe, and similar proposed requirements in
Canada, will phase-in about the same time as the proposed U.S. standards would phase in. These
required improvements in fuel quality will increase costs in these areas, as well as in the U.S.
This will support an increase in the price of gasoline in the U.S. sufficient to cover capital, as
well as operating costs.
V-60
-------
Chapter V: Economic Impact
A significant amount of gasoline is also imported into the U.S. from the Middle East and
South America. We do not expect gasoline sulfur standards to take effect in these gasoline
exporting countries in the near future. Thus, refiners in these countries could reblend their
gasoline to be able to export very low sulfur gasoline to the U.S., while selling higher sulfur
gasoline elsewhere. Under this scenario, their costs could be significantly less than those of
domestic refiners who essentially have to desulfurize their entire product. However, the
potential volume of low sulfur gasoline would be limited. Also, these refiners also export to
eastern Canada, which will likely have its own low sulfur specification. Thus, the ability of
these importers to flood the market with inexpensive, low sulfur gasoline appears to be limited.
While margins may improve which would help domestic refiners recover the cost of
meeting the proposed gasoline sulfur requirements, there are still differences between refiners
which would cause the per-gallon cost for some to be higher than others. This may be due to:
having to pay a premium for capital costs due to their location, starting from a higher sulfur
baseline, or facing diseconomies of scale due to small size. In order to remain profitable, high
cost refiners would be expected to take further steps to reduce their costs.
Refiners could adopt a whole array of changes which may help them meet the sulfur
standard, at a reduced overall cost. These changes include changing crude oil supply, optimizing
other feedstock use, cost cutting of existing operations, opting to use processing outside the
refinery, improvements in transportation and marketing of product, and changing the consumer
market.27 Refiners could choose to merge their refining operations with other refiners. Merging
of refinery downstream operations (the refining and marketing portions of the oil industry) is
already occurring across the industry as a means to reduce administrative costs and optimizing
the production and distribution of common products.28 This practice has already been occurring
because the return on investment for the refining portion of the industry has been low for some
time.
It is possible that the projected per-gallon cost for a specific refinery to desulfurize
gasoline may be high enough relative to their ability to pay that a refiner might conclude that it is
in their best financial interest to sell the refinery. Over the last several decades, there have been
numerous refinery sales as refiners have determined that they are no longer capable of making an
acceptable level of profit, and, thus, have put the refinery up for sale.29 Many of the refineries
sold have been purchased by independents (refiners who are not vertically integrated). Because
of their flexibility and the relative availability of crude oil and other feedstocks, such as residual
oil, these independents have been able to profitably operate these refineries. If a buyer is not
found, refiners might be compelled to close the refinery, if no provisions were available to
prevent such closures.
However, this proposed rule contains multiple provisions which are intended to prevent
refinery closures due to financial hardship. The small refiner provisions extend the time which
small refiners would have to meet the sulfur standards. Additional time would allow them to
improve their financial standing, obtain a loan or another financial source for their capital
expenditures, and employ desulfurization technology developed later on or take advantages of
improvements made with existing desulfurization technology. Similarly, refiners which do not
V-61
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
fall under the small refinery definition can enjoy some of these same temporal benefits through
the Averaging, Banking and Trading program (ABT). The ABT program allows a refiner to
phase-in the proposed program across its refineries to its best financial advantage, or gain even
more leeway through trades for sulfur credits.
Based on this qualitative review of cost recovery by the refining industry and the benefits
of the proposed small refiner and ABT provisions, we do not expect refineries to close as a result
of the implementation of the proposed sulfur standards.
c. Refinery Energy and Global Warming Impacts
We estimated the increase in energy consumption in refineries expected to occur from
desulfurizing gasoline to 30 ppm by analyzing the specific impact on PADD 3 refineries. Also,
consistent with our cost estimation methodology, we performed the analysis presuming that only
improved gasoline desulfurization technologies would be used. For this analysis, we first
established a baseline energy consumption value for PADD 3 refineries using 1994 Energy
Information Administration data, which is the most recent energy consumption data available.
We increased the 1994 energy consumption by 2.05 percent per year until 1997, which is the
base year of the analysis. (The value of 2.05 percent per year is the projected growth rate for
gasoline consumption). This energy consumption calculation is summarized below in Table V-
41.
V-62
-------
Chapter V: Economic Impact
Table V-41. Energy Consumed by PADD 3 Refineries in 1994, Projected to 1997
Energy Type
Crude Oil
LPG
Distillate
Residual Oil
Still Gas
Petroleum Coke
Natural Gas
Coal
Purchased Electricity
Purchased Steam
Hydrogen
Other Products
Total in 1994
Total in 1997
(Estimated)
Energy Consumed
OMBbls
660 MBbls
54 MBbls
998 MBbls
112,538 MBbls
3 8, 152 MBbls
487,115MMCuft
0 MStTons
20,602 MMKwH
ll,970MMLbs
68,962 MMScf
252 MBbls
BTU Value
-
3.64MMBtu/Bbl
5.83 MMBtu/Bbl
6.29 MMBtu/Bbl
6.00 MMBtu/Bbl
FOE
6.02 MMBtu/Bbl
FOE
1.03MBtu/CuFt
-
3.41 MBtu/KwH
0.809MBtu/Lb
0.305 MBtu/Scf
6.00 MBtu/Bbl FOE
MMMBTUs
Consumed
0
2399
315
6274
675,200
229,800
501,200
0
70.3
9680
21,000
1510
1,438,000
1,528,500
Table V-40 shows that the energy consumed by PADD 3 in 1997 is estimated to be 1,500 trillion
BTUs.
The increase in energy consumed by desulfurization of the FCC gasoline is calculated by
adding up the fuel gas, steam and electricity (in terms of British thermal units (BTUs)) consumed
during the desulfurization. First there is the energy consumed running both CDTECH and
Octgain processing units. Consistent with how the cost of desulfurization was estimated, each
desulfurization technology was presumed to handle half the PADDs desulfurization needs. Then
the octane and hydrogen demand had to be met. For both CDTECH and Octgain, extra reformer
capacity in PADD 3 was presumed to produce the octane and hydrogen needed for
desulfurization. The amount of additional reformer processing capacity needed was based on
hydrogen demand, which produced more octane makeup than needed. This estimation
methodology likely overestimates the energy consumed since most refiners would probably run
V-63
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
the reformers to make up the octane needed. They would then obtain the additional hydrogen
needed from excess hydrogen going to plant gas, and make up the refinery plant gas energy loss
due to the recovered hydrogen from cheap, unrefined petroleum streams, or by combusting
natural gas. However, we had insufficient data for estimating hydrogen recovery from plant gas.
Alternatively, reformers could obtain hydrogen from hydrogen plants which consume less
energy per quantity of hydrogen produced than the reformer. Finally, half of the Octgain
desulfurization processes installed are presumed to need splitters, or distillation columns, to
fractionate the FCC gasoline. This additional energy demand is accounted for as well. This
presumption may overestimate costs as well for two reasons. First, to get down to 30 ppm, many
refiners would likely feed the entire feed to the Octgain unit, and not use a splitter. Second, the
splitter data upon which we based our energy demand probably boils off the entire feed, which
would not be necessary in this case since only the light ends may have to be boiled off for
sending the heavier compounds to the Octgain desulfurization unit. A summary of the estimated
CDTECH and Octgain energy and hydrogen demands in PADD 3 is summarized in Tables V-42
and V-43, respectively.
Table V-42. Estimated Yearly Energy and Hydrogen Demand of CDTECH
Desulfurization Units in PADD 3
CDTECH Utility
Demands
Electricity
Fuel Gas
Hydrogen
Reformer
Electricity
Fuel Gas
Steam
Total
Process
Demand
0.5 KwH/Bbl
55 MBtu/Bbl
69 Scf/Bbl
2.6 KwH/Bbl
0.048 FOE/Bbl
75 Lb/Bbl
Yearly
Throughput
240 MMBbls
240 MMBbls
240 MMBbls
18 MMBbls
18 MMBbls
18 MMBbls
BTU Conversion
Factor
3.41 MBtu/KwH
-
-
3.41 MBtu/KwH
6 MMBtu/Bbl
0.809 MBtu/Lb
Energy and
Hydrogen
Consumed
415MMMBtu
13,400
MMMBtu
16,800 MMScf
160 MMMBtu
5240 MMMBtu
11 00 MMMBtu
20,300
MMMBtu
V-64
-------
Chapter V: Economic Impact
Table V-43. Estimated Yearly Energy and Hydrogen Demand of OCTGAIN
Desulfurization Units in PADD 3
OCTGAIN
Utility Demands
Electricity
Fuel Gas
Steam
Hydrogen
Splitter
Electricity
Fuel Gas
Steam
Reformer
Electricity
Fuel Gas
Steam
Total
Process
Demand
3.6KwH/Bbl
17MBtu/Bbl
50 Lb/Bbl
125 Scf/Bbl
2.5 KwH/Bbl
0.015 FOE/Bbl
10 Lb/Bbl
2.6 KwH/Bbl
0.048 FOE/Bbl
75 Lb/Bbl
Yearly
Throughput
190MMBbls
190MMBbls
190MMBbls
190MMBbls
190MMBbls
190MMBbls
190MMBbls
26 MMBbls
26 MMBbls
26 MMBbls
BTU Conversion
Factor
3.41 MBtu/KwH
-
0.809 MBtu/Lb
-
-
3.41 MBtu/KwH
6 MM Btu/Bbl
0.809 MBtu/Lb
3.41 MBtu/KwH
6 MM Btu/Bbl
0.809 MBtu/Lb
Energy and
Hydrogen
Consumed
2950 MMBtu
4080 MMMBtu
9710MMMBtu
23,700 MMScf
8 10 MMMBtu
8540 MMMBtu
770 MMMBtu
230 MMMBtu
7400 MMMBtu
8 10 MMMBtu
32,500
MMMBtu
As these tables show, the average increase in energy demand for the improved gasoline
desulfurization technologies, including other changes needed in the refinery to desulfurize
gasoline, is estimated to be about 53 trillion BTU's in 1997. This increase in energy use is about
3.4 percent higher than the baseline PADD 3 energy consumption. For the U.S. outside of
California, the refining industry is estimated to consume 3000 trillion BTUs per year.aa Thus the
increase in energy demand for the U.S. refining industry, based on PADD 3 and using the 3.4
percent factor calculated above, is estimated to be about 102 trillion BTUs per year. If the
additional energy consumed by refiners producing low sulfur gasoline for importing gasoline
into the U.S. is considered, the total increase in energy consumed increases to about 122 trillion
aaThis estimate is based on the presumption that PADD 3 consumes 50 percent of the energy in the U. S.
outside of California.
V-65
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
BTU's per year.
We next estimated the amount of global warming gas emissions that would be emitted to
meet the proposed 30 ppm gasoline sulfur standard. The basis for the estimate is an estimate of
carbon dioxide emissions emitted from the combustion of fuels, which is the source of most all
refinery energy and, thus, is presumed to be the source of most all refinery emissions of carbon
dioxide. The carbon dioxide emission factor is 65,000 grams of CO2 per million Btu of fuel
consumed, which is based on the combustion of half natural gas and half liquid petroleum gas
(LPG is presumed to emit the same quantity of carbon dioxide per volume fuel consumed as
refinery plant gas).30 For simplicity, this analysis assumes that all BTUs consumed in a refinery
are produced by these fuel sources. On this basis, in 2004, CO2 emissions from PADD 3
refineries would increase by 3.4 million tons under the proposed 30 ppm sulfur standard. Across
the entire domestic refining industry, carbon dioxide emissions in 2004 would increase by 6.9
million tons. Considering overseas refiners who export gasoline to the U.S., CO2 emissions
would increase by 7.5 million tons in 2004, or 2.1 million tons (1.9 million metric tons) of
carbon emissions.
This increase is a one-time step increase which represents 0.03 percent of the projected
worldwide CO2 emissions inventory in 2004 which is 29.4 billion tons of CO2 per year. This
increase also represents 1.2 percent of the total projected increase in worldwide CO2 emissions
in 2004 over 2003, which would be 652 million tons. After the step increase, the CO2 emissions
increase due to gasoline desulfurization for this program is expected to increase only at or
slightly lower than the rate of increase in gasoline demand, which is about two percent. This
further increase in CO2 emissions associated with gasoline desulfurization in 2005 and beyond
would represent only 0.02 percent of the projected annual growth in worldwide CO2 emissions.
V-66
-------
Chapter V: Economic Impact
5
4
"c"
Ł
"re ,
.0) 3
7/5
0)
& 2
•4-1
t/>
O
O
1
0
1
i
N
X
^_
"v
\
N^
D 40 80 120 160 200 240 280 320 360 400 440 480 520
Sulfur Concentration (ppm)
Figure V-2. Cost of Reducing Gasoline Sulfur in PADD 1
(Costs are Based on Improved Gasoline Desulfurization Technologies)
5
4
"Ł"
o
a 3
7/5
0)
S 2
*j
(A
O
O
1
0
(
^
•^
l^_
-^•^
' »
— •
•**-^.
— •
•»—
5 40 80 120 160 200 240 280 320 360 400 440 480 520
Sulfur Concentration (ppm)
Figure V-3. Cost of Reducing Gasoline Sulfur in PADD 2
(Costs are Based on Improved Gasoline Desulfurization Technologies)
V-67
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Cost (cents/gallon)
O -» NJ W 4^ Ol
^
•.
\
1
*--,
X
1^.
••^
X
^»
\
-^-.
^=
=5*6
**^
• Improved Tech
*• OCTGAIN125
3 40 80 120 160 200 240 280 320 360 400 440 480 520
Sulfur Concentration (ppm)
Figure V-4. Cost of Reducing Gasoline Sulfur in PADD 3
(Costs are Based on Improved Gasoline Desulfurization Technologies and Octgain 125)
ts/gallon)
W *>. Ol
0)
Ł 2
•4-1
(A
O
O
1
0
(
T
1
k
~i
V
"V
w^
\
\
x
) 40 80 120 160 200 240 280 320 360 400 440 480 520
Sulfur Concentration (ppm)
Figure V-5. Cost of Reducing Gasoline Sulfur in PADD 4
(Costs are Based on Improved Gasoline Desulfurization Technologies)
V-68
-------
Chapter V: Economic Impact
Cost (cents/gallon)
O -» NJ W 4^ Ol
\
^
*.
"••^
^
^^
^^
^^
"
^^
^ —
"^^
^^^
**^* _
) 40 80 120 160 200 240 280 320 360 400 440 480 520
Sulfur Concentration (ppm)
Figure V-6. Cost of Reducing Gasoline Sulfur in PADD 5 Outside of California
(Costs are Based on Improved Gasoline Desulfurization Technologies)
c
_o
"<5 ,
O) 3
|
0)
^ 2
to
o
o
40 80 120 160 200 240 280 320 360 400
Sulfur Concentration (ppm)
440 480
520
Figure V-7. National Cost of Reducing Gasoline Sulfur Outside of California
(Costs are Based on Improved Gasoline Desulfurization Technologies)
V-69
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
4. Per Vehicle Life-Cycle Fuel Costs
The additional cost of low sulfur gasoline is encountered by the average vehicle owner
each time the fuel tank is refilled. The impacts of the gasoline sulfur standard on the average
vehicle owner can therefore be calculated as the increased fuel production costs in cents per
gallon, multiplied by the total number of gallons used by a vehicle over a particular timeframe.
Thus we have calculated the in-use impact of our proposed gasoline sulfur standard on a per-
vehicle basis for both a single year and for an entire vehicle's lifetime.
To estimate the cost of low sulfur gasoline in one year for a single vehicle, it is necessary
to convert the annual miles traveled by a single vehicle into gallons of gasoline consumed. This
conversion requires the use of an average fuel economy factor. Although the current fleet-
average fuel economy is approximately 20.7 miles per gallon31, this value is expected to change
in the future for two reasons:
1) As the fleet turns over, those vehicles that were certified at lower fuel economy
levels drop out of the in-use fleet.
2) The light-duty vehicle fraction of the fleet is projected to drop as more and more
light-duty trucks come into the market.
We have projected that the light-duty vehicle portion of the fleet will level off to a fuel economy
of about 24.2 miles per gallon during the next decade, while the light-duty truck portion of the
fleet will level off to about 15.5 miles per gallon in the same timeframe. Using the projected
long-term distribution of 40 percent LDV and 60 percent LDT in the fleet32, we calculated the
fleet-average fuel economy to be 19.0 miles per gallon.
In a single year, the average in-use light-duty vehicle travels approximately 11,500
milesbb. Applying the average fuel economy factor of 19.0 miles per gallon and the initial cost
for low sulfur fuel of 1.68 eVgal leads us to a per-vehicle estimate of $10.17. This is the
additional cost that the average vehicle owner will incur in the first year of the program due to
the use of low sulfur gasoline.
The per-vehicle cost of low sulfur gasoline can also be calculated over the lifetime of a
vehicle. However, to calculate a lifetime cost for the average in-use vehicle, it is necessary to
account for the fact that individual vehicles experience different lifetimes in terms of years that
they remain operational. This distribution of lifetimes is the vehicle survival rate distribution,
for which we used data from the National Highway Transportation Safely Administration. The
costs of low sulfur gasoline incurred over the lifetime of the average fleet vehicle can then be
calculated as the sum of the costs in individual years as shown in the equation below:
bb Calculated from the annual miles traveled per vehicle for each year of a vehicle's life, multiplied by a
distribution of vehicle survival rates by year. Annual miles travelled from "MOBILE6 Fleet Characterization Input
Data," Tracie R. Jackson, Report Number M6.FLT.007. Estimate of 11,500 miles per year includes both LDV and
LDT.
V-70
-------
Chapter V: Economic Impact
Where:
LFC
(AVMT);
(SURVIVE);
C
FE
= [(AVMT); • (SURVIVE); • (C) - (FE)]
= Lifetime fuel costs in $/vehicle
= Annual vehicle miles travelled in year i of a vehicle's operational life33
= Fraction of vehicles still operating after i years of service34
= Cost of low sulfur gasoline in $/gal
= Fuel economy in miles per gallon. 24.2 for LDV, 15.5 for LDT
= Vehicle years of operation, counting from 1 to 25
The cost of low sulfur gasoline is a function of the year of refinery production as described in
Section V.B.; the initial cost of 1.68 eVgal applies only in the first year of low sulfur gasoline
production. In subsequent years, refiners are able to make use of their experience in order to
lower their operating expenses. As a result of these declining fuel costs over time, we
determined that it is appropriate to calculate total lifetime costs for two separate cases:
1) Near-term, representing a vehicle whose operational life begins at the same time
that low sulfur gasoline production begins
2) Long-term, representing a vehicle whose operational life begins six years after
low sulfur gasoline production begins
The sixth year for calculating long-term costs of low sulfur gasoline was chosen to be consistent
with the sixth year of vehicle manufacture, when the capital cost amortization period ends.
Details of the calculation of long-term vehicle costs are given in Section V.A.
We used the above equation to calculate lifetime fuel costs separately for LDV, LDT1,
LDT2, LDT3, and LDT4. The results are shown in Table V-44.
Table V-44. Undiscounted Per-vehicle Costs of
Low Sulfur Gasoline (In 1997 Dollars)
LDV
LDT1, LDT2
LDT3, LDT4
Near-term ($)
83.36
178.52
192.13
Long-term ($)
78.15
167.56
180.28
We then weighted the per-vehicle costs for the individual vehicle categories in Table V-44 by
the fleet fractions. As a result, the total cost incurred by the average in-use vehicle over its
lifetime due to the use of low sulfur gasoline was calculated to be $142.53 on a near-term basis
and $133.73 on a long-term basis.
V-71
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999 _
An alternative approach to calculating lifetime per-vehicle costs of low sulfur gasoline is
to discount future year costs. This approach leads to "net present value" lifetime fuel costs, and
is a useful means for showing what the average vehicle owner would have to spend in the first
year in order to pay for all future year fuel costs. It also provides a means for comparing the
program's costs to its emission reductions in a cost-effectiveness analysis, as described in
Section VI.
Discounted lifetime fuel costs are calculated in an analogous manner to the undiscounted
values, except that each year of the summation is discounted at the average rate of 7%. The
equation given above can be modified to include this annual discount factor:
[{(AVMT); • (SURVIVE); • (C) - (FE)}/(1.07)M]
Once again, we calculated lifetime fuel costs separately for LDV, LDT1, LDT2, LDT3, and
LDT4. These values are shown in Table V-45.
Table V-45. Discounted Per-vehicle Costs of
Low Sulfur Gasoline (In 1997 Dollars)
LDV
LDT1, LDT2
LDT3, LDT4
Near-term ($)
60.98
126.95
135.85
Long-term ($)
56.73
118.19
126.43
Once again, we then weighted the per-vehicle costs for the individual vehicle categories in Table
V-45 by the fleet fractions. As a result, the total discounted cost incurred by the average in-use
vehicle over its lifetime due to the use of low sulfur gasoline was calculated to be $101.92 on a
near-term basis and $94.86 on a long-term basis.
A summary of all per-vehicle fuel costs described in this section is given in Table V-46
below.
V-72
-------
Chapter V: Economic Impact
Table V-46. Fleet Average Per-vehicle Costs
Of Low Sulfur Gasoline (In 1997 Dollars)
First year
Lifetime, undiscounted, near-term
Lifetime, undiscounted, long-term
Lifetime, discounted, near-term
Lifetime, discounted, long-term
Cost per vehicle
($)
10.17
142.53
133.73
101.92
94.86
5. Aggregate Annual Fuel Costs
Aggregate fuel costs are those costs associated with the increased cost per gallon of
gasoline due to the proposed sulfur controls, multiplied by the total number of gallons of
gasoline consumed in any given year by both highway and non-road sources. The total gallons
of gasoline consumed by highway sources were calculated using the VMT projections used
throughout the analyses within this document, along with projected fuel economy estimates
(mpg) developed by Standard & Poor's Data Research International (DRI).35 The resultant
aggregate annual fuel costs are summarized in Table V-47. It is important to note that the capital
costs associated with the proposed sulfur controls have been amortized for this analysis. The
actual capital investment would occur up-front, prior to and during the initial years of the
program, as described previously in this chapter.
V-73
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-47. Increased Annualized Fuel Cost as a Result of Today's Proposed Tier 2
Gasoline Sulfur Controls
(SMillion)
Calendar
Year
2000
2004
2010
2015
2020
Including Non-Road and
Excluding California00
0
2,255
2,127
2,156
2,270
a. Methodology
The DRI develops projected fuel economy estimates for passenger cars (EPA's LDVs),
light trucks under 10,000 pounds, and heavy trucks over 10,000 pounds. The VMT projections
developed by EPA are for light-duty vehicles (LDV), light-duty trucks (LDT - under 8500
pounds), and heavy-duty gasoline (over 8500 pounds). Because of the inconsistency in
stratifying the fleet, the DRI fuel economy estimates for light trucks (under 10,000 pounds) were
used for both the EPA LDT (under 8500 pound) and for EPA's heavy-duty gasoline trucks from
8500 to 10,000 pounds. The DRI fuel economy estimates for over 10,000 pound trucks were
then used for EPA's over 10,000 pound heavy-duty gasoline trucks.
The stratification of EPA VMT projections between the 8500 to 10,000 trucks and the
over 10,000 trucks was done by using both DRI and EPA data. The DRI projections for the
2000 calendar year show that of all gasoline trucks, light and heavy, 2.1 percent are in the over
10,000 pound category. EPA projections show that of all gasoline trucks, light-duty and heavy-
duty, 9.1 percent are in the over 8500 pound category. Using these two projected population
percentages, the heavy-duty VMT projections were allocated 77 percent to the 8500 to 10,000
category, and 23 to the over 10,000 category. The same calculation was carried out and used for
each calendar year from 2000 to 2020, when the split is projected at 86 percent and 14 percent,
respectively. These results are shown in Table V-48.
The DRI fuel economy estimates also include both gasoline and diesel vehicles and
trucks. As a result, the truck fuel economy estimates may be slightly higher than a gasoline-only
estimate, as diesel vehicles and trucks tend to have higher fuel economy numbers than do
gasoline vehicles and trucks. There should be little effect on the fuel economy estimates for
ccThe aggregate fuel costs used in the economic impact analysis of today's proposal include gasoline
consumed by non-road sources and exclude gasoline consumed in the State of California.
V-74
-------
Chapter V: Economic Impact
passenger cars, because DRI estimates that 99.7 percent of passenger cars will be gasoline fueled
in the 2000 calendar year (although 96.5 percent in the 2020 calendar year). Even for light
trucks under 10,000 pounds, where more diesels would be expected, DRI estimates a split of 96
percent gasoline in the 2000 calendar year and 92.8 percent in the 2020 calendar year.
Therefore, the effect of diesel vehicles and trucks on the DRI under 10,000 pound fuel economy
estimates is considered negligible due to their low populations.
The effect of diesels on the over 10,000 pound heavy truck fuel economy estimates is
also considered negligible, at least where the total gasoline consumption is concerned. Although
the diesel population is relatively high in this category, where DRI estimates diesels at roughly
68 percent of the over 10,000 trucks, their effect is considered negligible because of the
insignificant amount of gasoline consumed by trucks over 10,000 pounds (less than 0.02 percent)
relative to the gasoline consumed by vehicles and trucks under 10,000 pounds.
The projected VMT values within each category (LDV, LOT, HDCX 10,000, and
HDG> 10,000) were then divided by the corresponding DRI projected fuel economy estimates to
derive the gasoline consumption for each category per year. These values were then added, in
each given year, to derive the total highway gasoline consumption for each year from 2004 to
2020. The results are shown in Table V-49.
b. Explanation of Results
The aggregate fuel costs used in the economic impact analysis of today's proposal
include the non-road contribution but exclude gasoline consumed within the State of California.
The total nationwide highway gasoline consumption was adjusted by eliminating 11 percent to
exclude the California contribution.dd The non-road contribution to the gasoline consumption
was then added in by multiplying the highway contribution by 6.4 percent, as non-road sources
are estimated to use 6.4 percent of the amount consumed by highway sources.36 The highway
gasoline consumption, including the non-road contribution and excluding the California
contribution, was then multiplied by the per gallon increase due to the proposed sulfur control
requirements to arrive at the estimated aggregate fuel cost for each individual year. The results
are shown in Table V-50.
The aggregate fuel costs used in the economic impact analysis of today's proposal
include non-road sources because gasoline used to power these sources will incur the increased
per gallon cost, but exclude California because today's proposal will not impact the cost of
gasoline in the State of California. The aggregate fuel costs used in the economic impact
analysis include Alaska and Hawaii as gasoline in those states will incur the increased per gallon
cost.
The aggregate fuel costs decrease during the early years due to the decreasing per gallon
ddBased on EPA VMT estimates that California accounts for approximately 11 percent of nationwide
VMT.
V-75
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
cost associated with improved refining techniques and the pay off of amortized capital costs.
The aggregate costs then increase in later years due both to the reinvestment in refinery
equipment (increased capital costs), which increases the per gallon cost, and because VMT is
projected to increase every year, which results in increasing fuel consumption.
V-76
-------
Chapter V: Economic Impact
Table V-48. Stratification of Heavy-Duty Gasoline Fleet using Vehicle Count Projec
(Counts are in Millions of Vehicles)
CY
1 997/6
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
S&P DRI
72.03
83.74
87.42
89.94
92.46
94.98
99.85
102.37
104.89
107.41
109.93
113.32
114.79
116.27
117.74
119.21
122.67
124.14
125.62
127.09
128.56
128.39
S&P
DRI
1.88
1.76
1.74
1.73
1.71
1.69
1.70
1.68
1.67
1.65
1.63
1.68
1.68
1.68
1.67
1.67
1.67
1.67
1.67
1.66
1.66
1.66
S&P DRI
total
truck
73.91
85.50
89.16
91.67
94.17
96.68
101.55
104.05
106.56
109.06
111.57
115.00
116.47
117.94
119.41
120.89
124.34
125.81
127.28
128.75
130.23
130.05
S&P DRI
%>10k
2.1%
2.0%
1.9%
1.8%
1.8%
1.7%
1.6%
1.6%
1.5%
1.5%
1.5%
1.4%
1.4%
1.4%
1.4%
1.3%
1.3%
1.3%
1.3%
1.3%
1.3%
AMD<6k
(2)
54.91
56.91
58.94
61.00
63.09
64.97
66.62
68.29
69.98
71.68
73.24
74.24
75.26
76.27
77.30
78.23
78.83
79.44
80.05
80.66
81.28
AMD 6k-
8500 (2)
19.52
20.23
20.95
21.68
22.43
23.09
23.68
24.28
24.87
25.48
26.03
26.39
26.75
27.11
27.48
27.81
28.02
28.24
28.46
28.67
28.89
AMD<8500
74.43
77.14
79.89
82.68
85.52
88.06
90.30
92.57
94.85
97.16
99.27
100.63
102.01
103.38
104.78
106.04
106.85
107.68
108.51
109.33
110.17
AMD 8500-
14k (2)
5.22
5.41
5.60
5.81
6.02
6.15
6.28
6.42
6.55
6.70
6.78
6.87
6.96
7.04
7.13
7.19
7.24
7.29
7.35
7.40
7.45
AMD>14k
(2)
2.26
2.34
2.43
2.52
2.61
2.67
2.72
2.78
2.84
2.90
2.94
2.98
3.02
3.05
3.09
3.12
3.14
3.16
3.18
3.21
3.23
AMD
Total
HDG
7.48
7.75
8.03
8.33
8.63
8.82
9.00
9.20
9.39
9.60
9.72
9.85
9.98
10.09
10.22
10.31
10.38
10.45
10.53
10.61
10.68
AMD Total
Truck
81.91
84.89
87.92
91.01
94.15
96.88
99.30
101.77
104.24
106.76
108.99
110.48
111.99
113.47
115.00
116.35
117.23
118.13
119.04
119.94
120.85
AMD
%>8500
9.1%
9.1%
9.1%
9.2%
9.2%
9.1%
9.1%
9.0%
9.0%
9.0%
8.9%
8.9%
8.9%
8.9%
8.9%
8.9%
8.9%
8.8%
8.8%
8.8%
8.8%
%HDG
8500-1 OK
77.5%
78.6%
79.4%
80.2%
80.9%
81.6%
82.2%
82.7%
83.2%
83.7%
83.6%
83.8%
84.0%
84.2%
84.4%
84.8%
85.0%
85.2%
85.4%
85.6%
85.6%
%HDG>10k
22.5%
21.4%
20.6%
19.8%
19.1%
18.4%
17.8%
17.3%
16.8%
16.3%
16.4%
16.2%
16.0%
15.8%
15.6%
15.2%
15.0%
14.8%
14.6%
14.4%
14.4%
(1) From S&P DRI World Energy Service U.S. Outlook, Table 17, April 1998; see memo fr. T.Sherwood to Docket A-97-10, 3/22/99
(2) Draft MOBILE6 Fleet Characterization Input Data, OMS/AMD/Jackson, August 1998; uses count projections where 99.2% of LDTs are gasoline & 0.8% are diesel in both the 2000CY & the 2020CY;
see memo fr. T.Sherwood to Docket A-97-10, 3/22/99
V-77
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-49. Calculation of Gasoline Consumption
AMD <8500
<8500 VMT
VMT ex nation
CA.AL.HI Bmiles %Truck
CY Bmiles (1) (2) %Car(1) (1)
1997
2000 2160 2455 54.7% 45.3%
2001 2210 2511 53.1% 46.9%
2002 2250 2557 51.5% 48.5%
2003 2290 2602 49.9% 50.1%
2004 2330 2648 48.4% 51.6%
2005 2380 2705 46.8% 53.2%
2006 2420 2750 45.2% 54.8%
2007 2460 2795 43.6% 56.4%
2008 2510 2852 42.0% 58.0%
2009 2550 2898 40.6% 59.4%
2010 2600 2955 39.4% 60.6%
2011 2650 3011 38.3% 61.7%
2012 2710 3080 37.3% 62.7%
2013 2770 3148 36.5% 63.5%
2014 2882 3275 35.8% 64.2%
2015 2888 3282 35.2% 64.8%
2016 2940 3341 34.7% 65.3%
2017 3000 3409 34.2% 65.8%
2018 3060 3477 33.9% 66.1%
2019 3130 3557 33.6% 66.4%
2020 3190 3625 33.4% 66.6%
PassCar
AMD PassCar
PassCar Gasoline
VMT S&P DRI Consump
nation PassCar nation
Bmiles (3) mpg (4) Bgal
1342 21.2 63.30
1333 21.3 62.59
1318 21.4 61.43
1300 21.6 60.19
1281 21.7 58.91
1266 21.9 57.81
1244 22.0 56.41
1220 22.2 54.95
1199 22.3 53.67
1177 22.5 52.35
1164 23.2 50.15
1153 23.4 49.19
1149 23.7 48.58
1149 23.9 48.07
1172 24.1 48.57
1155 24.6 46.93
1158 24.8 46.64
1167 25.1 46.57
1178 25.3 46.59
1195 25.5 46.82
1211 25.5 47.47
LDK8500
S&P
LOT DRI LOT <8500
AMD LOT Gasoline <10k Gasoline
VMT nation VMT nation Truck Consump
Bmiles (5) Bmiles (6) mpg (4) nation Bgal
1112 1104 15.9 69.41
1178 1169 16.0 73.05
1239 1229 16.1 76.17
1302 1292 16.3 79.38
1367 1356 16.4 82.60
1438 1427 16.5 86.48
1506 1494 16.6 89.81
1576 1563 16.8 93.17
1653 1640 16.9 96.95
1720 1707 17.1 100.07
1791 1777 17.3 102.70
1859 1844 17.5 105.63
1930 1915 17.6 108.71
1999 1983 17.8 111.59
2103 2086 17.9 116.35
2127 2110 18.2 115.95
2183 2165 18.4 117.91
2242 2224 18.5 120.02
2299 2280 18.7 121.98
2362 2343 18.9 124.19
2414 2395 19.0 126.06
HDG 8500- 10k
S&P 8500-1 Ok
AMD HDG 8500-10k DRI Gasoline
VMT ex HDG VMT HDG VMT <10k Consump
CA.AL.HI nation nation Truck nation
Bmiles (1) Bmiles (2) Bmiles (7) mpg (4) Bgal
0.518 0.589 0.456 15.9 0.03
0.533 0.606 0.476 16.0 0.03
0.548 0.623 0.494 16.1 0.03
0.563 0.640 0.513 16.3 0.03
0.578 0.657 0.531 16.4 0.03
0.593 0.674 0.550 16.5 0.03
0.610 0.693 0.569 16.6 0.03
0.627 0.713 0.589 16.8 0.04
0.645 0.733 0.610 16.9 0.04
0.662 0.752 0.630 17.1 0.04
0.679 0.772 0.645 17.3 0.04
0.690 0.784 0.657 17.5 0.04
0.705 0.801 0.673 17.6 0.04
0.721 0.819 0.690 17.8 0.04
0.736 0.836 0.706 17.9 0.04
0.752 0.855 0.725 18.2 0.04
0.767 0.872 0.741 18.4 0.04
0.783 0.890 0.758 18.5 0.04
0.798 0.907 0.774 18.7 0.04
0.814 0.925 0.791 18.9 0.04
0.829 0.942 0.806 19.0 0.04
HDG>10k
>10k
S&P Gasoline
>10kHDG DRI Consump
VMT nation >10k nation
Bmiles (7) mpg (4) Bgal
0.133 7.1 0.02
0.130 7.1 0.02
0.128 7.2 0.02
0.127 7.2 0.02
0.125 7.2 0.02
0.124 7.4 0.02
0.124 7.4 0.02
0.123 7.5 0.02
0.123 7.5 0.02
0.122 7.5 0.02
0.126 7.5 0.02
0.127 7.5 0.02
0.128 7.5 0.02
0.129 7.6 0.02
0.130 7.6 0.02
0.130 7.6 0.02
0.131 7.6 0.02
0.132 7.6 0.02
0.133 7.7 0.02
0.134 7.7 0.02
0.136 7.8 0.02
Totals
EPA Total S&PDRI Hwy
Hwy Gasoline
Gasoline Consump
Consump nation Bgal
nation Bgal (4), (8)
120.94
132.76 132.72
135.69 134.90
137.65 137.07
139.62 139.25
141.56 141.43
144.34 142.44
146.27 144.62
148.18 146.79
150.67 148.97
152.47 151.15
152.91 151.56
154.88 152.47
157.34 153.38
159.71 154.29
164.98 155.20
162.94 157.48
164.61 158.39
166.65 159.30
168.62 160.21
171.07 161.12
173.59 161.66
(1) OMS/AMD/Koupal; %Car & %Truck represent % of Light-duty VMT
(2) CA = 11% of nation; CA,AK,HI= 12% of nation
(3) Multiplies <8500 VMT nation by %Car
(4) From S&P DRI World Energy Service U.S. Outlook, Table 17 (mpg values include diesel), April 1998; see memo fr. T.Sherwood to Docket A-97-10, 3/22/99
(5) Multiplies <8500 VMT nation by %Truck
(6) Draft MOBILE6 Fleet Characterization Input Data, OMS/AMD/Jackson, August 1998; uses count projections where 99.2% of LDTs are gasoline & 0.8% are diesel in both the 2000CY & the 2020CY;
see memo fr. T.Sherwood to Docket A-97-10, 3/22/99
(7) Uses S&P DRI data for % of all gas trucks >10k, and AMD data for % of all gas trucks >8500, then calculates % of all >8500 gas trucks in the 8500-1 Ok category, and % of all >8500 gas trucks in the >10k category.
(8) Presented for comparison only. Discrepancy in later years due mainly to AMD's larger LOT VMT share (67% of LD VMT) vs S&P (-53% of <10k VMT)
V-78
-------
Chapter V: Economic Impact
Table V-50. Aggregate Annualized Fuel
CY
1997
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
EPA Total
Hwy
Gasoline
Consump
nation Bgal
132.76
135.69
137.65
139.62
141.56
144.34
146.27
148.18
150.67
152.47
152.91
154.88
157.34
159.71
164.98
162.94
164.61
166.65
168.62
171.07
173.59
Total Hwy
Gasoline
Consumption
excluding CA
Bgal(1)
118.16
120.76
122.51
124.26
125.99
128.47
130.18
131.88
134.09
135.70
136.09
137.84
140.03
142.14
146.83
145.01
146.50
148.32
150.07
152.25
154.50
Costs per Year from 2004 to 2020
Non-road
Gasoline
Consumption
excluding CA
Bgal (2)
7.56
7.73
7.84
7.95
8.06
8.22
8.33
8.44
8.58
8.68
8.71
8.82
8.96
9.10
9.40
9.28
9.38
9.49
9.60
9.74
9.89
Tier2 Cost ex
CA & incl
NonRoad $B
(3)
0
0
0
0
2.255
2.276
2.136
2.138
2.147
2.147
2.127
2.129
2.142
2.154
2.204
2.156
2.158
2.166
2.172
2.264
2.270
(1) CA = 11% of total nation; CA.AK.HI = 12% of nation
(2) OMS/VPCD/Todd Sherwood; NonRoad fraction = 6.4%;
see memo to Docket A-97-10, 2/19/99
(3) OMS/FED/Wyborny; Tier2 $/gal increase
CY
2004
2005
2006
2007
2008
2009
2010
2011
2012
Adj Cost
$/gal
0.01682
0.01665
0.01542
0.01523
0.01505
0.01487
0.01469
0.01452
0.01438
CY
2013
2014
2015
2016
2017
2018
2019
2020
Adj Cost
$/gal
0.01424
0.01411
0.01397
0.01385
0.01372
0.01360
0.01398
0.01381
V-79
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
C.
Combined Vehicle and Fuel Costs
Sections A. and B. of this section provide detailed cost analyses for Tier 2 vehicles and
low sulfur gasoline, respectively. The following sums the costs to consumers to provide total
incremental costs of the Tier 2 program. The per vehicle costs are provided first, followed by
the total annual nationwide costs.
1.
Combined Costs Per Vehicle
Table V-51 provides a summation of our estimated incremental per vehicle costs,
including increased costs for Tier 2 vehicles and for low sulfur gasoline over the life of the
vehicles. We use the cost estimates for our cost-effectiveness analysis presented in the following
Chapter. As described in the previous sections, we expect these costs to decrease over time as
manufacturers make production improvements and recover fixed costs. Table V-51 provides
estimates of near-term costs, which represent costs in the first years of the program, and long-
term costs which account for the cost decreases.
Table V-51. Total Incremental Per Vehicle Costs to Consumers
Over the Life of a Tier 2 Vehicle
LDV
($)
LDT1
($)
LDT2
($)
LDT3
($)
LDT4
($)
Near-term Costs
Vehicle costs
Fuel costs*
Total
80
61
141
73
127
200
136
127
263
274
136
410
270
136
406
Long-term Costs
Vehicle costs
Fuel costs*
Total
50
57
107
47
118
165
103
118
221
218
126
344
213
126
339
* Discounted lifetime fuel costs in 1997 dollars
2.
Combined Total Annual Nationwide Costs
Figure V-8 and Table V-52 summarize EPA's estimates of total annual costs to the nation
V-80
-------
Chapter V: Economic Impact
both for Tier 2 vehicles and low sulfur gasoline.ee The capital costs have been amortized for
these analyses. The actual capital investment would occur up-front, prior to and during the
initial years of the program, as described previously in this chapter. The fuel costs shown are for
all gasoline consumed nationwide, including both on-highway and nonroad. Annual aggregate
vehicle costs change as Tier 2 vehicle sales are phased-in and projected per-vehicle costs and
annual sales change over time. The aggregate fuel costs change as projected per gallon costs and
annual fuel consumption change over time. Increases in fuel consumption over time are
generally off-set by decreases in per gallon costs. The methodology we used to derive the
aggregate costs are described in detail in the sections A.3. and B.5. of this chapter. As shown
below, total annual costs increase over the phase-in period and peak at about $3.7 billion.
Annual costs then drop to about $3.5 billion, largely due to decreases in vehicle costs. Costs
increase gradually after 2012 due to the stabilization of vehicle costs in the long-term and
projected increases in vehicle sales and fuel consumption.
Total Annualized Costs of Tier 2 Vehicles and Low Sulfur Ga
4
(0
JO
"o
Q
"5
01
o
CO 1
Total
Fuels
—•—
Vehicles
2004 2006 2008 2010 2012 2014 2016 2018 2020
Calendar Year
Note: Capital costs have been amortized for purposes of this analysis
Figure V-8. Total Annualized Costs of Tier 2 Vehicles and Low Sulfur Gasoline.
ee Excluding vehicles and fuel sold in California.
V-81
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table V-52. Total Annualized Costs to the Nation for
Tier 2 Vehicles and Low Sulfur Gasoline
(Smillion)
Calendar Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
Vehicle Costs ($)
$257
$506
$815
$1,365
$1,589
$1,587
$1,496
$1,427
$1,359
$1,348
$1,346
$1,352
$1,359
$1,366
$1,373
$1,380
$1,387
Fuel Costs ($)
$2,255
$2,276
$2,136
$2,138
$2,147
$2,147
$2,127
$2,129
$2,142
$2,154
$2,204
$2,156
$2,158
$2,166
$2,172
$2,264
$2,270
Total ($)
$2,512
$2,782
$2,951
$3,503
$3,736
$3,734
$3,623
$3,556
$3,501
$3,502
$3,550
$3,508
$3,517
$3,532
$3,545
$3,644
$3,657
V-82
-------
Chapter V: Economic Impact
Chapter V. References
1. "Notice of Public Meeting to Consider the Status of Implementation of the Low Emission
Vehicle Program", California Air Resources Board, Mail-out #96-28, November 21,
1996.
2. "Proposed Amendments to California Exhaust and Evaporative Emission Standards and
Test Procedures for Passenger Cars, Light-duty Trucks, and Medium-duty Vehicles",
Staff Report: Initial Statement of Reasons, State of California Air Resources Board,
September 18, 1998.
3. Report Submitted for WA 2-9, Evaluation of the Costs and Capabilities of Vehicle
Evaporative Emission Control Technologies. ICF Consulting Group, March 22, 1999.
4. Cost Estimations for Emission Control-Related Components/Systems and Cost
Methodology Description," Leroy H. Lindgren, Rath and Strong for U.S. EPA, EPA-
460/3-78-002, March 1978.
5. "Cost Analysis, Compliance Assurance Program for Light-duty Vehicles and Light-duty
Trucks", March 1999, Docket A-96-50.
6. "Learning Curves in Manufacturing," Linda Argote and Dennis Epple, Science, February
23, 1990, Vol. 247, pp. 920-924.
7. J.M Dutton and A. Thomas, Academy of Management Review., Rev. 9, 235, 1984.
8. Cost Estimations for Emission Control-Related Components/Systems and Cost
Methodology Description," Leroy H. Lindgren, Rath and Strong for U.S. EPA, EPA-
460/3-78-002, March 1978.
9. "Update of EPA's Motor Vehicle Emission Control Equipment Retail Price Equivalent
(RPE) Calculation Formula," Jack Faucett Associates for U.S. EPA, Report No.
JACKFAU-85-322-3. September 4, 1985.
10. "Motor Vehicle Facts and Figures", American Automobile Manufacturers, 1997.
11. "Light-duty Truck Reference Guide", J.D. Power and Associates, July 1998.
12. "Annual Energy Outlook 1999 with Projections to 2020", Energy Information
Administration, Office of Integrated Analysis and Forecasting, U.S., Department of
Energy, DOE/EIA - 0383(99), p. 137, December 1999.
13. Gary, James H., Handwerk, Glenn E., Petroleum Refining: Technology and Economics,
Marcel Dekker, New York (1994).
14. Personal conversation with A. Judzis of Equilon.
V-83
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
15. Gary, James H., Handwerk, Glenn E., (1994).
16. Gary, James H., Handwerk, Glenn E., (1994).
17. Gary, James H., Handwerk, Glenn E., (1994).
18. U.S. Petroleum Refining, Meeting Requirements for Cleaner Fuels and Refineries,
Volume V-Refming Capability Appendix, National Petroleum Council, 1993.
19. Personal conversation with A. Judzis of Equilon.
20. Personal conversations with Bob King of Sun Company Inc, and A. Judzis of Equilon.
21. U.S. Petroleum Refining, Meeting Requirements for Cleaner Fuels and Refineries,
Volume V-Refming Capability Appendix, National Petroleum Council, 1993.
22. Personal conversation with Bob King of Sun Company Inc.
23. Presentation by Air Products and Chemicals, Hart's/IRI World Fuels Conference, San
Francisco, California, March 1998.
24. Petroleum 1996: Issues and Trends, Energy Information Administration, 1996.
25. Judzis, A, E-mail message to Wyborny, Lester, entitled: Copies of two slides, 6/26/1998.
26. Venner, S. F., Downstream Mergers and Acquisitions: Is There Really a Pot of Gold at
the End of the Rainbow?, NPRA Annual Meeting, San Francisco, CA, March 1998.
27. Stetzer, C. Martin, Price Waterhouse L.L.C., Redefining the Context of Refinery
Pacesetter Performance, 1997 NPRA Annual Meeting, March 16-18, 1997.
28. Venner, S. F., 1997.
29. Refining Industry Profile Study, 1998.
30. Wang, M.Q., GREET 1.4 - Transportation Fuel-Cycle Model, Center for Transportation
Research, Argonne National Laboratory.
31. U.S. DOT/ FHA, Highway Statistics for 1995. Total rural and urban miles traveled
divided by total gallons of gasoline consumed in transportation.
32. Technical memorandum from John Koupal to EPA Air Docket A-97-10, "Methodology
for Developing Light-Duty Emission Inventory Estimates in the Tier 2 NPRM," EPA
Report No. EPA420-R-99-005.
33. "MOBILE6 Fleet Characterization Input Data," Tracie R. Jackson, Report Number
M6.FLT.007.
V-84
-------
Chapter V: Economic Impact
34. Technical memorandum from John Koupal to EPA Air Docket A-97-10, "Methodology
for Developing Light-Duty Emission Inventory Estimates in the Tier 2 NPRM," EPA
Report No. EPA420-R-99-005.
35. Standard & Poor's DRI World Energy Service U.S. Outlook, Table 17, April 1998.
36. Memo from Todd Sherwood to Air Docket A-97-10, Non-Road Gasoline Consumption,
dated February 19,1999.
V-85
-------
Chapter VI: Cost Effectiveness
Chapter VI: Cost-Effectiveness
This section will present the cost-effectiveness analysis we completed for the combined
Tier 2 exhaust, Tier 2 evaporative, and gasoline sulfur standards. This analysis relies in part on
cost information from Section V and emissions information from Section III to estimate the cost-
effectiveness of the standards in terms of dollars per ton of total NOx + NMHC emission
reductions. Finally, this Section compares the cost-effectiveness of the new provisions with the
cost-effectiveness of other NOx and NMHC control strategies from previous and potential future
EPA emission control programs.
A. Overview of the Analysis
The cost-effectiveness analysis conducted for our proposed standards focused on the
costs and emission reductions associated with a single vehicle meeting the Tier 2 emission
standards, and operating on low sulfur fuel. Both costs and emission reductions were calculated
over the life of the vehicle and then discounted at a rate of seven percent. Costs and emission
reductions were measured relative to an NLEV baseline and average sulfur levels in the absence
of sulfur controls. The calculations were performed separately for each vehicle class and the
results weighted according to the expected fleet mix. Details on our approach to cost-
effectiveness follow.
1. Temporal and Geographic Applicability
We have taken a per-vehicle approach to our cost-effectiveness calculations that
produces $/ton values representing any controlled vehicle, no matter where that vehicle operates.
In effect, this means that emission reductions in both attainment and nonattainment areas are
included in our cost-effectiveness analysis. We believe that this is appropriate. Both the Tier 2
vehicle and gasoline sulfur programs are proposed to apply nationwide, so that the same
emission reductions will occur regardless of where the vehicle operates. Attainment area
emission reductions also produce health benefits. In general, the benefits of NMHC reductions
in ozone attainment areas include reductions in emissions of air toxics, reductions in the
contribution from NMHC emissions to the formation of fine particulate matter, and reductions in
damage to agricultural crops, forests, and ecosystems from ozone exposure. Emission reductions
in attainment areas help to maintain clean air as the economy grows and new pollution sources
come into existence. Also, ozone health benefits can result from reductions in attainment areas,
although the most certain health effects from ozone exposure below the NAAQS appear to be
both transient and reversible. The closure letter from the Clean Air Science Advisory
Committee (CAS AC) for the recent review of the ozone NAAQS states that there is no apparent
threshold for biological responses to ozone exposure1.
In the Regulatory Impact Analysis for a recent rulemaking for highway heavy-duty diesel
engine standards2, EPA also presented a regional ozone control cost-effectiveness analysis in
VI-1
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
which the total life-cycle cost was divided by the discounted lifetime NOx + NMHC emission
reductions adjusted for the fraction of emissions that occur in the regions expected to impact
ozone levels in ozone nonattainment areas. (Air quality modeling indicates that these regions
include all of the states that border on the Mississippi River, all of the states east of the
Mississippi River, Texas, California, and any remaining ozone nonattainment areas west of the
Mississippi River not already included.) The results of that analysis show that the regional cost-
effectiveness values were 13 percent higher than the nationwide cost-effectiveness values.
Because of the small difference between the two results, EPA is presenting only nationwide cost-
effectiveness results for this analysis.
Despite the fact that a per-vehicle approach to cost-effectiveness allows us to avoid the
arbitrary choice of a specific year in which to conduct the analysis, there is some value in
examining different points in time after the program is first implemented. The costs of the
program will be higher immediately after it is implemented than they will be after several years,
since both vehicle manufacturers and refiners can take advantage of decreasing capital and
operating costs over time. For the purposes of this proposed rulemaking, therefore, we will
present cost-effectiveness of our proposed program on both a near-term and long-term basis.
More details concerning per-vehicle costs are given in Section VI.B.l.
2. Baselines
There are two broad approaches to cost-effectiveness that can be taken, each of which
requires a different baseline. These two approaches can be termed "incremental" and "average."
Both incremental and average approaches to cost-effectiveness provide a measure of how much
more stringent than the existing standards our proposed standards can be before they cease to be
cost-effective.
An incremental approach to cost-effectiveness requires that we evaluate a number of
different potential standards, each of which is compared to the potential standards closest to it.
Using this approach, the cost-effectiveness of our proposed standards would be calculated with
respect to another set of potential standards which is less stringent than our proposed standards.
In this way, the $/ton values represent the last increment of control, highlighting any
nonlinearities that exist in either the costs or emission reductions.
Incremental cost-effectiveness will produce different $/ton values than an average
approach to cost-effectiveness only if the costs or emission reductions are nonlinear. In the case
of our proposed standards, both the emission reductions and the fuel costs are nearly linear,
though the vehicle costs do contain some nonlinearity.
An average approach to cost-effectiveness, on the other hand, requires that we compare
the costs and emission reductions associated with our proposed standards to those for the
previous set of standards that are being met by manufacturers. In this case, the $/ton values
represent the full range of control from the last applicable standard to our proposed standards.
VI-2
-------
Chapter VI: Cost Effectiveness
Since today's proposed program includes both fuel standards and vehicle standards, it
was necessary for us to define a baseline for both fuels and vehicles. For fuels, there are no
previous controls applicable to sulfur (apart from an ASTM limit of 1000 ppm). As a result, we
have determined that the sulfur baseline should represent the national average sulfur level that
would exist at the time our proposed sulfur standard would go into effect. The national average
sulfur content of current conventional gasoline is approximately 330 ppm. We are not projecting
the sulfur level of conventional gasoline to change over the next ten years in the absence of
specific sulfur controls. For Phase II reformulated gasoline (RFG), the average sulfur content is
projected to be 150 ppm in the summer and 300 ppm in the wintei^. Based on seasonal volume
data, we estimate that 40 vol% of the annual pool is summer gasoline, with the remainder being
winter gasoline, producing an annual Phase II RFG sulfur level of 240 ppm. Because estimating
the number of areas that will continue to be in the RFG program by the middle of the next
decade is highly speculative, we have assumed that the current volume split between RFG and
conventional gasoline will continue indefinitely. Thus we estimated that Phase II RFG will
account for 26.7 percent of the total gasoline pool. As a result, we calculated the national
average sulfur level for the next decade to be 305 ppm. This is the baseline sulfur level used in
our calculations.
For the exhaust emission standards applicable to light-duty vehicles and trucks, there are
two potentially valid baselines that could be used. The Clean Air Act (CAA) suggests that Tier 2
vehicle standards should be compared to the previous set of federal light-duty standards, termed
Tier 1 standards. However, the language does not explicitly require that the cost-effectiveness
determination use Tier 1 standards as the baseline. Since the passage of the CAA Amendments
of 1990, the National Low Emission Vehicle (NLEV) program has gone into effect. NLEV
includes light-duty standards that are more stringent that Tier 1 for LDV, LDT1, and LDT2.
NLEV did not exist in 1990 and was not envisioned by the authors of the CAA Amendments of
1990. Had NLEV existed, either in concept or as a formal program, we believe that it could have
been identified in the CAA as the point of comparison for evaluating Tier 2 standards. In
addition, NLEV standards represent the most recent set of standards with which manufacturers
must comply. For our proposal, therefore, we have decided to make NLEV the baseline on
which the vehicle side of our cost-effectiveness calculations are based. Further, these NLEV
vehicles would be SFTP compliant since they would be sold in 2004 (the first year of our
proposed Tier 2 program).
The NLEV program did not include new standards for evaporative emissions, and so
cannot be used as the baseline for evaluating the cost-effectiveness of our proposed Tier 2
evaporative emission standards. Instead, the 2.0 gram/test standards under the enhanced
evaporative procedure, initially implemented in 1996, have been used as the baseline.
B. Costs
ff Based on a consensus opinion of the multi-party Phase II RFG Implementation Team, and summarized
in a report entitled, "Phase II RFG Report on Performance Testing." Contact: Deborah Wood, Office of Mobile
Sources.
VI-3
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
The costs used in our cost-effectiveness calculations are the sum of the costs of
compliance with the Tier 2 exhaust, Tier 2 evaporative, and gasoline sulfur standards on a per-
vehicle basis. Costs are given in 1997 dollars, and result from discounting over the lifetime of a
vehicle at a seven percent discount rate. In addition, all costs represent the fleet-weighted
average of light-duty vehicles and trucks.
1. Near and Long-Term Cost Accounting
Since the costs of complying with both the Tier 2 exhaust and gasoline sulfur standards
will vary over time, we determined that it is appropriate to consider both near-term and long-
term costs in our cost-effectiveness analysis. First, the capital costs associated with the
manufacture of vehicles that meet the proposed Tier 2 standards are generally amortized over
five years. Thus in the sixth year of production, a portion of the capital costs become zero and
the total costs of production drop. Manufacturers also gain knowledge about the best way to
meet new standards as time goes on, and as a result their operating costs decrease over time. As
described in a recent rulemaking setting standards for non-road compression ignition engines, we
have determined that the cost-implications of this "learning curve" can be estimated as a 20
percent drop in operating costs in the third year of production.
Thus near-term costs represent the highest costs of the program, as they include all
capital costs and no cost savings due to the manufacturer's learning curve. Long-term costs, on
the other hand, represent the lowest costs of the program which occur after a portion of capital
cost amortizations have ended and all learning curve cost savings have been accounted for. For
the purposes of this proposed rulemaking, therefore, we will present cost-effectiveness of our
proposed program on both a near-term and long-term basis.
Because of our per-vehicle approach to cost-effectiveness, near-term and long-term costs
are not associated with any specific year of our proposed Tier 2 program. For instance, the costs
associated with our proposed gasoline sulfur control program will decline steadily due to rotating
capital expenditures and continuous improvements in catalyst design. Vehicle costs, however,
decline over a different schedule. Not only are the vehicle-related capital costs amortized over
five years instead of the longer, rotating schedule for gasoline sulfur, but the phase-in schedule
for the Tier 2 exhaust standards varies depending on vehicle class. Therefore, the near-term
costs actually represent a conservative view of the costs of our proposed program, since they
consider the highest vehicle and fuel costs as if they occurred at the same time for all vehicle
classes. The long-term costs, on the other hand, represent the case for some later year of the Tier
2/gasoline sulfur program in which a majority of the fleet is meeting our proposed standards. In
this case, the phase-in schedule for light-duty vehicles and trucks is no longer evident in the fleet
mix, a portion of capital cost amortizations have ended, and most learning curve cost savings
will have been taken into account. Details about the calculation of near and long-term vehicle
and fuel costs can be found in Sections V.A.I and V.B.2.
2. Vehicle and Fuel Costs
VI-4
-------
Chapter VI: Cost Effectiveness
The per-vehicle costs used in our cost-effectiveness calculations were derived and
presented in the preceding sections. Vehicle costs were presented in Table V-12 for the five
vehicle categories affected by our proposed standards. For the purposes of calculating cost-
effectiveness, we first subtracted out the costs attributable to compliance with our proposed
evaporative emission standards, then weighted the remaining costs for those five individual
vehicle categories by the expected fleet fractions to obtain fleet-average costs for our proposed
exhaust emissions standards. Also, we treated first-year production costs as the "near-term"
costs, and sixth-year production costs as the "long-term" costs. Costs associated with
compliance with our proposed evaporative emission standards were constant across all vehicle
classes at $4.10 per vehicle. For low sulfur gasoline, we used the discounted lifetime costs
presented in Table V-41. The costs used in our cost-effectiveness calculations are repeated in
Table VI-1.
Table VI-1. Fleet-average, Per-vehicle Costs Used in Cost-effectiveness
Near-term
Long-term
Vehicle-exhaust
($)
124.04
89.47
Vehicle-evap
($)
4.10
4.10
Fuel
($)
101.92
94.86
Total costs
($)
230.06
188.43
Note that the total costs in Table VI-1 were used for establishing "uncredited" cost-effectiveness
values. As described in the next section, the costs from Table VI-1 were also adjusted to
produce "credited" cost-effectiveness values.
3. Cost Crediting for PM and SO2
The object of our cost-effectiveness analysis is to compare the costs to the emission
reductions in an effort to assess the program's efficiency in helping to attain and maintain the
NAAQS. Thus we recognize that the primary purpose of our proposed standards is to reduce
emissions of hydrocarbon and oxides of nitrogen emissions from the affected vehicles. That is
why we determined that cost-effectiveness should be calculated on the basis of total NOx +
NMHC emissions. However, we also believe that reductions in other pollutants which produce
health or welfare benefits should be included in the cost-effectiveness assessment, since they
also represent a value of our proposed program.
The reduction in gasoline sulfur levels that would result from our proposed standards will
necessarily result in reductions in sulfur-containing compounds that exit the tailpipe. These
compounds are limited to sulfur dioxide (SO2) and sulfate particulate matter. To account for
reductions in emissions of these two pollutants in our cost-effectiveness calculations, we have
calculated a second set of $/ton values in which we credit some of the costs to SO2 and direct
sulfate PM, with the remaining costs being used to calculate $/ton NOx+NMHC. As a result, we
VI-5
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
have produced both "credited" and "uncredited" $/ton NOx+NMHC values; the former takes into
account the SO2 and direct PM emission reductions associated with our proposed standards,
while the latter does not.
Cost-effectiveness values for the control of SO2 and direct PM represent conservative
estimates of the cost of measures that will need to be implemented in the future in order for all
areas to reach attainment. Such cost-effectiveness values are therefore an appropriate source for
estimating the amount of the costs to credit to these pollutants. As a result, we credited some
costs to SO2 and direct PM through the application of cost-effectiveness ($/ton) values for these
two pollutants withdrawn from other sources.
In concept, we would consider the most expensive program needed to reach attainment a
good representation of the ultimate value of PM or SO2 However, in this rulemaking, we chose
to simplify by using more conservative approaches to establish crediting values for PM and SO2.
The potential future programs evaluated as part of the NAAQS revisions rulemaking (discussed
in more detail in Section VI.D below) provided a reasonable source for identifying the value of
SO2 and direct PM in terms of their cost-effectiveness.
Out of the nine SO2 control programs evaluated in the NAAQS revisions rule, eight were
actually used in the modeling of ambient concentrations of PM based on their contribution to
secondary PM (sulfate) levels in PM nonattainment areas. The modeling showed that these eight
programs, along with other PM control programs as described above, permitted 70 percent of
counties not meeting the annual 8-hour PM standard to come into attainment. The
cost-effectiveness of the eight SO2 control programs ranged from $1600/ton to $111,500/ton. In
this particular rulemaking, rather than attempt to identify a more precise credit value for SO2
based on the last measures needed for attainment, we have for simplicity's sake used the average
cost effectiveness of the eight SO2 control programs, calculated to be $4800 a ton. This average
value of $4800/ton was used in the crediting of some costs to SO2, and represents a conservative
valuation of SO2.
The NAAQS revisions rule also evaluated PM control strategies, accounting for both
PM10 and PM2 5. The average cost-effectiveness for the PM control strategies considered in the
NAAQS revisions rule ranged from $2,400/ton (for PM10) to $12,900/ton (for PM25). However,
a recent rulemaking setting standards for urban busses3 determined that the cost-effectiveness of
PM control for these heavy-duty diesel engines was $10,000 - $16,000/ton. In this particular
rulemaking, rather than attempt to identify an more precise credit value for PM based on the last
measures needed for attainment, we have for simplicity's sake used $10,000/ton as a
conservative but reasonable crediting value for PM for our proposed standards.
The cost crediting was applied after all costs associated with compliance with our
proposed standards were calculated and summed. The per-vehicle tons reduced of both direct
PM and SO2 were multiplied by the respective cost-effectiveness values of $10,000/ton and
$4800/ton (see Sections VI.C.3 and VI.C.4 below for tons calculations). As a result, $53.73 of
the total costs were apportioned to SO2, while $3.96 was apportioned to direct PM. These
amounts are independent of whether we are considering a near-term or long-term cost-
VI-6
-------
Chapter VI: Cost Effectiveness
effectiveness calculation, since the total tons reduced for these two compounds is the same, on a
per-vehicle basis, in any year of the program. A summary of the costs used in our cost-
effectiveness calculations is given below in Table VI-2.
Table VI-2. Fleet Average Per-vehicle Costs Used in Cost-effectiveness
Total uncredited costs
SO2 credit allocation
Direct PM credit allocation
Total credited costs
Near-term costs
($)
230.06
-53.73
-3.96
172.37
Long-term costs
($)
188.43
-53.73
-3.96
130.74
c.
Emission Reductions
In order to determine the overall cost-effectiveness of the standards we are proposing, it
was necessary to calculate the lifetime tons of each pollutant reduced on a per vehicle basis.
This section will describe the steps involved in these calculations. In general, emission
reductions were calculated for NOx, NMHC, primary PM, and SO2 in a manner analogous to the
discounted lifetime fuel costs described in Section V.B.4.
1.
NOx and NMHC
Our proposed standards are intended primarily to reduce emissions of NOx and NMHC.
We have determined that the cost-effectiveness of our proposed standards should be determined
for both NOx and NMHC. Several past rulemakings which produced reductions in both of these
pollutants have taken an approach to cost-effectiveness that sums the NOx and NMHC emission
reductions. This approach leads to $/ton NOx+NMHC. In addition, many standards for mobile
sources have been established in terms of NOx+NMHC caps. Thus we believe that this
approach to cost-effectiveness is appropriate for our proposed Tier 2 standards as well, because
we are proposing more stringent exhaust standards for both NOx and NMHC (separately). This
approach also allows for a direct comparison to previous programs for which NOx and NMHC
were summed in the cost-effectiveness analyses.
The discounted lifetime tonnage numbers for NOx, exhaust NMHC, and evaporative
NMHC were based on average in-use emission levels developed for EPA's proposed MOBILE6
on-highway inventory model. These in-use emission levels were expressed in terms of average
gram/mile emissions for each year in a vehicle's life, up to 25 years. From this basis, lifetime
tonnage estimates were developed using the following procedure:
VI-7
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
1) Annual mileage accumulation levels proposed for MOBILE6 were applied to the
in-use emission rates for each year in a vehicle's life to generate total mass emissions
produced in each year by that vehicle.
2) The resultant mass emissions were multiplied by the probability of survival in the
appropriate year, known as the "survival" rate, from estimates for cars and trucks
published by NHTSA4.
3) A seven percent annual discount factor, compounded from the first year of the
vehicle's life, was then applied for each year to allow calculation of net present value
lifetime emissions.
Converting to tons and summing across each year results in the total discounted lifetime
per-vehicle tons. This calculation can be described mathematically as follows:
LE = [{(AVMT); • (SURVIVE); • (ER); • (K)}/(1.07)M]
Where:
LE = Discounted lifetime emissions in tons/vehicle
(AVMT); = Annual vehicle miles traveled in year i of a vehicle's operational life
(SURVIVE); = Probability of vehicle survival after i years of service
(ER); = Emission rate, g/mi in year i of a vehicle's operational life
K = Conversion factor, 1.102 x 10"6 tons/gram
i = Vehicle years of operation, counting from 1 to 25
For NOx and exhaust NMHC, we generated discounted lifetime tonnage values for each
vehicle class (LDV, LDT1, LDT2, LDT3, LDT4) using the above equation. This was done
separately for the baseline and control cases. The baseline case included the NLEV vehicle
program (LEV for LDV, LDT1 and LDT2; Tier 1 for LDT3 and LDT4) and the in-use fuel
program (RFG in the appropriate areas, modeled at 150 ppm sulfur for the summer; conventional
gasoline in the remaining areas, modeled at 330 ppm sulfur). The control case entailed the Tier
2 vehicle program (0.07 g NOx/mi and 0.09 g NMHC/mi for all vehicle classes) and fuel
program (30 ppm nationwide). Baseline and controlled sulfur levels also included the maximum
sulfur levels that would be seen by a vehicle over its lifetime in order to estimate the impacts of
catalyst irreversibility as described in Section VI. C. 2 below. Thus the actual number of sulfur
cases was four: two for the average baseline and control sulfur levels, and two more for the
maximum baseline and control sulfur levels. For each permutation of vehicle and fuel program,
tonnage estimates were also developed for EVI and non-IM areas to allow generation of a
nationwide composite tonnage estimate. The tonnage values that we calculated according to this
procedure are presented in Appendix VI-A.
Before using the tonnage values to calculate the cost-effectiveness of our proposed
program, it was necessary for us to combine the values for IM vs. no-IM areas and RFG vs.
conventional gasoline areas in an effort to represent the national scope of our proposed program.
VI-8
-------
Chapter VI: Cost Effectiveness
The weighting factors were based on an analysis of the fraction of the population in the 47 state
area (U.S. excluding California, Alaska, and Hawaii) which was located within or outside of IM
and RFG areas5. We also made a distinction between summer and winter RFG, since summer-
grade Phase II RFG having approximately 150 ppm sulfur will be used for only 40 percent of the
year, while winter-grade Phase II RFG having approximately 300 ppm sulfur will be used for the
remaining 60 percent of the year. 1998 population data was used to determine these population
fractions by state, and then nationwide weighting factors were produced from the sum of these
fractional by-state populations. The geographical results are shown in Table VI-3.
For evaporative NMHC, we based the baseline tonnage values on gram/mile emissions
projected by MOBILESb. To model our control case, we projected the gram/mile emissions
using the version of MOBILESb which was modified to reflect the benefits of our proposed Tier
2 controls. We used gram/mile emission factors from 2030 to reflect a baseline fleet consisting
entirely of Enhanced Evaporative vehicles, and a control fleet consisting of essentially all Tier 2
vehicles6. The evaporative tonnage values are presented in Appendix VI-B.
Table VI-3. Weighting Factors for NOx and NMHC Lifetime Tonnage Values
RFG program area?
Yes
Yes
No
No
IM program area?
Yes
No
Yes
No
Fraction of population
0.248
0.019
0.228
0.505
The final step before calculating the cost-effectiveness of our proposed program was to
weight the discounted lifetime tonnage values for each vehicle class by their respective fraction
of the fleet. These fractions were developed based on our projection that LDT sales will
stabilize at 60 percent of the light-duty market by 2008. This value is based on sales data
projected by auto manufacturers for 1998 model year certification. Table VI-4 presents the final
weighting factors we used to develop fleet-average tonnage values.
VI-9
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table VI-4. Vehicle Class Sales Weighting Factors
LDV
LDT1
LDT2
LDT3
LDT4
0.4
0.11
0.34
0.10
0.05
The final discounted lifetime tonnage values in the absence of sulfur irreversibility effects for an
average fleet vehicle meeting either the standards for NLEV or our proposed Tier 2 standards are
shown in Tables VI-5 and VI-6, respectively.
Table VI-5. Fleet-average, Per-vehicle Discounted
Lifetime Tons for the NLEV Baseline
Sulfur
(ppm)
800gg
305
NOx
(tons)
0.13925
0.11295
Exhaust
NMHC
(tons)
0.03623
0.03285
Evap
NMHC
(tons)
0.04192
0.04192
Table VI-6. Fleet-average, Per-vehicle Discounted
Lifetime Tons for Proposed Tier 2 Standards
Sulfur
(ppm)
80
30
NOx
(tons)
0.03557
0.02738
Exhaust
NMHC
(tons)
0.02366
0.02247
Evap
NMHC
(tons)
0.03887
0.03887
The values in Tables VI-5 and VI-6 were not used in the cost-effectiveness calculations directly.
Instead, the effects of irreversibility were first calculated according to the methodology
described in Section VI.C.2 below using the tonnage values from the tables above.
gg Tonnage values at 800 ppm and 80 ppm sulfur were used for estimating the impacts of irreversibility.
See Section VI.C.2 for details.
VI-10
-------
Chapter VI: Cost Effectiveness
2. Irreversibility
As described in Appendix B, we believe that vehicles meeting the SFTP and/or NLEV
standards will exhibit an increased tendency towards sulfur poisoning of their catalysts. As a
result of sulfur poisoning, catalyst efficiency is reduced and emissions increase. Since all
vehicles are currently certified on low sulfur fuel, current in-use emissions can be expected to be
higher than certification levels.
The increased emissions that result when an SFTP-compliant NLEV or Tier 2 vehicle is
run on high sulfur fuel is a function of the "sulfur sensitivity" of the catalyst. This aspect of
sulfur poisoning has been taken into account in our cost-effectiveness analysis by virtue of the
fact that the change in lifetime tons reduced is a function of our proposed gasoline sulfur
standard. The impacts of the sulfur sensitivities on emissions for pre-SFTP and post-SFTP
compliant vehicles are described in an EPA Technical Report7.
However, one aspect of sulfur poisoning requires special treatment in our cost-
effectiveness analysis. In SFTP-compliant vehicles, some sulfur poisoning due to the use of high
sulfur fuel often extends well beyond the time that high sulfur fuel is actually used. When an
SFTP-compliant vehicle returns to using low sulfur gasoline after having been operated on high
sulfur fuel, a degree of poisoning remains. This effect is termed "irreversibility," and is
described in detail in Appendix B. We have estimated that the irreversibility effect for SFTP-
compliant vehicles will be about 50 percent, meaning that 50 percent of the emission reductions
that would otherwise occur when changing from high to low sulfur fuel are lost due to permanent
sulfur poisoning of the catalyst. That is to say, 50 percent of the sensitivity effect is permanent
or "irreversible" regardless of the fuel sulfur level.
Since our cost-effectiveness analysis makes use of emissions summed over the life of a
vehicle, we must account for the fact that there may have been hundreds of refuelings in that
time frame with repeated switches between low and high sulfur fuel. Since the higher sulfur
fuels will be widely available, we expect vehicles to be exposed to such fuels early in their lives.
As a result, the irreversibility effect will be present for most of these vehicles' lifetimes.
Irreversibility effects on lifetime emissions can then be calculated as the difference between
lifetime emissions at high sulfur fuel and lifetime emissions at the average fuel sulfur level.
While it is possible that the irreversibility effect can be reduced or eliminated under
certain driving conditions, such as high temperature/high load driving, we believe that this is
unlikely for SFTP-compliant vehicles. The data regarding catalyst cleanup conditions for future
vehicles is quite limited. Lacking data to support the recovery of full catalyst functionality, our
analysis treats irreversibility as a permanent effect.
Under our proposed gasoline sulfur program, the average sulfur level will be 30 ppm and
the maximum allowable level will be 80 ppm. Per-vehicle lifetime emissions at these two sulfur
levels were used to determine the effect of irreversibility on Tier 2 vehicles. The Tier 2 lifetime
tonnage values for NOx and exhaust NMHC at 30 ppm, which included the effects of
VI-11
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999 _
irreversibility and which was actually used in our cost-effectiveness analysis, was calculated
from the following equation:
ILE30 = (IE).(LE80-LE30) + LE30
Where:
ILE30 = Irreversibility-impacted, discounted lifetime emissions of Tier 2 vehicles at
30 ppm sulfur in tons/vehicle, for each vehicle class
IE = Irreversibility impact, 0.50
LE80 = Discounted lifetime emissions of Tier 2 vehicles at 80 ppm sulfur in
tons/vehicle, for each vehicle class
LE30 = Discounted lifetime emissions of Tier 2 vehicles at 30 ppm sulfur in
tons/vehicle, for each vehicle class
For the NLEV vehicles forming our baseline, the average sulfur level was established as
305 ppm as described in Section VI. A. 3 above. Apart from an ASTM maximum allowable value
of 1000 ppm, there is no regulated in-use maximum value for gasoline sulfur. However, after the
year 2000, we project that more than 95 percent of gasoline will contain sulfur levels below 800
ppm. We have therefore chosen 800 as the maximum sulfur level on which NLEV vehicles will
be operated. It could be argued that 1000 ppm is a more appropriate value to represent the
maximum (or even higher, as a few in-use batches of gasoline exceed the ASTM limit). We
believe that a maximum of 800 ppm is more representative of the maximum sulfur level that the
average NLEV vehicle will be operated on, since very few vehicles will ever see sulfur levels as
high as 1000 ppm.
Per-vehicle lifetime emissions at 305 ppm and 800 ppm were used to determine the effect
of irreversibility on vehicles meeting NLEV standards. Unlike for Tier 2 vehicles, however,
NLEV standards only apply to LDV, LDT1, and LDT2, while LDT3 and LDT4 meet Tier 1
standards as well as the SFTP. As discussed in Appendix B, we believe that irreversibility
applies to any SFTP-compliant vehicle, including Tier 1 vehicles produced after the year 2000.
Thus the calculations followed the same procedure as that used for Tier 2 vehicles:
5 — (IE) • (LE800 - LE305) + LE305
Where:
ILE305 = Irreversibility-impacted, discounted lifetime emissions of SFTP-complaint
NLEV vehicles at 305 ppm sulfur in tons/vehicle, for each vehicle class
IE = Irreversibility impact, 0.50
LE800 = Discounted lifetime emissions of NLEV vehicles at 800 ppm sulfur in
tons/vehicle, for each vehicle class
LE305 = Discounted lifetime emissions of NLEV vehicles at 305 ppm sulfur in
tons/vehicle, for each vehicle class
VI-12
-------
Chapter VI: Cost Effectiveness
After assessing the impact of irreversibility on both Tier 2 and NLEV vehicles, we were
able to develop a final set of discounted lifetime tonnage values that were actually used in our
cost-effectiveness analysis. These values are given in Table VI-7.
Table VI-7. Fleet-average, Per-vehicle Discounted
Lifetime Tons Used in Cost-effectiveness Analysis
Baseline: NLEV
at 305 ppm
Proposal: Tier 2
at 30 ppm
NOx (tons)
0.12610
0.03148
Exhaust NMHC
(tons)
0.03454
0.02307
EvapNMHC
(tons)
0.04192
0.04020
Total NOx +
NMHC (tons)
0.20256
0.09475
3. Primary Particulate Matter
Vehicles meeting our proposed standards will produce reductions in both primary and
secondary parti culate matter. As described in Section VLB. 3 above, we are accounting for
reductions in primary (sulfate) PM in our cost-effectiveness analysis. Although secondary PM
reductions are not being accounted for in our cost-effectiveness analysis, they have been
included in our analysis of the health and welfare benefits of our proposed program, as described
in Section VII.
Primary PM emission reductions result from the removal of sulfur in gasoline, which
produces a commensurate reduction in the amount of sulfate PM emitted from the tailpipe. To
calculate the reduction, we have assumed that sulfate PM accounts for 1 percent of all sulfur
exiting the tailpipe on a molar basis. Primary sulfate PM exists almost entirely as sulfuric acid,
and is generally hydrated. We have assumed seven hydrations, consistent with the approach
taken in the development of EPA's NON-ROAD emissions model.
Discounted lifetime tons of primary PM reduced as a result of our proposed gasoline
sulfur standard are calculated according to the following equation:
LE = [{(AVMT); • (SURVIVE); - (FE) • (D) • (SUL) • (F) • (MC) • (K)}/(1.07)M]
Where:
LE = Discounted lifetime emissions of primary PM in tons/vehicle
(AVMT); = Annual vehicle miles traveled in year i of a vehicle's operational life
(SURVIVE); = Fraction of vehicles still operating after i years of service
FE = Fuel economy by vehicle class (see Section VLB. 4)
VI-13
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
D = Density of gasoline, 6.17 Ib/gal
SUL = Change in gasoline sulfur concentration, 2.75xlO"4 Ib sulfur/lb fuel (275 ppm)
F = Fraction of total sulfur which exits the tailpipe as primary PM, 0.01
MC = Molar conversion factor, 7 Ib sulfuric acid per Ib sulfur
K = Conversion factor, 5.0 x 10"4 tons/lb
i = Vehicle years of operation, counting from 1 to 25
After applying the above equation separately for each vehicle class and weighting the
resulting tonnage values according to the factors presented in Table VI-4, we determined that the
fleet-average, per-vehicle discounted lifetime tons of primary PM reduced is 0.000396. This is
the value that was used to determine the PM-based credit that was applied to the total costs as
described in Section VLB.3 and summarized in Table VI-2.
4. Sulfur Dioxide
The sulfur contained in gasoline exists the tailpipe as either sulfuric acid, a component of
primary particulate matter, or as sulfur dioxide (SO2). As described in Section VI. C. 2 above, we
have assumed that sulfate PM, as hydrated sulfuric acid, accounts for 1 percent of all sulfur
exiting the tailpipe on a molar basis. Thus the remaining 99 percent of sulfur exiting the tailpipe
is in the form of SO2.
Discounted lifetime tons of SO2 reduced as a result of our proposed gasoline sulfur
standard are calculated according to the following equation:
LE = [{(AVMT); • (SURVIVE); - (FE) • (D) • (SUL) • (F) • (MC) • (K)}/(1.07)M]
Where:
LE = Discounted lifetime emissions of SO2 in tons/vehicle
(AVMT); = Annual vehicle miles traveled in year i of a vehicle's operational life
(SURVIVE); = Fraction of vehicles still operating after i years of service
FE = Fuel economy by vehicle class (see Section VLB. 4)
D = Density of gasoline, 6. 17 Ib/gal
SUL = Change in gasoline sulfur concentration, 2.75xlO"4 Ib sulfur/lb fuel (275 ppm)
F = Fraction of total sulfur which exits the tailpipe as SO2, 0.99
MC = Molar conversion factor, 2 Ib SO2 per Ib sulfur
K = Conversion factor, 5.0 x 10"4 tons/lb
i = Vehicle years of operation, counting from 1 to 25
After applying the above equation separately for each vehicle class and weighting the
resulting tonnage values according to the factors presented in Table VI-4, we determined that the
fleet-average, per-vehicle discounted lifetime tons of SO2 reduced is 0.01 1 19. This is the value
that was used to determine the SO2-based credit that was applied to the total costs as described in
Section VLB. 3 and summarized in Table VI-2.
VI- 14
-------
Chapter VI: Cost Effectiveness
D.
Results
We calculated the cost-effectiveness of our proposed standards for Tier 2 exhaust, Tier 2
evaporative, and gasoline sulfur as the total per-vehicle, discounted lifetime costs divided by the
total per-vehicle, discounted lifetime tons reduced. Costs are given in Table VI-2. The tons
reduced are calculated from the values in Table VI-7 as the difference between our NLEV
baseline at our baseline sulfur level of 305 ppm, and our proposed Tier 2 standards at our
proposed sulfur standard of 30 ppm. The results are given in Table VI-8.
Table VI-8. Cost-effectiveness of the Proposed Standards
Near term
Long term
Credited
costs ($)
172.37
130.74
Uncredited
costs ($)
230.06
188.43
Tons
NOx+NMHC
0.10781
0.10781
Credited
$/ton
1599
1213
Uncredited
$/ton
2134
1748
We also evaluated the cost effectiveness of a number of alternative control options using
the methodology described in this Section. The options evaluated were:
The proposed Tier 2 emission standards with no reduction in gasoline sulfur levels;
The proposed Tier 2 emission standards with the sulfur controls proposed by API and
NPRA, which include average sulfur standards of 150 ppm in the NOx Control Region
(nominally the eastern two-thirds of the U.S.) and 300 ppm elsewhere (i.e., the West)
starting in 2004;
The proposed Tier 2 emission standards with average sulfur standard of 30 ppm in
API/NPRA NOx Control Region and 150 ppm in the West starting in 2004;
The proposed Tier 2 emission standards with an 80 ppm nationwide sulfur standard
starting in 2004;
The proposed 30 ppm nationwide sulfur standard with California Phase 2 LEV emission
standards (excluding the ZEV mandate); because these standards change from year to
year, we chose to evaluate the 2010 model year standards;
All of these alternative control options are evaluated relative to the same baseline which was
used to evaluate the cost effectiveness of the proposed Tier 2 and sulfur standards. The results
are shown in Table VI-9 below.
VI-15
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table VI-9. Alternative program options evaluated by EPA
Credited
costs ($)
Uncredited
costs ($)
Tons
NOx+NMHC
Credited
$/ton
Uncredited
$/ton
Tier 2 vehicle standards with 80 ppm nationwide
Near term
Long term
155.37
115.63
202.58
162.84
0.09921
0.09921
1566
1166
2042
1641
Tier 2 vehicle standards with no change in sulfur
Near term
Long term
128.14
93.57
128.14
93.57
0.07319
0.07319
1751
1278
1751
1278
Tier 2 vehicle standards with 150 ppm in API region, 300 ppm in non-API region
Near term
Long term
136.80
99.89
161.81
124.90
0.08719
0.08719
1569
1146
1856
1433
Tier 2 vehicle standards with 30 ppm in API region, 150 ppm in non-API region
Near term
Long term
165.20
124.51
216.57
175.88
0.10263
0.10263
1610
1213
2110
1714
California LEV-II NMOG emission standards with 30 ppm nationwide
Near term
Long term
205.02
158.75
262.70
216.43
0.11168
0.11168
1836
1421
2352
1938
As can be seen, the cost effectiveness of the five alternatives are all quite similar to that
of the proposed program. The long-term credited cost per ton of the alternatives are all within
$50 per ton of that for the proposed program, with the exception of the California LEV-II
NMOG standards. The long-term credited cost effectiveness of this program is roughly $150 per
ton higher than that of the proposed program. For reasons cited elsewhere in this Draft RIA and
in the preamble to the proposed rule, EPA chose not to propose any of these alternative control
programs in lieu of the proposed standards.
VI-16
-------
Chapter VI: Cost Effectiveness
Because the primary purpose of cost-effectiveness is to compare our proposed program to
alternative programs, we made a comparison between the values in Table VI-8 and the cost-
effectiveness of other programs. Table VI-10 summarizes the cost effectiveness of several
recent EPA actions for controlled emissions from mobile sources.
Table VI-10. Cost-effectiveness of Previously Implemented
Mobile Source Programs (Costs Adjusted to 1997 Dollars)
Program
2004 Highway HD Diesel stds
Non-road Diesel engine stds
Tier 1 vehicle controls
NLEV
Marine SI engines
On-board diagnostics
$/ton
NOx+NMHC
300
410-650
l,980-2,690gg
1,859
1,128-1,778
2,228
By comparing the values from Table VI-8 to those in Table VI-10, we can see that the
cost effectiveness of the Tier 2/gasoline sulfur standards falls within the range of these other
programs. Engine-based standards (the 2004 highway heavy-duty diesel standards, the non-road
diesel engine standards and the marine spark-ignited engine standards) have generally been less
costly than our proposed Tier 2/gasoline sulfur standards. Vehicle standards, most similar to
today's proposal, have comparable or higher values than our proposed Tier 2/gasoline sulfur
program.
The primary advantage of making comparisons to previously implemented programs is
that their cost-effectiveness values were based on a rigorous analysis and are generally accepted
as representative of the efficiency with which those programs reduce emissions. Unfortunately,
previously implemented programs can be poor comparisons because they may not be
representative of the cost-effectiveness of potential future programs. For instance, it is tempting
to look at the engine standards and conclude that more reductions at a similar low cost
effectiveness should still be available. This is especially true for the two largest categories
(highway and non-road diesel engines) where new standards have been adopted that were highly
cost effective. However, cost effectiveness was not a limiting consideration in either case.
Rather, the level of the standards selected was based on technical feasibility in the time
available.
We do not believe that significant further control is available from highway or non-road
gg Cost-effectiveness of Tier 1 standards was originally calculated separately for NOx and NMHC. A
combined cost-effectiveness was recalculated for our proposal. See internal memorandum from David Korotney to
Docket A-97-10, "Calculation of Tier 1 vehicle cost-effectiveness in terms of $/ton NOx+NMHC," document
number II-B-03.
VI-17
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
diesel engines through more stringent standards at the cost effectiveness levels shown in Table
VI-10. Based on current knowledge, the next generation of controls for these diesel engines
would require advanced after-treatment devices, still in the research and development phase.
Such controls have not yet been employed and when they become available will be more costly
and will have difficulty functioning without changes to diesel fuel.
On the vehicle side, the last two sets of standards were Tier 1 and NLEV, which had cost
effectiveness comparable to or higher than our proposed Tier 2/gasoline sulfur standards.
Compared to engines, these levels reflect the advanced (and more expensive) state of vehicle
control technology, where standards have been in effect for a much longer period than for
engines. Based on these results, Tier 2/gasoline sulfur appears to be a logical and consistent next
step in vehicle control.
The most complete source of information on the cost-effectiveness of potential future
programs is the rulemaking which revised the PM and ozone National Ambient Air Quality
Standards (NAAQS). The Regulatory Impact Analysis (RIA) associated with that rulemaking
contained a listing of potential future emission control programs and their cost-effectiveness8.
The listing categorizes control programs by mobile, point, and area source strategies for a total
of 236 potential future programs. Although the majority of the programs in this list would most
likely be implemented on a local or regional basis, they still provide the most complete
information available on alternative programs and their associated cost-effectiveness.
Of the 236 potential future programs in the NAAQS RIA, 112 produced NOx reductions
with an average cost-effectiveness of $13,000/ton, while 55 programs produced NMHC
reductions with an average cost-effectiveness of $5,000/ton. These values confirm that future
controls will be more expensive than past controls. In fact, for the purposes of evaluating the
capability of potential future controls for bringing all areas into attainment, an upper limit of
$10,000/ton was established. As a result of the analyses conducted in the context of the NAAQS
revisions rulemaking, it was determined that some areas would be willing to pay up to
$10,000/ton for local control measures in order to achieve attainment.
We recognize that the cost effectiveness calculated for our proposed program is not
strictly comparable to the $10,000/ton limit established in the NAAQS analyses since the
technologies identified there can be targeted at the specific nonattainment areas of concern,
while the proposed Tier 2/gasoline sulfur program would apply nationwide. However, we are
not using cost effectiveness to portray Tier 2 as a control strategy to select as an alternative to
local controls because of its lower cost effectiveness. Rather, the program we are proposing
today is likely one of several programs, both national and local in nature, that will be necessary
for attainment and maintenance of the NAAQS.
In summary, given the array of controls that will have to be implemented to make
progress toward attaining and maintaining the NAAQS, we believe that the weight of the
evidence from alternative means of providing substantial NOx + NMHC emission reductions
indicates that our Tier 2/gasoline sulfur proposal is cost-effective. This is true from the
perspective of other mobile source control programs or from the perspective of other stationary
VI-18
-------
Chapter VI: Cost Effectiveness
source technologies that might be considered.
VI-19
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
APPENDIX VI-A : Discounted Lifetime Tonnage Values for Exhaust
Emissions
StandardVeh clasIM case Sulfur Fuel
NOx tons NMHC tons
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
IM
IM
IM
IM
IM
IM
IM
IM
IM
No
No
No
No
No
No
No
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
No
No
No
No
No
No
No
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
No
No
No
No
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.04614
.06296
.10032
.13343
.04494
.06132
.07523
.09463
.12953
.06646
.08716
.13312
.18824
.06478
.08495
.10209
.12597
.16619
.07705
.08783
.10639
.11983
.07503
.08552
.09307
.10225
.11660
.09894
.11092
.13155
.14896
.09642
.10809
.11650
.12671
.14218
.15696
.15929
.17147
.18512
.15282
.15508
.15830
.16546
.17755
.18307
.18545
.19794
.22195
.17836
.18068
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
01839
01989
02252
02424
01565
01694
01787
01901
02061
03540
03669
03906
04059
03000
03110
03192
03297
03424
02205
02329
02535
02668
01878
01984
02059
02147
02271
03943
04051
04241
04306
03344
03436
03502
03586
03688
05429
05585
06451
07818
04632
04765
04960
05410
06646
07525
07659
08400
09850
06396
06510
VI-20
-------
Chapter VI: Cost Effectiveness
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
LDT3
LDT3
LDT3
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT1
LDT2
LDT2
No
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
No
No
No
No
No
No
No
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
No
No
No
No
No
No
No
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
No
No
No
No
No
No
No
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.18399
.19134
.20478
.23321
.23669
.25494
.28329
.22703
.23042
.23525
.24598
.27546
.26188
.26534
.28349
.30934
.25512
.25849
.26330
.27397
.30272
.03043
.04183
.06714
.08982
.02963
.04073
.05016
.06330
.08723
.03939
.05250
.08161
.11664
.03839
.05116
.06201
.07713
.10329
.02183
.02903
.04500
.05863
.02128
.02828
.03424
.04253
.05683
.04163
.05338
.07948
.11031
.04060
.05206
.06180
.07537
.09735
.02033
.02685
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
06677
07062
08046
06443
06632
07682
09441
05498
05659
05895
06442
08026
08646
08807
09702
11431
07351
07489
07690
08155
09452
01124
01224
01400
01517
00957
01043
01106
01182
01291
01892
01983
02146
02264
01605
01683
01740
01812
01903
01839
01989
02252
02470
01565
01694
01787
01901
02055
03540
03669
03906
04041
03000
03110
03192
03297
03416
01832
01982
VI-21
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT2
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT3
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDT4
LDV
LDV
LDV
LDV
LDV
LDV
LDV
IM
IM
IM
IM
IM
IM
IM
No
No
No
No
No
No
No
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
No
No
No
No
No
No
No
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
No
No
No
No
No
No
No
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
300
800
30
80
330
800
30
80
150
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
04133
05357
01982
02617
03157
03909
05191
04101
05236
07756
10723
04000
05106
06047
07357
09464
02730
03626
05614
07307
02661
03533
04274
05307
07083
05087
06519
09700
13467
04961
06358
07544
09198
11874
02970
03954
06139
08008
02894
03853
04667
05802
07763
05402
06935
10338
14375
05268
06763
08032
09802
12673
01364
01831
02868
03766
01328
01783
02170
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
02242
02459
01559
01687
01780
01893
02045
03535
03663
03898
04033
02996
03105
03187
03291
03413
02130
02302
02602
02848
01813
01960
02066
02197
02369
04114
04260
04528
04681
03486
03611
03703
03822
03979
02152
02326
02631
02883
01831
01981
02089
02221
02398
04138
04286
04559
04714
03506
03633
03728
03848
04006
01124
01224
01400
01556
00957
01043
01106
VI-22
-------
Chapter VI: Cost Effectiveness
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
2
2
2
2
2
2
2
2
2
2
2
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
LDV
IM
IM
No
No
No
No
No
No
No
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
300
800
30
80
330
800
30
80
150
300
800
RFC
RFC
Conventional
Conventional
Conventional
Conventional
RFC
RFC
RFC
RFC
RFC
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
02709
03653
02237
02905
04389
06155
02181
02832
03386
04157
05435
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
01182
01293
01892
01983
02146
02244
01605
01683
01740
01812
01930
VI-23
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
APPENDIX VI-B :
Discounted Lifetime Tonnage Values for
Evaporative Emissions
Standard
Veh class IM case
Fuel
NMHC tons
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
2 . 0 gpt
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
Tier 2
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
enhanced
LDT1
LDT1
LDT1
LDT1
LDT2
LDT2
LDT2
LDT2
LDT3
LDT3
LDT3
LDT3
LDT4
LDT4
LDT4
LDT4
LDV
LDV
LDV
LDV
LDT1
LDT1
LDT1
LDT1
LDT2
LDT2
LDT2
LDT2
LDT3
LDT3
LDT3
LDT3
LDT4
LDT4
LDT4
LDT4
LDV
LDV
LDV
LDV
IM
IM
No
No
IM
IM
No
No
IM
IM
No
No
IM
IM
No
No
IM
IM
No
No
IM
IM
No
No
IM
IM
No
No
IM
IM
No
No
IM
IM
No
No
IM
IM
No
No
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
IM
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
Conventional
RFC
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
.02835
.01793
.06791
.03537
.02835
.01793
.06791
.03537
.03216
.01972
.08730
.04301
.03216
.01972
.08730
.04301
.02184
.01208
. 04722
.02268
.02612
.01622
.06595
.03389
.02612
.01622
.06595
.03389
.02994
.01797
.08551
.04168
.02994
. 01797
.08551
.04168
.02028
.01101
.04567
.02158
VI-24
-------
Chapter VI: Cost Effectiveness
Chapter VI. References
1. U.S. EPA; Review of NAAQS for Ozone, Assessment of Scientific and Technical
Information, Office of Air Quality Planning and Standards Staff Paper; document
number: EPA-452VR-96-007
2. "Final Regulatory Impact Analysis: Control of Emissions of Air Pollution from Highway
Heavy-Duty Engines." September 16, 1997. Alan Stout, U.S. EPA, OAR/OMS/EPCD.
3. "Control of Air Pollution from New Motor Vehicles and New Motor Vehicle Engines;
Particulate Emission Regulations for 1993 Model Year Buses, Particulate Emission
Regulations for 1994 and Later Model Year Urban Buses, Test Procedures for Urban
Buses, and Oxides of Nitrogen Emission Regulations for 1998 and Later Model Year
Heavy-Duty Engines." March 24,1993. 58 FR 15781.
4. "Updated Vehicle Survivability and Travel Mileage Schedules", November 1995, U.S.
Department of Transportation / National Highway Traffic Safety Administration
(NHTSA). Tables 10-13. EPA Air Docket A-97-10.
5. See memorandum from David J. Korotney to EPA Air Docket A-97-10, "Nationwide and
regional population fractions," document No. II-B-07.
6. "Development of Light-Duty Emission Inventory Estimates in the Notice of Proposed
Rulemaking for Tier 2 and Sulfur Standards", Koupal. EPA Report No. EPA420-R-99-
005.
7. "Development of Light-Duty Emission Inventory Estimates in the Notice of Proposed
Rulemaking for Tier 2 and Sulfur Standards," Koupal. EPA Air Docket A-97-10.
8. Regulatory Impact Analysis for final rule revising the NAAQS for PM and ozone.
Appendix B, "Summary of control measures in the PM, regional haze, and ozone partial
attainment analyses." Contact: Scott Mathias, U.S. EPA, OAR/OAQPS.
VI-25
-------
Chapter VII: Benefit-Cost Analysis
Chapter VII: Benefit-Cost Analysis
While relative cost-effectiveness is the principal economic policy criterion established for
potential Tier 2 standards in the Clean Air Act, further insight regarding the merits of the
proposed standards can be provided by benefit-cost analysis (BCA). In its traditional
application, BCA estimates the economic "efficiency" of proposed standards by defining and
quantifying the various expected consequences and representing those consequences in terms of
dollars. Expressing the effects of the potential standards in dollar terms provides a means for
comparing the expected benefits of our proposed standards to the expected costs.
The basic question we sought to answer in the BCA was: "What are the net yearly
economic benefits to society of the reduction in mobile source emissions likely to be achieved by
today's proposed standards?" In designing an analysis to answer this question, we adopted an
analytical structure and sequence similar to that used in the so-called "Section 812 studies"1* to
estimate the total benefits and costs of the entire Clean Air Act. Moreover, we used many of the
same data sets, models, and assumptions actually used in the Section 812 studies and/or the
recent Regulatory Impact Analyses (RIAs) for the PM and Ozone NAAQS, and the NOX SIP
Call." By adopting the major design elements, data sets, models, and assumptions developed for
recent RIAs, we have largely relied on methods which have already received review by other
Federal Agencies, and the public. Furthermore, the data sets adopted from the Section 812
studies have received extensive review by the independent Science Advisory Board and the
public.
The BCA that we performed for our proposed standards can be thought of as having four
parts, each of which will be discussed separately in the Sections that follow. These four steps
are:
1. Calculation of the impact that our proposed standards will have on the nationwide
inventories for NOx, NMHC, SO2, and PM.
2. Air quality modeling to determine the changes in ambient concentrations of
various pollutants that will result from our proposed standards.
3. A benefits analysis to determine the changes in human health and welfare, both in
terms of number of incidences and monetary value, that result from the changes in
ambient concentrations of various pollutants.
** The "Section 812 studies" refers to (1) USEPA, Report to Congress: The Benefits and Costs of the
Clean Air Act, 1970 to 1990, October 1997 (also known as the "Section 812 Retrospective); and (2) the first in the
ongoing series of prospective studies estimating the total costs and benefits of the Clean Air Act, expected to be
published later in 1999.
""Regulatory Impact Analysis for the NOX SIP Call, FIP, and Section 126 Petitions" September 1998,
EPA-452/R-98-003
VII-1
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
4. Calculation of the costs of our proposed standards for purposes of comparison to
the monetized benefits.
Our BCA does contain a number of limitations common to all BCAs. Critical limitations
on the availability, validity, or reliability of data; limitations in the scope and capabilities of
environmental and economic effect models; and controversies and uncertainties surrounding key
underlying scientific and economic literature all contribute to an inability to estimate the
economic effects of environmental policy changes in exact and unambiguous terms. Under these
circumstances, we consider it most appropriate to view BCA as a tool to inform, but not dictate,
regulatory decisions such as the ones reflected in today's proposal. The limitations of the
assessment of benefits will be discussed in each of the following Sections as appropriate.
Despite these important uncertainties, we believe the preliminary BCA is indicative of
the range of benefits and costs associated with the standards proposed today. This is because the
analysis focuses on estimating the economic effects of the changes in air quality conditions
expected to result from today's proposed rules, rather than focusing on developing a precise
prediction of the absolute levels of air quality likely to prevail at some particular time in the
future. An analysis focusing on the changes in air quality can give useful insights into the likely
economic effects of emissions reductions of the magnitude expected to result from today's
proposed rule.
A. Emissions
In order to determine the air quality impact of our proposed standards, we first calculated
the reductions in vehicle emissions that are expected to occur as a result of those standards, and
then determined the impact of those emission reductions on the nationwide" inventories for NOx,
NMHC, SO2, and PM. This Section describes how these inventory impacts were determined.
Our analysis used the Section 812 post-CAAA scenario for 2010 as the baseline emission
estimates. This baseline inventory was also used to produce the control inventory through the
application of the estimated changes in emissions associated with our proposed Tier 2 rule.kk
We also updated the fugitive dust PM10 and PM25 emissions for the Section 812 inventory using
the National Pollutant Inventory in order to reflect significant changes to the base year
methodologies for fugitive dust categories. These changes reduced the estimates of primary PM
emissions. Fugitive dust PM10 and PM2 5 emitters whose 1990 emissions estimates were revised
include agricultural tilling, paved and unpaved roads, prescribed burning, construction activity,
•" For the purposes of air quality modeling, 'nationwide' is taken to mean the contiguous 48-states. Also,
the proposed Tier 2/gasoline sulfur standards are assumed to have no effect on vehicle emissions in California,
though air quality in California may be affected through meteorological boundary conditions.
kkPreparation of the baseline inventory is described in some detail by Woolfolk et al. (1998). E.H. Pechan
(1999) provide emissions data reflecting the incorporation of the Tier 2 rule.
VII-2
-------
Chapter VII: Benefit-Cost Analysis
and wind erosion.
The Tier 2/gasoline sulfur program we are proposing has various emission-related
components which begin at various times and in some cases phase in over time. This means that
during the early years of the program there will not be a consistent match between costs and
benefits. This is due to the fact that the full vehicle cost is incurred at the time of vehicle
purchase, while the fuel cost along with the emission reductions and benefits occur throughout
the lifetime of the vehicle. In order to more appropriately match the costs and emission
reductions of our proposed program, therefore, our BCA assumes some future year when the
fleet is fully turned over. For today's proposal this stability does not occur until well into the
future. However, for the purpose of the benefit calculations, we have no available baseline data
set beyond the year 2010, since the Section 812 inventory was developed only for this year. We
have therefore made adjustments to allow the use of 2010 as a surrogate for a future year in
which the fleet consists entirely of Tier 2 vehicles.
For emissions, we calculated reductions by treating 2010 as if the fleet had already turned
over. We did this by applying the control case emission reductions from a fully turned over fleet
(for the year 2040) to the fleet mileages for this year. Clearly, this approach does not, nor is it
intended to, predict actual expected emission reductions for 2010. This is not its purpose. It is
intended to portray the characteristics of the vehicle fleet after it is fully turned over, within the
constraint that 2010 was the latest year for which we could perform the analysis.
The resulting analysis represents a snapshot of benefits and costs in a future year in
which the light-duty fleet consists entirely of Tier 2 vehicles. As such, it depicts the maximum
emission reductions (and resultant benefits) and among the lowest costs that would be achieved
in any one year by the program on a "per mile" basis. (Note, however, that net benefits would
continue to grow over time beyond those resulting from this analysis, but only because of growth
in vehicle miles traveled.) Thus, based on the long-term costs for a fully turned over fleet, the
resulting benefit-cost ratio will be close to its maximum point (for those benefits which we have
been able to value).
At the time that we undertook the development of the benefit estimates for this rule, we
did not have quantitative estimates of the VOC emission reductions that would result from the
evaporative emission standards in the proposal. Therefore, the benefit estimates do not include
the value of the evaporative emission standard. Consistent with this, the program cost estimates
also exclude the evaporative emission control cost. Since the evaporative emission reductions
and costs are both relatively small compared to the rest of the program, they are not expected to
significantly affect the overall cost-benefit ratio.
For the purposes of assessing benefits, we estimated that the proposed Tier 2/gasoline
sulfur standards would reduce NMHC emissions by 214,443 tons and NOX emissions 1,789,318
tons for a hypothetical fully turned-over fleet of light-duty gasoline vehicles and trucks
(Korotney, 1998). These reductions would occur in all States except California, which already
meets this standard. Measured from the Section 812 2010 post-CAA scenario emission
estimates for highway vehicles, these reductions translate into a 6.1 percent and 48.9 percent
VII-3
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
annual reduction in VOC and NOX emissions, respectively, for these States and vehicle
categories. These percent reductions were used to estimate the 47-State VOC and NOX emission
reductions from light-duty vehicles and trucks in every county. Reductions (based on percent
VOC) were also used to estimate the soluble organic aerosol (SOA) emissions.
A reduction to the SO2 inventory was also made to account for expected gasoline sulfur
reduction. SO2 reductions are based on reducing the gasoline sulfur content from 330 parts per
million (ppm) to 30 ppm. There are some uncertainties introduced by the SO2 emission
estimation methods. For one, the baseline emission estimates do not account for the lower sulfur
levels in Federal or California reformulated gasoline. Thus, the baseline emission estimates
likely overestimate gasoline vehicle emitted SO2 nationwide by about 10 percent (in comparison
to the combined conventional + reformulated gasoline baseline sulfur of 305 ppm, as described
in Section VI. A.2), and in California by a factor of 10 (in comparison to their average sulfur
limit of 30 ppm). These differences are expected to have only a modest impact on SO2 benefits
attributed to the Tier 2 rule, however, because no motor vehicle SO2 benefit was estimated for
California, and 47 State benefits are only slightly overstated in Federal reformulated gasoline
areas.
Table VII-1 summarizes the emissions inventories in the 47 contiguous states for both the
baseline and control scenarios.
VII-4
-------
Chapter VII: Benefit-Cost Analysis
Table VII-1. Emission Estimates by Vehicle Type and Reductions Associated with Adoption of the Tier 2 Rule
Continental U.S. minus California — Section 812 2010 CAA Highway
Vehicle type
Light duty gas vehicle
Light duty gas truck 1
Light duty gas truck 2
Heavy duty gas vehicle
Motorcycle
Light duty diesel vehicle3
Light duty diesel trucks
Heavy duty diesel vehicle
2010 baseline emissions
Reductions due to Tier 2 rule
47-state emission estimates under Tier 2 rule
VOC
2,197,781
743,149
574,236
136,919
40,697
7
365
139,013
3,832,166
214,443
3,617,723
Vehicle Emissions
NOX
2,296,033
750,514
609,133
272,760
14,467
24
870
1,297,002
5,240,802
1,789,318
3,451,484
CO
22,746,343
7,681,457
5,947,424
1,526,289
221,551
22
819
2,123,937
40,247,842
0
40,247,842
S02
153,912
55,797
28,430
12,416
396
0
58
107,054
358,062
228,137
129,925
PM10
65,117
20,062
10,072
5,840
453
0
65
78,764
180,372
0
180,372
PM25
37,491
12,010
6,095
3,837
227
0
50
65,856
125,566
0
125,566
SOA
13,406
4,533
3,503
1,000
248
0
9
3,295
25,994
1,308
24,686
NH3
271,483
70,314
32,084
2,564
45
0
1
439
376,930
0
376,930
California emissions
Light duty gas vehicle
Light duty gas truck 1
Light duty gas truck 2
Heavy duty gas vehicle
Motorcycle
Light duty diesel vehicle
Light duty diesel trucks
Heavy duty diesel vehicle
California emissions for baseline and under Tier
2
48-state emission estimates for control scenario
65,841
17,450
8,756
5,250
3,647
2
39
12,740
113,725
3,731,448
106,110
33,335
22,425
23,561
2,030
7
146
145,980
333,595
3,785,079
965,593
304,932
154,846
112,979
24,311
11
198
84,364
1,647,234
41,895,076
24,105
8,177
4,167
1,635
60
0
13
13,013
51,170
181,095
10,198
2,936
1,474
769
68
0
14
9,599
25,059
205,431
5,872
1,752
892
507
37
0
11
8,030
17,101
142,667
402
106
53
38
22
0
1
302
925
25,611
42,528
10,302
4,701
338
7
0
0
53
57,930
434,860
"Future year emissions of SO2, PM10, PM2 5, SOA,
levels.
and NH3 from light-duty diesel vehicles are projected to be zero due to low projected vehicle mile traveled (VMT)
VII-5
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
B. Air Quality Impacts
EPA has used a regional-scale version of the Urban Airshed Model (UAM-V) to estimate
ozone air quality. Our analysis uses a Source-Receptor Matrix (S-R Matrix) based on the
Climatological Regional Dispersion Model (CRDM) is used to estimate nitrogen deposition, PM
air quality, and visibility degradation.
Section VII.B. 1 covers the estimation of ozone air quality using UAM-V. Section
VII.B.2 covers the estimation of particulate matter air quality, and Section VII.B.3 discusses the
estimation of nitrogen deposition. Finally, Section VII.B.4 covers the estimation of visibility
degradation.
1. Ozone Air Quality Estimates
EPA used the emissions inputs with a regional-scale version of UAM-V to estimate
ozone air quality. Because it accounts for spatial and temporal variations as well as differences
in the reactivity of emissions, the UAM-V is useful for evaluating the air-quality effects of the
Tier 2 rule.11
Our analysis applies the modeling system for a base-year of 1990 and for two future-year
scenarios: a 2010 baseline and a control scenario. As discussed later, we used the two separate
years because ambient air quality observations from 1990 are used to calibrate the model. The
UAM-V modeling system requires a variety of input files that contain information pertaining to
the modeling domain and simulation period. These include gridded, day-specific emissions
estimates and meteorological fields, initial and boundary conditions, and land-use information.
The model divides the U.S. into two regions: East and West. The model then segments
the area in each region into grids, each of which has several layers of air conditions that are
considered in the analysis. Using this data, the UAM-V model generates predictions of hourly
ozone concentrations for every grid. We then used the results of this process to develop 2010
ozone profiles at monitor sites by applying derived adjustment factors to the actual 1990 ozone
data at each monitor site. For areas (grids) without ozone monitoring data, we interpolated
ozone values using data from monitors surrounding the area. After completing this process, we
calculated daily and seasonal ozone metrics as inputs for the health and agriculture benefits
analysis. The Sections below provide a more detailed discussion of each of the steps in this
evaluation and a summary of the results.
"Douglas and Iwamiya (1999) provide further information on the UAM-V modeling used in this analysis.
VII-6
-------
Chapter VII: Benefit-Cost Analysis
a. Modeling Domain
For the eastern U.S., the domain is the same as the eastern U.S. domain used in EPA's
(1998b) recent analysis, "Regulatory Impact Analysis for the NOX SIP Call, FIP, and Section 126
Petitions." The domain encompasses most of the eastern U.S. and consists of two grids, as
illustrated in Figure VII-1. The shaded area of Figure VII-1 uses a relatively fine grid of 12 km,
which consists of seven vertical layers. The unshaded area of Figure VII-1 has less resolution, as
it uses a 36 km grid, which consists of five vertical layers. The top of the modeling domain is
4000 meters above ground level, for both the shaded and unshaded regions.
Figure VII-1. UAM-V Modeling Domain for Eastern U.S.
The modeling domain used to obtain results for the western U.S. comprises the
contiguous 48 states. Note that although the domain includes the entire contiguous 48 states,
results using this domain configuration were only used to estimate the effects of the Tier 2 rule
in the West (defined as the region not shown in Figure VII-1). The domain extends from 126
degrees west longitude to 66 degrees west longitude, and from 24 degrees north latitude to 52
degrees north latitude. The analysis used a grid cell size of 2/3 longitude by 1A latitude
(approximately 56 by 56 km) resulting in a 90 by 56 grid (5,040 cells) for each vertical layer,
with eight vertical layers in all.
VII-7
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
b. Simulation Periods
A simulation period is generally characterized by high ozone concentrations in one or
more portions of the U.S.; exceedances of the 1-hour National Ambient Air Quality Standard for
ozone were recorded at monitors during these periods. This study used three multi-day
simulation periods to prepare the future-year ozone profiles. For the eastern U.S. ozone analysis,
we modeled two simulation periods: 20-30 July 1993 and 7-18 July 1995. For the western U.S.
the simulation period was 1-10 July 1990.
c. UAM-V Model Output
Standard output from the UAM-V modeling system includes: (1) hourly, surface-layer
ozone concentrations (provided as hourly averages) for each grid cell; and (2) instantaneous
ozone values for all grid cells and layers for each hour of the simulation. This study extracted
hourly, surface-layer ozone concentrations for each grid-cell from the file containing hourly
average ozone values. We then used this information to calculate a set of adjustment factors for
forecasting 2010 ozone concentrations, as described in the following Section.
d. Converting Episode Estimates to Full-Season Profiles
The UAM-V runs generate surface layer hourly average ozone concentration estimates
for the limited modeled episodes which are used in conjunction with actual 1990 concentrations
to generate ozone concentrations for the entire ozone season."^ We mapped individual monitors
onto the gridded UAM-V output, and used the modeled concentrations of the corresponding grid
cells to calculate an adjustment factor.
We multiplied hourly ozone concentrations for 1990 by the adjustment factors to estimate
2010 ozone concentrations. Using the calculated adjustment factors and the observed monitor
concentrations, we created a data set containing modified observed hourly ozone concentrations
for each of the two scenarios. The Technical Support Document for this analysis details the
steps involved.
"^ The five-month ozone season for this analysis is defined as May to September for health benefits. For
agricultural benefits for some crops, the relevant growing season extends into April and into October and
November. In this analysis, no changes in ozone concentrations are assumed to occur outside the five-month ozone
season. However, the ozone metric used to estimate certain crop yield benefits requires that the baseline level of
ozone concentrations be estimated for months outside the five-month ozone season.
VII-8
-------
Chapter VII: Benefit-Cost Analysis
e. Extrapolating from Monitored to Unmonitored Locations
To model whole U.S., we needed ozone data for every location. Since actual ozone data
is only available from limited monitor sites, we needed a method to extrapolate to unmonitored
locations, in order to estimate the effects of several ozone-related health and welfare effects.™1
Given available ozone monitoring data, we obtained ozone measures (e.g., daily average) for
each location in the contiguous 48 states in two steps: (1) we converted hourly data to an ozone
measure of interest, such as the daily average, and (2) we used monitor-specific ozone measures
to extrapolate ozone measures to a grid of eight km by eight km population grid-cells. The
conversion from hourly data to ozone measures of interest is straightforward. The estimation of
ozone measures at each grid-cell uses a Voronoi Neighbor Averaging (VNA) spatial
interpolation procedure.00
The VNA procedure interpolates air
quality estimates from the monitors to the
center of each population grid-cell. The VNA
procedure is a generalization of planar
interpolation. Rather than limit the selection of
monitors to, say, three, VNA identifies the set
of monitors that best "surrounds" the center of
each grid-cell. The result of VNA is illustrated
in Figure VII-2. VNA determines the set of
monitors that best surround the grid-cell by
identifying which monitor is closest
(considering both angular direction and
horizontal distance) in each direction from the
grid-cell center. Each selected monitor will
likely be the closest monitor for multiple
directions. The set of monitors found using this
approach forms a polygon around the grid-cell
center.
The analysis of ozone impacts on agriculture adjusts the VNA approach slightly.
Because calculating the benefits for this welfare category is best accomplished by using air
Figure VII-2. VNA Spatial Interpolation
""The Technical Support Document (Abt Associates, 1999) has a map of the location of ozone monitors in
the U.S. The map shows that some areas of the country do not have many ozone monitors in close proximity to
each other.
""Interpolation between monitors is conducted using the same method as used by Abt Associates (1998) for
the NOX SIP call analysis; previously termed the "convex polygon" method, it is more accurately described as
Voronoi Neighbor Averaging (VNA) spatial interpolation, which will be used throughout this document.
VII-9
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
quality data at the county level, we used the VNA approach to estimate ozone measures for the
center of each county, rather than the eight km by eight km population grid-cell level. To
provide estimates for all counties, the analysis includes monitors that are up to 400 km from a
county centroid. (Using a shorter distance would result in some county centroids not receiving
an estimate.)
f. Ozone Air Quality Results
A summary of the ozone air quality profiles used to assess the benefits of the proposed
standards is presented in Table VII-2. The change in seasonal ozone values across the U.S.
ranges from an increase of 0.0016 ppm to -0.0028 ppm from the base case to the control, with a
spatial average of -0.0008 ppm. The population-weighted average change is somewhat lower,
-0.0004 ppm, which reflects that urban regions have smaller reductions in ozone than less
populated rural regions. The air quality technical support document for this Regulatory Impact
Assessment (RIA) (Abt Associates, 1999) contains maps showing the base case ozone
concentrations and ozone concentration changes for the control scenario. These maps only
convey information about the five-month ozone season used for the health benefits analysis. The
change in the ozone index used in the agriculture analysis (termed "SUM06" and defined in
Table VII-2) ranges from -0.0132 to 0.0087 ppm, with a spatial average of-0.0025 ppm and a
population weighted average of-0.0026 ppm.
VII-10
-------
Chapter VII: Benefit-Cost Analysis
Table VII-2. Summary of UAM-V Derived Hourly Ozone Air Quality
Statistic
2010 Base Case"
Change"
Percent Change
Seasonal Average
Minimum (ppm) b
Maximum (ppm) b
Spatial Average (ppm)
Population- Weighted Average (person-ppm) °
0.0168
0.0611
0.0305
0.0302
- 0.0028
0.0016
- 0.0008
- 0.0004
-16.7%
2.6%
-2.6%
-1.3%
Seasonal SUM06d
Minimum (ppm) b
Maximum (ppm) b
Spatial Average (ppm)
Population- Weighted Average (person-ppm) °
0.0000
0.1052
0.0122
0.0193
-0.0132
0.0087
- 0.0025
- 0.0026
0.0%
8.3%
-20.5%
-13.5%
a All values are calculated at the county centroid, using VNA spatial interpolation and allowing all monitors with a
maximum distance of 400 km. The change is defined as the control case value minus the base case value.
b The base case minimum (maximum) is the value for the county with the lowest (highest) seasonal average, where the
season is defined as May through September and all hours are included in the calculation. The change relative to the
base case picks the minimum (maximum) from the set of changes in all counties.
0 Calculated by summing the product of the projected 2010 county population and the estimated 2010 county centroid
seasonal (or SUM06) ozone concentration, and then dividing by the total population.
d SUM06 is defined as the cumulative sum of hourly ozone concentrations over 0.06 ppm that occur from Sam to 8pm
in the months of May through September.
2. PM Air Quality Estimates
Changes in concentrations of PM have an important effect on people's health and welfare.
Our analysis uses the S-R Matrix model to evaluate the air quality effects of the Tier 2 rule. The
S-R Matrix reflects the relationship between annual average PM concentration values at a single
receptor in each county (a hypothetical monitor sited at the county population centroid) and the
contribution by PM species to this concentration from each emission source (E.H. Pechan,
1996). The modeled receptors include all U.S. county centroids plus receptors in ten Canadian
provinces and 29 Mexican cities/states. The methodology used in this RIA for estimating PM air
quality concentrations using the S-R Matrix is similar to the method used in the July 1997 PM
and Ozone NAAQS RIA (U.S. EPA, 1997e). Below is a detailed discussion of the steps taken to
run the S-R Matrix and to derive the resulting changes in PM air quality.
VII-11
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
a. Climatological Regional Dispersion Model
The CRDM uses assumptions similar to the Industrial Source Complex Short Term
model (ISCST3), an EPA-recommended short range Gaussian dispersion model. CRDM
incorporates terms for wet and dry deposition and chemical conversion of SO2 and NOX to PM,
and uses climatological summaries (annual average mixing heights and joint frequency
distributions of wind speed and direction) from 100 upper air meteorological sites throughout
North America. The analysis used meteorological data for 1990 coupled with emissions data
from version 2.0 of the 1990 National Particulate Inventory to develop the S-R Matrix.
b. Development of the S-R Matrix
To develop the S-R Matrix, we modeled a nationwide total of 5,944 sources (i.e.,
industrial point, utility, area, nonroad, and motor vehicle) of primary and precursor emissions
with CRDM. In addition, we modeled secondary organic aerosols formed from anthropogenic
and biogenic VOC emissions, as well as natural sources of PM10 and PM25 (i.e., wind erosion
and wild fires). We modeled emissions of SO2, NOX, and ammonia in order to calculate
ammonium sulfate and ammonium nitrate concentrations, the primary particulate forms of
sulfate and nitrate. The CRDM produced a matrix of transfer coefficients for each of these
primary and precursor emissions. These coefficients can be applied to the emissions of any unit
(area source or individual point source) to calculate a particular source's contribution to a county
receptor's total annual average PM10 or PM2 5 concentration. Each individual unit in the
inventory is associated with one of the modeled source types (i.e., area, point sources with
effective stack height of 0 to 250 m, 250 m to 500 m, and individual point sources with effective
stack height above 500 m) for each county.
The relative concentrations in the atmosphere of ammonium sulfate and ammonium
nitrate depend on complex chemical reactions. In the presence of sulfate and nitric acid (the gas
phase oxidation product of NOX), ammonia reacts preferentially with sulfate to form particulate
ammonium sulfate rather than react with nitric acid to form particulate ammonium nitrate. We
adjusted the S-R Matrix transfer coefficients to reflect concentrations of secondarily-formed
particulates (Latimer, 1996). First, we multiplied the transfer coefficients for SO2, NOX, and
ammonia by the ratios of the molecular weights of sulfate/SO2, nitrate/nitrogen dioxide and
ammonium/ammonia to obtain concentrations of sulfate, nitrate and ammonium.pp Ammonium
nitrate forms under conditions of excess ammonium and low temperatures. For each county
receptor, the sulfate-nitrate-ammonium equilibrium is estimated based on the following
simplifying assumptions:
pp Ratio of molecular weights: Sulfate/SO2= 1.50; nitrate/nitrogen dioxide = 1.35; ammonium/ammonia =
1.06.
VII-12
-------
Chapter VII: Benefit-Cost Analysis
1. All sulfate is neutralized by ammonium;
2. Ammonium nitrate forms only when there is excess ammonium;
3. Because ammonium nitrate forms only under relatively low temperatures, annual
average particle nitrate concentrations are divided by four assuming that
sufficiently low temperatures are present only one-quarter of the year.
Finally, we calculated the total particle mass of ammonium sulfate and ammonium nitrate.qq
c. Fugitive Dust Adjustment Factor
The 1990 CRDM predictions for fugitive dust are not consistent with measured ambient
data. The CRDM-predicted average fugitive dust contribution to total PM2 5 mass is 31 percent
in the East and 32 percent in the West (as cited in: U.S. EPA, 1998b, p. 3-15). Monitoring data
from the IMPROVE network show that minerals (i.e., crustal material) comprise approximately
five percent of PM25 mass in the East and approximately 15 percent of PM25 mass in the West
(U.S. EPA, 1996b). These disparate results suggest a systematic overestimate in the fugitive
dust contribution to total PM. This overestimate is further complicated by the recognition that
the 1990 National Particulate Inventory (NPI) significantly overestimates fugitive dust
emissions. The most recent National Emissions Trends inventory indicates that the NPI
overestimates fugitive dust PM10 and PM2 5 emissions by 40 percent and 73 percent respectively"
(U.S. EPA, 1997d).
To address this bias, we applied a multiplicative factor of 0.25 nationally to fugitive dust
emissions as a reasonable first-order attempt to reconcile differences between modeled
predictions of PM25 and actual ambient data. This is the same adjustment that was used in the
NOX SIP call analysis (U.S. EPA, 1998b). This adjustment results in a fugitive dust contribution
to modeled ambient PM25 concentrations of 10 percent to 17 percent.88 Even after this
adjustment the fugitive dust fraction of total eastern PM25 mass is 10.4 percent, which is still
greater than the five percent indicated by IMPROVE monitors. However, given that the
adjustment factor appears to bring the modeled fugitive dust contribution to PM2 5 mass more
qq To calculate total particle mass of ammonium sulfate and ammonium nitrate, the anion concentrations of
sulfate and nitrate are multiplied by 1.375 and 1.290 respectively.
" Natural and man-made fugitive dust emissions account for 86 percent of PM10 emissions and 59 percent
of PM2 5 emissions in the most recent 1990 estimates in the National Emission Trends Inventory.
88 See U.S. EPA (1997b, page 6-5) for a map delineating modeling regions. Using 0.25 multiplicative
factor, fugitive dust as percentage of PM25 mass for: Central U.S. = 17.2 percent; Eastern U.S.= 10.4 percent;
Western U.S.= 10.6 percent. By comparison, without using a multiplicative factor, fugitive dust as a percentage of
PM2 5 mass for: Central U.S. = 44.6 percent; Eastern U.S. = 30.9 percent; Western U.S. = 31.5 percent.
VII-13
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
within the range of values reported from monitoring data, we adjusted the fugitive dust
contribution to total PM that is estimated by the S-R Matrix by this factor. This factor still may
result in an overprediction of the fugitive dust contribution.
d. Normalizing S-R Matrix Results to Measured Data
In an attempt to further ensure comparability between S-R Matrix results and measured
annual average PM values, the analysis calibrated the S-R results using factors developed for the
PM and Ozone NAAQS RIA (U.S. EPA, 1997 e). For the NAAQS RIA, a "calibration factor"
was developed for each monitored county." This analysis calibrated all S-R Matrix predictions
regardless of overprediction or underprediction relative to monitored values. We applied this
factor equally across all particle species contributing to the annual average PM value at a county-
level receptor.
The calibration procedure employed 1993 - 1995 PM10 ambient monitoring data from the
AIRS database following the assumptions of data completeness discussed above. The PM10 data
represent the annual average of design value monitors averaged over three years (U.S. EPA,
1997f). We eliminated the standardization for temperature and pressure from this concentration
data based upon proposed revisions to the reference method for PM10.UU
Because there is little PM2 5 monitoring data available, we developed a general linear
model to predict PM25 concentrations directly from the 1993 - 1995 PM10 values (U.S. EPA,
1996a). The analysis used a SAS™ general linear model (i.e., GLM) procedure to predict PM2 5
values as a function of season, region, and measured PM10 value. We then used these derived
PM25 data to calibrate model predictions of annual average PM25.
tt The normalization procedure was conducted for county-level modeled PM10 and PM2 5 estimates falling
into one of four air quality data tiers. The tiering scheme reflects increasing relaxation of data completeness criteria
and therefore increasing uncertainty for the annual design value (U.S. EPA, 1997c). Nationwide, Tier 1 monitored
counties cover the 504 counties with at least 50 percent data completeness and therefore have the highest level of
certainty associated with the annual design value. Tier 2 monitored counties cover 100 additional counties with at
least one data point (i.e., one 24-hour value) for each of the three years during the period 1993 -1995. Tier 3
monitored counties cover 107 additional counties with missing monitoring data for one or two of the three years
1993 - 1995. In total, Tiers 1, 2 and 3 cover 711 counties currently monitored for PM10 in the 48 contiguous states.
In 1997 the PM10 monitoring network consisted of approximately 1600 individual monitors with a coverage of
approximately 711 counties in the 48 contiguous states. Tier 4 covers the remaining 2369 non-monitored counties.
uu See Appendix J - Reference Method for PM10, Final Rule for National Ambient Air Quality Standards
for Paniculate Matter (Federal Register, Vol. 62, No. 138, p. 41, July 18, 1997).
VII-14
-------
Chapter VII: Benefit-Cost Analysis
e. Development of Annual Median PM25 Concentrations
The CRDM procedure does not directly produce estimates of daily 24-hour average PM
concentrations or annual median PM concentrations. Some health benefits have concentration-
response (C-R) functions that rely on estimates of either the daily 24-hour average or annual
median concentrations. Using historical data, EPA therefore developed 24-hour average
estimates corresponding to the 99th percentile value for PM10 and the 98th percentile value for
PM25 reflecting forms of PM10 and PM25 daily standards.
Peak-to-mean ratios (i.e., ratio of the 24-hour average value to annual average value) are
established from actual PM10 monitor data for 1993 to 1995. For PM10, the peak value is defined
exactly the way it is for the new PM10 NAAQS, i.e., the value corresponding to the 99th
percentile value of the distribution of actual daily 24-hour average PM10 values. For PM25, the
peak value is also defined exactly the way it is for the new PM2 5 NAAQS, i.e., the value
corresponding to the 98th percentile value of the distribution of estimated daily 24-hour average
PM2 5 values. In this analysis, we assumed that these historical peak-to-mean ratios hold for the
2010 model year, and applied them to the annual average PM estimates generated by the S-R
Matrix.
Starting with the annual mean and peak values developed from the S-R Matrix, we used
maximum likelihood to estimate the parameters of a distribution that are most consistent with the
S-R Matrix results. Using the parameters of the distribution, we then estimated the annual
median concentration and other representative concentrations in the distribution (e.g., 5th
percentile).
f. PM Air Quality Results
Table VII-3 provides a summary of the predicted ambient PM10 and PM2 5 concentrations
used in this study. The concentration changes are generally very small. The technical support
document for this RIA (Abt Associates, 1999) contains maps showing the base case PM
concentrations and PM concentration changes generated.
VII-15
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table VII-3. Summary of S-R Matrix Derived PM Air Quality
Statistic
PM10
Minimum Annual Mean PM10 (/^g/m3) b
Maximum Annual Mean PM10 (,ug/m3) b
Average Annual Mean PM10 (/^g/m3)
Population- Weighted Average Annual Mean PM10 (person-
Mg/m3) c
2010 Base Case
5.96
63.18
22.46
28.31
Change"
-0.64
0.00
-0.14
-0.20
Percent Change
-10.7%
0.0%
-0.6%
-0.7%
PM25
Minimum Annual Mean PM2 5 (/^g/m3) b
Maximum Annual Mean PM2 5 (/^g/m3) b
Average Annual Mean PM2 5 (/^g/m3)
Population- Weighted Average Annual Mean PM2 5 (person-
Mg/m3) c
0.86
28.02
10.75
13.00
-0.64
0.00
-0.14
-0.20
-74.4%
0.0%
-1.3%
-1.5%
a The change is defined as the control case value minus the base case value.
b The base case minimum (maximum) is the value for the county with the lowest (highest) annual average. The change
relative to the base case picks the minimum (maximum) from the set of changes in all counties.
0 Calculated by summing the product of the projected 2010 county population and the estimated 2010 county PM
concentration, and then dividing by the total population in the 48 contiguous states.
3. Nitrogen Deposition Estimates
The analysis used RADM to generate nitrogen deposition estimates. The RADM was
developed over a ten year period, 1984 - 1993, under the auspices of the National Acid
Precipitation Assessment Program (NAPAP), to address policy and technical issues associated
with acidic deposition. The model provides a scientific basis for predicting changes in
deposition and air quality resulting from changes in precursor emissions and to predict the levels
of acidic deposition in certain sensitive receptor regions. To do so requires that RADM be a
multipollutant model that predicts the oxidizing capacity of the atmosphere, including the
prediction of ozone, and chemical transformations involving oxides of sulfur and nitrogen.
NAPAP has extensively documented the development, application, and evaluation of the
RADM (Chang et al., 1987; Chang et al., 1990; Dennis et al., 1990). Several recent studies of
acidic deposition have used RADM, including EPA's 1995 Acid Deposition Standard Feasibility
Study Report to Congress (U.S. EPA, 1995), EPA's 1997 Deposition of Air Pollutants to the
Great Waters Report to Congress (U.S. EPA, 1997a), work estimating the nitrogen deposition
airshed of the Chesapeake Bay watershed (Dennis, 1997), and in the NOX SIP call (U.S. EPA,
VII-16
-------
Chapter VII: Benefit-Cost Analysis
1998a)
RADM estimates deposition in units of kilograms per hectare (kg/ha). The model
estimates wet deposition in the form of SO42", NO3", NH3, H+. It estimates dry deposition in the
form of SO2, SO4 as aerosol, O3, HNO3, NO2, H2O2. The model then maps the deposition
estimates to specific East Coast and Gulf Coast estuaries and their watersheds, which are subject
to eutrophication problems. Land-deposited nitrogen in each watershed is multiplied by a factor
of 10 percent to obtain the nitrogen load delivered via export (pass-through) to the corresponding
estuary.
Table VII-4 provides a summary of the change in nitrogen deposition estimates for
selected estuaries as a result of the Tier 2 rule1. The results represent a 10.8 percent reduction in
the average annual deposition across these estuaries.
Table VII-4. Summary of 2010 Nitrogen Deposition in Selected Estuaries
(million kg/year)
Estuary
Albemarle/Pamlico Sound
Cape Cod Bay
Chesapeake Bay
Delaware Bay
Delaware Inland Bays
Gardiners Bay
Hudson River/Raritan Bay
Long Island Sound
Massachusetts Bay
Narragansett Bay
Sarasota Bay
Tampa Bay
All Selected Estuaries
2010 Base Case
11.87
3.96
18.05
3.37
0.44
1.24
3.95
5.78
1.33
1.17
0.37
2.27
53.8
Change"
-1.27
-0.42
-1.91
-0.34
-0.04
-0.13
-0.45
-0.66
-0.14
-0.12
-0.04
-0.28
-5.8
Percent Change
-10.7%
-10.6%
-10.6%
-10.1%
-9.1%
-10.5%
-11.4%
-11.4%
-10.5%
-10.3%
-10.8%
-12.3%
-10.8%
' Change is defined here as the emissions level after implementing the Tier 2 rule minus the base case emissions.
VII-17
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
4. Visibility Degradation Estimates Using the S-R Matrix
Visibility degradation is often directly proportional to decreases in light transmittance in
the atmosphere. Scattering and absorption by both gases and particles decrease light
transmittance. To quantify changes in visibility, our analysis used a light-extinction coefficient,
based on the work of Sisler (1996), which shows the total fraction of light that is decreased per
unit distance.
The light extinction coefficient accounts for the scattering and absorption of light by both
particles and gases, and a number of factors are included in its estimation. Because fine particles
are much more efficient at light scattering than coarse particles, the analysis specifies several
fine particle species, whereas coarse particles are kept as one category. Fine particles with
significant light-extinction efficiencies include sulfates, nitrates, organic carbon, elemental
carbon (soot), and soil (Sisler, 1996).
Once we determined the light-extinction coefficient, we calculated a unitless visibility
index, called a "deciview," which we used in the valuation of visibility. The deciview metric
provides a linear scale for perceived visual changes over the entire range of conditions, from
clear to hazy. Under many scenic conditions, the average person can generally perceive a
change of one deciview.
The analysis generated visibility degradation estimates in "recreational" (e.g., federally
designated Class I areas such as national parks and recreation areas) and "residential" (non-Class
I areas) areas at the county level using the results of the S-R Matrix. The visibility benefits
analysis (see Section VII.C) distinguishes between general regional visibility degradation and
visibility degradation in certain Federally-designated Class I areas (i.e., national parks, forests,
recreation areas, wilderness areas, etc.). Therefore we separated visibility degradation estimates
into "residential" and "recreational" categories depending upon the geographic area covered by
the estimate, and summed from the county-level to one of six regions (defined in part by the
underlying study) and the nation.
Table VII-5 provides a summary of the visibility degradation estimates in terms of
deciviews. The valuation methodology for recreational visibility requires separate treatment of
visibility changes in the different regions in the U.S. Table VII-5 provides residential and
recreational visibility degradation estimates for each region. All predicted visibility changes are
small (less than one deciview), with the largest changes occurring in the Southeast and
Northeast. The air quality technical support document for this RIA (Abt Associates, 1999)
contains maps showing the base case visibility degradation and visibility degradation changes.
VII-18
-------
Chapter VII: Benefit-Cost Analysis
Table VII-5. Summary of 2010 Visibility Degradation Estimates
(deciviews)
Visibility Degradation
2010 Base Case
Change"
Percent Change
Southeast
Annual Average— Residential
Annual Average— Recreational1"
23.44
21.65
-0.19
-0.23
-0.8%
-1.1%
Southwest
Annual Average— Residential
Annual Average— Recreationalb
17.89
18.69
-0.08
-0.08
-0.4%
-0.4%
California & Nevada
Annual Average— Residential
Annual Average— Recreational1"
19.29
19.93
-0.04
-0.06
-0.2%
-0.3%
Northeast
Annual Average— Residential
Annual Average— Recreational13
21.80
17.66
-0.17
-0.06
-0.8%
-0.3%
North Central
Annual Average— Residential
Annual Average— Recreational1"
18.55
19.13
-0.11
-0.08
-0.6%
-0.4%
Northwest
Annual Average— Residential
Annual Average— Recreational1"
20.70
21.65
-0.21
-0.15
-1.0%
-0.7%
National
Annual Average— Residential
Annual Average— Recreational13
21.77
19.51
-0.16
-0.09
-0.7%
-0.5%
a The change is defined as the control case deciview level minus the base case deciview level.
b Recreational visibility averages are from the 41 Class I areas used in the benefits analysis. See Table VII-14 for
list of Class I areas.
c.
Benefits Assessment
The changes in ozone, PM, nitrogen oxides, and visibility levels described in Section
VII.B will result in changes in the health and welfare impacts associated with elevated ambient
concentrations of these pollutants. This Section describes the methods for estimating the
VII-19
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
physical magnitude and monetary value of these impacts.
Section VII.C. 1 provides an overview of the benefits methodology. Section VII.C.2
discusses issues in estimating health effects. Section VII.C.3 discusses methods and provides
estimated values for avoided incidences and monetary benefits for ozone- and PM-related health
effects. Section VII.C.4 discusses methods and provides estimated values for air pollution-
related welfare effects. Section VII.C.5 discusses the aggregation of health and welfare benefits,
and presents an estimate of total benefits. Section VII.C.6 presents sensitivity analyses, and
Section VII.C.7 discusses potential benefit categories that are not quantified due to data and/or
methodological limitations, and provides a list of analytical uncertainties, limitations, and biases.
1. Overview of Benefits Estimation
Most of the specific methods and information used in this benefit analysis are similar to
those used in the §812 Retrospective of the Benefits and Costs of the Clean Air Act and
forthcoming §812 Prospective EPA Reports to Congress, which were reviewed by EPA's
Science Advisory Board (U.S. EPA, 1997g), as well as the approach used by EPA in support of
revising the ozone and PM NAAQS (U.S. EPA, 1997e; U.S. EPA, 1997h) and the Regional
NOx SIP call (U.S. EPA, 1998b). Prior to describing the details of the approach for the benefits
analysis, it is useful to provide an overview of the approach. The overview is intended to help
the reader better identify the role of each issue described later in this Section.
The general term "benefits" refers to any and all outcomes of the regulation that
contribute to an enhanced level of social welfare. The value of "benefits" refers to the dollar
value associated with all the expected positive impacts of the regulation; that is, all regulatory
outcomes that lead to higher social welfare. If the benefits are associated with market goods and
services, the monetary value of the benefits is approximated by the sum of the predicted changes
in "consumer (and producer) surplus." If the benefits are non-market benefits (such as the risk
reductions associated with environmental quality improvements), however, other methods of
measuring benefits must be used as discussed in the text. The total value of such a good is the
sum of the dollar amounts that all those who benefit are willing to pay.
In addition to benefits, regulatory actions may also lead to unintended nonmarket costs,
that some might term "disbenefits." An example of a disbenefit of reduced ozone concentrations
is that there will be less protection from UV radiation. In order to quantify the impact of a
regulatory action, both the benefits and disbenefits should be included. However, like many
benefits, disbenefits are difficult to quantify. EPA's approach is to present as complete a set of
quantified estimates of benefits and disbenefits as possible, given the state of science at the time
of the analysis.
This conceptual economic foundation raises several relevant issues and potential
VII-20
-------
Chapter VII: Benefit-Cost Analysis
limitations for the benefits analysis of the regulation. First, the standard economic approach to
estimating environmental benefits is anthropocentric- all benefits values arise from how
environmental changes are perceived and valued by people. Thus, all near-term as well as
temporally distant future physical outcomes associated with reduced pollutant loadings need to
be predicted and then translated into the framework of present-day human activities and
concerns. Second, as noted below, it is not possible to quantify or to value all of the benefits
resulting from environmental quality improvements.
Conducting a benefits analysis for anticipated changes in air emissions is a challenging
exercise, as it requires a series to steps to be specified and understood. Figure VII-3 illustrates
these steps, which include: (1) institutional relationships and policy-making; (2) the technical
feasibility of pollution abatement; (3) the physical-chemical properties of air pollutants and their
consequent linkages to biological or ecological responses in the environment, and (4) human
responses and values associated with these changes.
VII-21
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Tier 2 Rule
Evaluate Changes in Vehicles and Fuels
Estimate Reductions in
Pollutant Emissions
Model Changes in Ambient Air Quality
Estimate Changes in Plant Damage,
Crop Yields, and Other Welfare
Effects
Estimate Changes in Adverse Human
Health Symptoms and Risk
Estimate Changes in Supply and
Value of Crops, Changes in
Visibility Levels, and Other Welfare
Effects
Estimate Value of Averted Adverse
Human Health Symptoms and Risk
Figure VII-3. Example Benefits Analysis Method
Our analysis mainly uses a "damage function" approach to estimate the adverse physical
effects from air pollution that will be avoided in the United States due to implementation of the
emission reductions required by the Tier 2 rulew This approach examines individual physical
effects, such as, say, hospital admissions, that may be affected by reductions in specific
pollutants. The total value for a given physical effect is simply the product of the number of
incidences avoided and the value per incidence avoided. The damage function approach
assumes that the benefits from individual effects are additive and independent, i.e., benefits for
one effect do not depend on benefits for a separate effect. Alternative approaches include
market-based measures include: hedonic prices, which measure the total value of a reduction in
approach.
vThe exception to this is the estimation of nitrogen deposition benefits, which uses an avoided cost
VII-22
-------
Chapter VII: Benefit-Cost Analysis
air pollution using a single metric, such as the price of a house, or contingent valuation, which
asks individuals for their total willingness to pay (WTP) for a reduction in air pollution. If the
single metric approach successfully captures the full WTP for a reduction in air pollution, then
the damage function approach should yield an estimate that is less than or equal to the estimate
from the single metric approach. All monetized estimates of benefits presented are in 1997
dollars.™
Some of the estimates of the economic value of avoided health and welfare effects are
derived from contingent valuation (CV) studies. Concerns about the reliability of value
estimates that come from CV studies have dominated debates about the methodology, since
research has shown that bias can be introduced easily into these studies, especially if they are not
carefully done. Accurately measuring willingness to pay for avoided health and welfare losses
depends on the reliability and validity of the data collected. There are several issues to consider
when evaluating study quality, including but not limited to 1) whether the sample estimates of
WTP are representative of the population WTP, 2) whether the good to be valued is
comprehended and accepted by the respondent, 3) whether the WTP elicitation format is
designed to minimize strategic responses, 4) whether WTP is sensitive to respondent familiarity
with the good, to the size of the change in the good, and to income, 5) whether the estimates of
WTP are broadly consistent with other estimates of WTP for similar goods, and (6) the extent to
which responses are consistent with established economic principles. This benefits analysis does
not attempt to list the individual strengths and weaknesses of each CV study used. However, in
some instances, such as for valuation of chronic bronchitis and residential visibility, when the
CV study reliability is questionable, we adopt alternative estimates as conservative measures of
benefits, which are presented in the low-end estimate of the range of monetized benefits. In
other instances, for example the study used to value changes in visibility at Class I areas, we
recognize potential weaknesses, but do not alter the estimates presented in the study.
In this analysis, the valuation of avoided incidences of health effects and avoided
degradation of welfare effects relies on benefits transfer. The benefits transfer approach takes
values or value functions generated by previous research and transfers them from the study to the
policy of interest. For example, we obtained the value of reduced mortality from a distribution
of values of statistical life based on 26 wage-risk and contingent valuation studies. None of the
values for the health and welfare categories valued in this benefit analysis were generated
specifically in the context of the Tier 2 rule. The validity of this approach relies on the
correlation between attributes of the policy and the studies from which the values were obtained.
Where possible, we selected studies that valued effects matching those in the policy analysis.
wwRecent analyses have been presented in 1990 dollars, such as the NOX SIP call (U.S. EPA, 1998b) and
the §812 Retrospective of the Benefits and Costs of the Clean Air Act (U.S. EPA, 1997g). The method of adjusting
from 1990 dollars to 1997 dollars depends on the basis of the benefits estimates. Benefits estimates based on cost-
of-illness are adjusted by using the consumer price indexes (CPI-Us) for medical care, while benefits estimates
based directly on estimates of WTP have been adjusted using the CPI-U for "all items."
VII-23
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
When studies were not available that exactly matched the studied effect and the policy effect, we
selected studies that matched as closely as possible, and note the differences (and where known,
potential drawbacks to their application) in the text.
The first step in a benefits analysis using this approach is the identification of the types or
categories of benefits associated with the anticipated changes in ambient air quality conditions.
The second step is the identification of relevant studies examining the relationships between air
quality and these benefit categories and studies estimating the value of avoiding damages. The
most prominent avoided damages are those related to human health risk reductions, effects on
crops and plant life, visibility, and materials damage.
It is difficult to identify all the types of benefits that might result from environmental
regulation and to value those benefits that are identified, due to the non-market nature of many
benefits categories. Since many pollution effects (e.g., adverse health or ecological effects)
traditionally have not been traded as market commodities, economists and analysts cannot look
to changes in market prices and quantities to estimate the value of these effects. This lack of
observable markets may lead to the omission of significant benefits categories from an
environmental benefits analysis. It is not possible to quantify the magnitude of this
underestimation. The more important of these omitted effect categories are shown in Table
VII-6.
VII-24
-------
Chapter VII: Benefit-Cost Analysis
Table VII-6. Unquantified Benefit Categories*
Unquantified Benefit Categories Associated with
Ozone and Nitrogen Oxides
Unquantified Benefit Categories
Associated with PM
Health
Categories
Airway responsiveness.
Pulmonary inflammation.
Increased susceptibility to respiratory infection.
Acute inflammation and respiratory cell damage.
Chronic respiratory damage/premature aging of
lungs.
Ultraviolet-B radiation (cost).
Changes in pulmonary function.
Morphological changes.
Altered host defense mechanisms.
Cancer.
Other chronic respiratory disease.
Welfare
Categories
Ecosystem and vegetation effects in Class I areas
(e.g., national parks).
Damage to urban ornamentals (e.g.,grass, flowers,
shrubs, and trees in urban areas).
Fruit and vegetable crops.
Reduced yields of tree seedlings, commercial and
non-commercial forests.
Damage to ecosystems.
Materials damage (other than consumer cleaning
cost savings).
Nitrates in drinking water.
Brown clouds.
Materials damage (other than consumer cleaning
cost savings ).
Damage to ecosystems (e.g., acid sulfate
deposition).
Nitrates in drinking water.
Brown clouds.
* Note that there are other pollutants that are reduced in conjunction with the Tier 2 rule that are not considered in this
analysis, such as carbon (a pollutant associated with global climate change).
Within each effect category, there may be several possible estimates of health and
welfare effects. Each of these possibilities represents a health or welfare "endpoint." The basic
structure of the analysis is to create a set of benefit estimates reflecting key assumptions
concerning environmental conditions and the responsiveness of human health and the
environment to changes in air quality. Total benefits are presented as the sum of non-
overlapping endpoints, to avoid double-counting benefits.
We made subj ective judgements in our analysis because of a lack of information. To
reflect the range of uncertainty regarding key assumptions- such as the appropriate PM
threshold- this analysis uses two suites of assumptions. This RIA has adopted the approach of
presenting a range of monetized benefits that reflects these uncertainties by selecting alternative
values for each of several key assumptions. Taken together, these alternative sets of assumptions
define a "high end" and a "low end" estimate for the benefits that have been monetized in this
analysis.
Table VII-7 lists the specific health and welfare effects that are included in at least one of
the assumptions sets, indicating the specific effect categories that are included in the plausible
range of benefits. This table also includes the estimates of mean WTP, or "unit values" used to
monetize the benefits for each effect. Table VII-8 highlights the key differences between the
assumption sets.
VII-25
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table VII-7. Quantified and Monetized Primary Health and Welfare Effects
Effect
Pollutant
Value per incident ($1997)
LOW HIGH
Health Effects in the Benefits Analysis
Mortality, long-term exposure - over age 30
Mortality, short-term exposure
Chronic bronchitis - all ages
Hospital admissions - all respiratory, all ages
Hospital admissions - congestive heart failure
Hospital admissions - ischemic heart disease
Any of 19 acute respiratory symptoms -adult
Acute bronchitis - children
Lower respiratory symptoms (LRS) - children
Upper respiratory symptoms (URS) - children
Work loss days (WLD) - adult
Minor restricted activity days (MRAD) - adult
PM25
Ozone
PM10
Ozone &
PM2,
PM10
PM10
Ozone
PM2,
PM10
PM10
PM2.5
PM25
$2,730,000
$0
$74,500
$9,672 (Ozone)
$9,142 (PM)
$11,931
$14,854
$22
$55
$15
$23
$102
$47
$5,894,400
$5,894,400
$319,280
$9,672 (Ozone)
$9,142 (PM)
$11,931
$14,854
$22
$55
$15
$23
$102
$47
Welfare Effects in the Benefits Analysis
Agriculture - select commodity crops
Household soiling (annual value)
Nitrogen deposition: (annual value)
Albemarle-Pamilico
Sound
Chesapeake Bay
Tampa Bay
Average nine estuaries
Decreased worker productivity
Visibility - residential
In-region recreational visibility: (annual value)
California
Southwest
Ozone
PM10
NOX
Ozone
PM and gases
PM and gases
n/a
$3.09/household/
ug/m3 change in PM10
$90/kg of nitrogen
$59/kg of nitrogen
$238/kg of nitrogen
$129/kg of nitrogen
$l/worker/10%
change in ozone
not valued
$6.43/household
/deciview
$8. 4 I/household
n/a
$3.09/household/ug/
m3 change in PM10
$90/kg of nitrogen
$59/kg of nitrogen
$238/kg of nitrogen
$129/kg of nitrogen
$l/worker/10%
change in ozone
$17/household
per deciview
$12.89/household
/deciview
$16.82/household
VII-26
-------
Chapter VII: Benefit-Cost Analysis
Effect
Southeast
Out-of-region recreational visibility: California
Southwest
Southeast
Pollutant
Value per incident ($1997)
LOW HIGH
$3.99/household
/deciview
$4.48/household
/deciview
$6 . 76/household
/deciview
$2.46/household
/deciview
$7.98/household
/deciview
$8.96/household
/deciview
$13. 51 /household
/deciview
$4. 91 /household
/deciview
Table VII-8. Key Differences Between Low and High Assumption Sets
Assumption
Threshold for PM effect
PM Mortality
Ozone-related short-term exposure mortality
Agriculture
Visibility
Infant mortality
Low
15 ug/m3
value of statistical life
year lost
excluded
low crop sensitivity to
ozone
no residential visibility
valuation
excluded
High
background
value of statistical life
included
high crop sensitivity to
ozone
recreational and
residential visibility
valued
included
2. Issues in Estimating Changes in Health Effects
This benefits analysis relies on concentration-response (C-R) functions estimated in
published epidemiological studies relating adverse health to ambient air quality. The specific C-
R functions used are included in Table VII-9. While a broad range of adverse health effects have
been associated with exposure to elevated ozone and PM levels (as noted for example in Table
VII-6), in this quantified benefit analysis only a subset of health effects are included. Health
effects are excluded from the current analysis for three reasons: (1) the possibility of double
counting (such as hospital admissions for specific respiratory diseases); (2) uncertainties in
applying effect relationships based on clinical studies (where human subjects are exposed to
various levels of air pollution in a carefully controlled and monitored laboratory situation) to the
affected population; or (3) a lack of an established C-R relationship.
When a single published study is selected as the basis of the C-R relationship between a
VII-27
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
pollutant and a given health effect, or "endpoint," applying the C-R function is straightforward.
This is the case for most of the endpoints selected for inclusion in the benefits analysis. A single
C-R function may be chosen over other potential functions because the underlying
epidemiological study used superior methods, data or techniques, or because the C-R function is
more generalized and comprehensive. For example, the study that estimated the effects of PM
on hospital admissions for all ages and all respiratory diseases is selected over studies limited to
the over age 65 population or specific categories of respiratory diseases.
An exception to the "single study" selection in the benefits analysis is mortality
associated with exposure to ozone. Estimates of premature mortality associated with short-term
exposure to PM25 and PM10, are also based on multiple estimates of the relationship between PM
and mortality, but are presented as a sensitivity analysis. When several estimated C-R
relationships between a pollutant and a given health endpoint have been selected, they are
combined or pooled to derive a single estimate of the relationship. A separate technical support
document provides details of the procedures used to combine multiple C-R functions (Abt
Associates, 1999).
Whether the C-R relationship between a pollutant and a given health endpoint is
estimated by a single function from a single study or by a pooled function of C-R functions from
several studies, we apply that same C-R relationship everywhere in the benefits analysis.
Although the C-R relationship may in fact vary somewhat from one location to another (for
example, due to differences in population susceptibilities or differences in the composition of
PM), location-specific C-R functions are generally not available. While a single function
applied everywhere may result in overestimates of incidence changes in some locations and
underestimates of incidence changes in other locations, these location-specific biases will to
some extent cancel each other out when the total incidence change is calculated. It is not
possible to know the extent or direction of the bias in the total incidence change based on
application of a single C-R function everywhere.
The remainder of this Section discusses two key issues involving the use of C-R
functions to estimate the benefits of the Tier 2 rule: baseline incidences and health effect
thresholds, i.e. levels of pollution below which changes in air quality have no impacts on health.
VII-28
-------
Chapter VII: Benefit-Cost Analysis
Table VII-9. PM and Ozone Health Concentration-Response Function Summary Data
Endpoint
Pollutant
Concentration-Response Function
Source
Functional
Form"
Averaging Time
Studied
Applied
Population1'
Pollutant
Coefficienf
Mortality
Mortality (long-term exposure) -
PM25
Mortality (short-term exposure)
PM25
Ozone
Ozone
Ozone
Ozone
Popeetal. (1995)
Kinneyetal.,(1995)
Ito and Thurston (1996)
Moolgavkar et al. (1995)
Sametetal. (1997)
log-linear
log-linear
log-linear
log-linear
log-linear
annual median
daily 1 -hour max
1 -day average
1-day average
1-day average
annual median0
daily 1 -hour max
1 -day average
1-day average
1-day average
ages 30+
all
all
all
all
0.006408
0.000000
0.000677
0.000611
0.000936
Hospital Admissions
All respiratory illnesses
Congestive heart failure
Ischemic heart disease
All respiratory illnesses
PM25/PM10
PM10
PM10
Ozone
Thurston et al. (1994)
Schwartz & Morris
(1995)
Schwartz & Morris
(1995)
Thurston et al. (1992)
linear
log-linear
log-linear
linear
1-day average
2 -day average
1-day average
daily 1 -hour max
1-day average
1-day average
1-day average
daily 1 -hour max
all
age 65+
age 65+
all
3.45 X10'8
0.00098
0.00056
0.00137
VII-29
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Endpoint
Pollutant
Concentration-Response Function
Source
Functional
Form"
Averaging Time
Studied
Applied
Population11
Pollutant
Coefficienf
Respiratory Symptoms/Illnesses not requiring hospitalization
Development of chronic
bronchitis
Acute bronchitis
Upper respiratory symptoms
(URS)
Lower respiratory symptoms
(LRS)
Any of 19 acute respiratory
symptoms
Minor restricted activity days
(MRAD)
Work loss days (WLD)
Decreased worker productivity
PM10
PM2,
PM10
PM10
Ozone
PM25
PM25
Ozone
Schwartz (1993)
Dockeryetal.(1989)
Popeetal. (1991)
Schwartz et al. (1994)
Krupnick et al. (1990)
Ostro and Rothschild
(1989)
Ostro (1987)
Crocker & Horst (1981)
and EPA (1994)
logistic
logistic
log-linear
logistic
logistic
log-linear
log-linear
percent
change
annual mean
annual mean
1 -day average
1-day average
daily 1 -hour max
2-week average
2-week average
1-day average
annual mean
annual mean4
1 -day average
1-day average
daily 1 -hour max
1-day average
1-day average
1-day average
all
ages 10-12
asthmatics,
ages 9-11
ages 8-12
ages 18-65
ages 18-65
ages 18-65
laborers
0.012
0.0298
0.0036
0.01823
0.00014
0.00741
0.0046
n/a
a The log-linear is the most common concentration-response relationship; in this case, the relationship between a change in pollutant level, APM, and the change in
incidence of the health effect, Ay, is: Ay = population* incidence rate*[exp(B*APM)-l].
b The population examined in the study and to which this analysis applies the reported concentration-response (C-R) relationship. In general, epidemiological studies
analyzed the C-R relationship for a specific age group (e.g., ages 65+) in a specific geographical area. This analysis applies the reported pollutant coefficient to all
individuals in the age group nationwide.
0 A single pollutant coefficient reported for several studies indicates a pooled analysis; see text for discussion of pooling C-R relationships across studies.
VII-30
-------
Chapter VII: Benefit-Cost Analysis
a. Baseline Incidences
The epidemiological studies of the association between pollution levels and adverse
health effects generally provide a direct estimate of the relationship of air quality changes to the
relative risk of a health effect, rather than an estimate of the absolute number of avoided cases.
For example, a typical result might be that a ten |ig/m3 decrease in daily PM2 5 levels might
decrease hospital admissions by three percent. The baseline incidence of the health effect is
necessary to convert this relative change into a number of cases.
Because most PM and ozone studies that estimate C-R functions for mortality considered
only non-accidental mortality, we adjusted county-specific baseline mortality rates used in the
estimation of PM- and ozone-related mortality to provide a better estimate of county-specific
non-accidental mortality. We multiplied each county-specific mortality rate by the ratio of
national non-accidental mortality to national total mortality (0.93). We estimated county-
specific baseline mortality incidences among individuals aged 30 and over- necessary for PM2 5-
related long-term exposure mortality, estimated by Pope et al. (1995)- by applying national age-
specific death rates to county-specific age distributions, and adjusting the resulting estimated
age-specific incidences so that the estimated total incidences (including all ages) equals the
actual county-specific total incidences.
County-level incidence rates are not available for other endpoints. The analysis used
national incidence whenever possible, because these data are most applicable to a national
assessment of benefits. However, for some studies, the only available incidence information
come from the studies themselves; in these cases, incidence in the study population is assumed to
represent typical incidence at the national level.
b. Thresholds
A very important issue in applied modeling of changes in PM is whether to apply the C-R
functions to all predicted changes in ambient concentrations, even small changes occurring at
levels approaching "anthropogenic background". Different assumptions about whether to model
thresholds, and if so, at what level, can have a major effect on the resulting benefits estimates.
We use two thresholds- a different threshold for the low, primary, and high sets of assumptions-
which are set respectively at: 1) 15 |ig/m3 for all effects except those that have a lowest observed
level higher than 15 |ig/m3; and 2) the background level of the pollutant (i.e., the pollutant level
that would occur after removing all anthropogenic emissions).
3. PM- and Ozone-related Health Effects
This Section discusses the methods used to estimate the change in the incidence of PM-
and ozone-related health effects due to the Tier 2 rule and the methods used to value this change.
VII-31
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
a. Premature Mortality
Both ozone and particulate matter have been associated with increased risk of premature
mortality, which is a very important health endpoint in this economic analysis due to the high
monetary value associated with risks to life. There are two types of exposure to elevated levels
of air pollution that may result in premature mortality. Acute (short-term) exposure (e.g.,
exposure on a given day) to peak pollutant concentrations may result in excess mortality on the
same day or within a few days of the elevated exposure. Chronic (long-term) exposure (e.g.,
exposure over a period of a year or more) to levels of pollution that are generally higher may
result in mortality in excess of what it would be if pollution levels were generally lower. The
excess mortality that occurs will not necessarily be associated with any particular episode of
elevated air pollution levels. Both types of effects are biologically plausible, and there is an
increasing body of consistent corroborating evidence from animal toxicity studies indicating that
both types of effects exist.
There are, similarly, two basic types of epidemiological studies of the relationship
between mortality and exposure to pollutants. Long-term studies (e.g., Pope et al., 1995)
estimate the association between long-term (chronic) exposure to air pollution and the survival
of members of a large study population over an extended period of time. Such studies examine
the health endpoint of concern in relation to the general long-term level of the pollutant of
concern- for example, relating annual mortality to some measure of annual pollutant level.
Daily peak concentrations would impact the results only insofar as they affect the measure of
long-term (e.g., annual) pollutant concentration. In contrast, short-term studies relate daily
levels of the pollutant to daily mortality. By their basic design, daily studies can detect acute
effects but cannot detect the effects of long-term exposures. A chronic exposure study design (a
prospective cohort study, such as the Pope study) is best able to identify the long-term exposure
effects, and may detect some of the short-term exposure effects as well. Because a long-term
exposure study may detect some of the same short-term exposure effects detected by short-term
studies, including both types of study in a benefit analysis would likely result in some degree of
double counting of benefits.
Another major advantage of the long-term study design concerns the issue of the degree
of prematurity of mortality associated with air pollution. It is possible that the short-term studies
are detecting an association between air pollution and mortality that is primarily occurring
among terminally ill people. Critics of the use of short-term studies for policy analysis purposes
correctly point out that an added risk factor that results in a terminally ill person dying a few
days or weeks earlier than they otherwise would have (known as "short-term harvesting") is
potentially included in the measured air pollutant mortality "signal" detected in such a study.
As the short-term study design does not examine individual people (it examines daily mortality
rates in large populations, typically a large city population), it is impossible to know anything
about the overall health status of the specific population that is detected as dying early. While
VII-32
-------
Chapter VII: Benefit-Cost Analysis
some of the detected excess deaths may have resulted in a substantial loss of life (measuring loss
of life in terms of lost years of remaining life), others may have lost a relatively short amount of
lifespan.
While the long-term study design is preferred, these types of studies are expensive to
conduct and consequently there are relatively few well designed long-term studies. For PM,
there has only been one high quality study accepted by the Science Advisory Board, and for
ozone, no acceptable long-term studies have been published. For this reason, our analysis used
short-term ozone mortality studies as the basis for determining ozone-related mortality benefits.
The next two Sections provide details on the measurement of changes in incidences of premature
mortality associated with changes in PM and ozone arising from implementation of the Tier 2
rule.
Estimating PM-related Premature Mortality
The benefits analysis estimated PM-related mortality using the PM2 5 relationship from
Pope et al. (1995). This decision reflects the Science Advisory Board's explicit recommendation
for modeling the mortality effects of PM in both the completed §812 Retrospective Report to
Congress and the ongoing §812 Prospective Study. The Pope et al. study estimates the
association between long-term (chronic) exposure to PM25 and the survival of members of a
large study population. This relationship is selected for use in the benefits analysis instead of
short-term (daily pollution) studies for a number of reasons.
We selected the Pope et al. (1995) long-term study as providing the best available
estimate of the relationship between PM and mortality. It is used alone- rather than considering
the total effect to be the sum of estimated short-term and long-term effects- because summing
creates the possibility of double-counting a portion of PM-related mortality. We selected the
Pope et al. study in preference to other available long-term studies because it uses better
statistical methods, has a much larger sample size, the longest exposure interval, and more
locations (51 cities) in the United States, than other studies. It is unlikely that the Pope et al.
study contains any significant amount of short-term harvesting. First, the health status of each
individual tracked in the study is known at the beginning of the study period. Persons with
known pre-existing serious illnesses were excluded from the study population. Second, the
statistical model used in the Pope study examines the question of survivability throughout the
study period (ten years). Deaths that are premature by only a few days or weeks within the
ten-year study period (for example, the deaths of terminally ill patients, triggered by a short
duration PM episode) are likely to have little impact on the calculation of the average probability
of surviving the entire ten year interval. In relation to the "Six-cities" study by Dockery et al.
(1993), the Pope et al. study found a smaller increase in excess mortality for a given PM air
quality change.
Estimating Ozone-related Premature Mortality
VII-33
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
The literature on the possible relationship between exposure to ambient ozone and
premature mortality has been evolving rapidly. Of the 28 time-series epidemiology studies
identified in the literature that report results on a possible association between daily ozone
concentrations and daily mortality (see (see: U.S. EPA, 1997e, Appendix J), 21 were published
or presented since 1995. In particular, a series of studies published in 1995 through 1997 (after
closure on the ozone Criteria Document) from multiple cities in western Europe has significantly
increased the body of studies finding a positive association. Fifteen of the 28 studies report a
statistically significant relationship between ozone and mortality, with the more recent studies
tending to find statistical significance more often than the earlier studies. The ozone-mortality
datasets have also tended to become larger in more recent studies as longer series of air quality
monitoring data have become available over time. This suggests that it may take many years of
data before the ozone effect can be separated from the daily weather and seasonal patterns with
which it tends to be correlated.
In 1997, as a part of the ozone NAAQS promulgation RIA, EPA staff reviewed this
recent literature. They identified nine studies that met a defined set of selection criteria, and
conducted a meta-analy sis of the results of the nine studies (U.S. EPA, 1997e). Our analysis
implements the same basic approach to quantifying ozone mortality as the NAAQS, with the
exception that a subset of four of the nine studies is used, representing only U.S. based
analyses.xx In a post-NAAQS RIA review of the methodology for assessing ozone mortality
effects, it was determined that the relationships between ambient ozone and mortality in the non-
U.S. study locations included in the original NAAQS-related analysis may not be representative
of the range of ozone-mortality C-R relationships in the United States. To reduce the potential
for applying inappropriate C-R functions of the ozone mortality benefits from the Tier 2 rule, the
analysis only included U.S. studies, based on the assumption that demographic and
environmental conditions on average would be more similar between the study and policy sites.
However, the full body of peer-reviewed ozone mortality studies should be considered when
evaluating the weight of evidence regarding the presence of an association between ambient
ozone concentrations and premature mortality.
Because of differences in the averaging times used in the underlying studies (some use
daily average ozone levels, while others use 1-hour daily maximum values), it is not possible to
conduct a meaningful analysis directly on the coefficients of the C-R functions. Instead, the
analysis translated each C-R function into a set of predicted mortality incidence changes that
would be estimated by that C-R function, given the set of air quality changes. We then
combined these studies to estimate the impact of ozone on mortality incidence. The technical
support document for this analysis provides additional details of this approach (Abt Associates,
1999)
Infant Mortality
xxThe U.S. study-only approach has been implemented previously in the NOX SIP call RIA (U.S. EPA,
1998b).
VII-34
-------
Chapter VII: Benefit-Cost Analysis
Woodruff et al. (1997) found a significant association between annual PM10 levels and
post-neonatal mortality (deaths of infants aged 28-51 weeks). This estimate should not overlap
with the Pope et al. (1995) estimate because the Pope et al. function is based on a population
over the age of 30. The SAB recently advised the §812 Prospective project, however, to not
include this in the §812 primary analysis at this time, primarily because the study is of a new
endpoint and the results have not been replicated in other studies in the U.S. Consequently, our
analysis includes infant mortality in the high set of assumptions.
Valuing Premature Mortality
To value the benefit of reducing premature mortality, we employ two approaches to the
calculated change in incidence. One approach, the "value of statistical lives lost" (VSL)
approach, uses information from several value-of-life studies to determine a reasonable benefit
of preventing mortality. The mean value of avoiding one statistical death is estimated to be $5.9
million in 1997 dollars (or $4.8 million in 1990 dollars as has been used in previous EPA
analyses). This represents an intermediate value from a variety of estimates that appear in the
economics literature, and is a value that EPA has frequently used in RIAs for other rules. This
estimate is the mean of a distribution fitted to the estimates from 26 value-of-life studies
identified in the §812 study as "applicable to policy analysis." The approach and set of selected
studies mirrors that of Viscusi (1992) (with the addition of two studies), and uses the same
criteria used by Viscusi in his review of value-of-life studies. The $5.9 million estimate is
consistent with Viscusi's conclusion (updated to 1997$) that "most of the reasonable estimates
of the value of life are clustered in the $3.7 to $8.6 million range." Five of the 26 studies are
contingent valuation (CV) studies, which directly solicit WTP information from subjects; the rest
are wage-risk studies, which base WTP estimates on estimates of the additional compensation
demanded in the labor market for riskier jobs. The 26 studies used to form the distribution of the
value of a statistical life are listed in Table VII-10.
The second approach for valuing premature mortality is the value of statistical life-years
lost" (VSLY) approach, which incorporates assumptions to account for the age-distribution of
the affected population. Moore and Viscusi (1998) suggest one approach for determining the
value of a statistical life-year lost. They assume that the willingness to pay to save a statistical
life is the value of a single year of life times the expected number of years of life remaining for
an individual. They suggest that a typical respondent in a mortal risk study may have a life
expectancy of an additional 35 years. Using a mean estimate of $4.8 million (1990 dollars), their
approach would yield an estimate of $137,000 per life-year lost or saved. If an individual
discounts future additional years using a standard discounting procedure. Using a 35 year life
expectancy, a $4.8 million value of a statistical life, and a 5 percent discount rate, the implied
value of each life-year lost is $293,000. A higher discount rate would produce a greater value
per life-year, and a lower discount rate would produce a lower value per life-year. The Moore
and Viscusi procedure is identical to this approach, but uses a zero discount rate. In addition to
the VSLY, the expected number of life-years saved is necessary to determine the appropriate
value for an avoided incidence of premature mortality. Based on adjustments to reflect age-
VII-35
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
specific relative premature mortality is determined to be 9.8 years. Using 9.8 years, the value of
an avoided incidence of PM-related premature mortality is then $2.2 million (1990$). Thus, for
the low-end estimate of premature mortality in this analysis we apply the value of $2.7 million in
1997 dollars per life-year saved, and the high-end estimate applies $5.9 million per life to the full
estimate of incidence.
Table VII-10. Summary of Mortality Valuation Estimates"
Study
Kneisner and Leeth (1991) (US)
Smith and Gilbert (1984)
Dillingham(1985)
Butler (1983)
Miller and Guria (1991)
Moore and Viscusi (1988)
Viscusietal. (1991)
Gegaxetal. (1985)
Marin and Psacharopoulos (1982)
Kneisner and Leeth (1991) (Australia)
Gerkingetal. (1988)
Cousineau et al. (1988)
Jones-Lee (1989)
Dillingham(1985)
Viscusi (1978; 1979)
R.S. Smith (1976)
V.K. Smith (1983)
Olson (1981)
Viscusi (1981)
R.S. Smith (1974)
Moore and Viscusi (1988)
Kneisner and Leeth (1991) (Japan)
Herzog and Schlottman (1987)
Leigh and Folson (1984)
Leigh (1987)
Garen(1988)
Type of Estimate
Labor Market
Labor Market
Labor Market
Labor Market
Contingent Valuation
Labor Market
Contingent Valuation
Contingent Valuation
Labor Market
Labor Market
Contingent Valuation
Labor Market
Contingent Valuation
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Labor Market
Valuation per Statistical Life
(millions of 1990 $)
0.7
0.9
1.1
1.4
1.5
3.1
3.3
4.1
3.4
4.1
4.2
4.4
4.7
4.8
5.0
5.6
5.8
6.4
8.0
8.8
9.0
9.3
11.2
11.9
12.8
16.6
1 Based on Viscusi (1992). The values in Viscusi have been updated to 1997 $, as detailed in (Abt Associates, 1999).
VII-36
-------
Chapter VII: Benefit-Cost Analysis
b. Chronic Bronchitis
There are a limited number of studies that have estimated the impact of air pollution on
chronic bronchitis. An important hindrance is the lack of long-term health data and the
associated air pollution levels. Schwartz (1993) and Abbey et al.(1993; 1995) provide the
evidence that long-term PM exposure gives rise to the development of chronic bronchitis in the
U.S. Following the NOX SIP call analysis (U.S. EPA, 1998b), our analysis uses the Schwartz
study to develop a C-R function linking PM to chronic bronchitis.
It should be noted that Schwartz used data on the prevalence of chronic bronchitis, not its
incidence. To use Schwartz's study and still estimate the change in incidence, there are at least
two possible approaches. The first is to simply assume that it is appropriate to use the baseline
incidence of chronic bronchitis in a C-R function with the estimated coefficient from Schwartz's
study, to directly estimate the change in incidence. The second is to estimate the percentage
change in the prevalence rate for chronic bronchitis using the estimated coefficient from
Schwartz's study in a C-R function, and then to assume that this percentage change applies to a
baseline incidence rate obtained from another source. (That is, if the prevalence declines by 25
percent with a drop in PM, then baseline incidence drops by 25 percent with the same drop in
PM.) Following work in the retrospective analysis of the Clean Air Act (U.S. EPA 1997a, pg.
D-24), our analysis uses the former approach, and estimates the change in incidence using an
annual incidence rate of 0.6 percent.
Valuing Chronic Bronchitis
PM-related chronic bronchitis is the only measured morbidity endpoint that may be
expected to last from the initial onset of the illness throughout the rest of the individual's life.
WTP to avoid chronic bronchitis would therefore be expected to incorporate the present
discounted value of a potentially long stream of costs (e.g., medical expenditures and lost
earnings) and pain and suffering associated with the illness. Two studies, Viscusi et al. (1991)
and Krupnick and Cropper (1992), provide estimates of WTP to avoid a case of chronic
bronchitis.
The Viscusi et al. and the Krupnick and Cropper studies were experimental studies
intended to examine new methodologies for eliciting values for morbidity endpoints. Although
these studies were not specifically designed for policy analysis, we believe the studies provide
reasonable estimates of the WTP for chronic bronchitis. As with other contingent valuation
studies, the reliability of the WTP estimates depends on the methods used to obtain the WTP
values. Some specific attributes of the studies may raise some questions regarding their
reliability. An alternative approach that can be use is the cost of illness (COI) approach, which
considers only the expenditures on the illness as a valuation method. This approach, however,
underestimates the true value of a change in incidence because it does not consider other
components of the valuation such as the amount an individual would be willing to pay to avoid
the illness even if they did not have medical expenses to consider. As such, it can serve as a
VII-37
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
lower bound of the value for chronic bronchitis. Therefore, this analysis values chronic
bronchitis by using the COI approach in the low-end estimate and the WTP approach for the
high-end estimate.
The COI approach for valuing chronic bronchitis uses average annual lost earnings and
average annual medical expenditures reported in Krupnick and Cropper (1990). Using a 5
percent discount rate and assuming that (1) lost earnings continue to age 65, (2) medical
expenditures are incurred until death, and (3) life expectancy is unchanged by chronic bronchitis,
the present discounted value of the stream of medical expenditures and lost earnings associated
with an average case of chronic bronchitis is estimated to be about $94,500 for a 30 year old,
$about $71,200 for a 40 year old, about $73,000 for a 50 year old, and about $50,300 for a 60
year old. The midpoint of the COI estimates across the range of ages is $72,400 per case, which
is used to value the low-end estimate of benefits for reduce incidence of chronic bronchitis.
For the WTP approach, we use two studies. The study by Viscusi et al. uses a sample
that is larger and more representative of the general population than the study by Krupnick and
Cropper (which selects people who have a relative with the disease). Thus, the valuation for the
high-end estimate is based on the distribution of WTP responses from Viscusi et al. (1991). The
WTP to avoid a case of pollution-related chronic bronchitis is derived by starting with the WTP
to avoid a severe case of chronic bronchitis, as described by Viscusi et al. (1991)yy, and adjusting
it downward to reflect (1) the decrease in severity of a case of pollution-related CB relative to
the severe case described in the Viscusi et al. study, and (2) the elasticity of WTP with respect to
severity reported in the Krupnick and Cropper (Krupnick et al., 1992) study. The technical
support document describes the adjustment procedure in more detail (Abt Associates, 1999).
The mean value of the adjusted distribution is $319,280. This is the WTP for chronic bronchitis
we used in our benefits analysis.
As expected, the WTP estimate is greater than the full COI estimate in part because it
reflects the willingness to pay to avoid the pain and suffering associated with the illness. Thus,
the COI approach has a known downward bias because it does not include a measure of an
individual's willingness to pay some amount to avoid the illness even if no medical expenses and
no loss of earnings occurred. The WTP estimate of $319,280 is from 3.4 times the COI estimate
for 30 year olds to 6.3 times the estimate for 60 year olds.
57 As previously mentioned, the Schwartz (1993) study defines a case of chronic bronchitis for the purpose
of estimating a concentration-response function. This function only examines the relationship of chronic bronchitis
and PM without differentiating between severity of cases. Therefore, an adjustment for severity is necessary to
value the benefits of reduced incidences.
VII-38
-------
Chapter VII: Benefit-Cost Analysis
c. Hospital Admissions
Both ozone and particulate matter have been associated with increased risk of premature
mortality. Each is discussed below.
Estimating Ozone-related Hospital Admissions
Our analysis estimates ozone-related hospital admissions for "all respiratory diseases,"
using a C-R function based on the work of Thurston et al. (1992). Thurston et al. examined
hospital admissions for all ages in the population. Because of the comprehensiveness of the
Thurston et al. study, it is selected over other available studies that are restricted to limited age
ranges (e.g., the population aged 65 years and older), and/or specific diagnoses (e.g., hospital
admissions for pneumonia). The age- and disease-specific effect categories are subsets of the
all-age, all-respiratory disease hospital admission category. Therefore, the benefits of avoided
hospital admissions for respiratory illnesses for all ages should be larger than the benefits for
more restricted categories. However, that is not true for the estimated benefits, based on the
available studies. The estimated relationship produces fewer benefits than either of the two
available alternatives: all respiratory disease admissions for the population over 65; or the sum of
pneumonia and chronic obstructive pulmonary disease (COPD) admissions for the population
over 65. Clearly adding the results for these study types would involve a serious amount of
double counting. Therefore, selecting the Thurston et al. study may underestimate the total
benefits of hospital admissions.
Estimating PM-related Hospital Admissions
The benefits analysis includes three PM-related hospital admissions, due to all
respiratory illnesses (Thurston et al., 1994), congestive heart failure (Schwartz and Morris,
1995), and ischemic heart disease (Schwartz and Morris, 1995). As with ozone-induced hospital
admissions, the benefits analysis relies on a study of all respiratory hospital admissions for all
age groups, rather than studies examining the population over 65.
Valuing Hospital Admissions
An individual's WTP to avoid a hospital admission will include, at a minimum, the
amount of money they pay for medical expenses (i.e., what they pay towards the hospital charge
and the associated physician charge) and the loss in earnings. In addition, however, an
individual is likely to be willing to pay some amount to avoid the pain and suffering associated
with the illness itself. That is, even if they incurred no medical expenses and no loss in earnings,
most individuals would still be willing to pay something to avoid the illness.
Because medical expenditures are to a significant extent shared by society, via medical
insurance, Medicare, etc., the medical expenditures actually incurred by the individual are likely
to be less than the total medical cost to society. The total value to society of an individual's
VII-39
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
avoidance of hospital admission, then, might be thought of as having two components: (1) the
cost of illness (COI) to society, including the total medical costs plus the value of the lost
productivity, as well as (2) the individual's WTP to avoid the illness itself.
In the absence of estimates of social WTP to avoid hospital admissions for specific
illnesses (components 1 plus 2 above), estimates of total COI (component 1) are typically used
as conservative (lower bound) estimates. Because these estimates do not include the value of
avoiding the illness itself (component 2), they are biased downward. Some analyses adjust COI
estimates upward by multiplying by an estimate of the ratio of WTP to COI, to better
approximate total WTP. Other analyses have avoided making this adjustment because of the
possibility of over adjusting — that is, possibly replacing a known downward bias with an
upward bias. The previous RIAs for PM and ozone, as well as the revised RIA for ozone and
PM NAAQS, did adjust the COI estimate upward. The COI values used in this benefits analysis
will not be adjusted to better reflect the total WTP. This is consistent with the guidance offered
by the §812 Science Advisory Board (SAB) committee.
The COI estimates used in our analysis consist of three components: estimated physician
charges (based on the average length of a hospital stay for the illness), the estimated opportunity
cost of time spent in the hospital, and estimated hospital charges.
Our analysis assumes that physician charges associated with hospital care for asthma and
chronic obstructive pulmonary disease (COPD) (two endpoints not estimated for this analysis)
provide reasonably good estimates of average physician charges associated with hospital stays
for the illness categories considered here. Abt Associates (1992) estimated that physician
charges for the first day of hospital care for asthma (in 1988) or COPD (in 1989) averaged $135
(in 1997 $); physician charges for subsequent days of hospital care averaged $50. Estimated
physician charges for a hospital stay of n days for any of the illness categories discussed below,
then, would be $135 + $50(n-l).
The opportunity cost of a day spent in the hospital is estimated, for people in the
workforce, as the value of the lost daily wage. This is estimated at $102. The study on PM and
work loss days from which this value is derived (Ostro, 1987), however, considers only
individuals 18 to 65 years old, while two of the hospital admission studies used in this analysis
("all respiratory, all ages", Thurston et al., 1994; and Thurston et al., 1992), considers all ages
for both ozone and PM. It should be noted that, because the value of a PM-related work loss day
(WLD) is elsewhere added into the total benefits analysis as a separate health endpoint,
including it as a component of the WTP to avoid a PM-related hospital admission associated
would be double counting. Additionally, because there is a not a separate work loss function for
ozone, the lost productivity is included in the cost of an ozone hospital admission, but not for
PM.
To derive estimates of the opportunity cost of a day spent in the hospital for respiratory
illness based on Thurston et al. (1994) or Thurston et al. (1992), which considered individuals of
VII-40
-------
Chapter VII: Benefit-Cost Analysis
all ages, we assumed that half of the PM- or ozone-related hospital admissions are among
individuals who are not employed, including the young and the elderly.zz We therefore estimated
the expected opportunity cost of a day spent in the hospital for an individual randomly selected
from among those admitted to the hospital for PM- or ozone-related respiratory illnesses to be
(0.5)($102) +(0.5)($51) = $76.50. However, because the value of work loss days for those in the
labor force is a separate component of the total benefit for PM, only the second component of
opportunity cost enters the PM-related "all respiratory" hospital admissions benefit, which is,
then, (0.5)($51) = $25.50.
To estimate the opportunity cost of a day spent in the hospital for an individual aged 65
or older (necessary for the ischemic heart disease and congestive heart failure hospital admission
functions for individuals 65 years and over), we assumed that such an individual is not in the
workforce. Although the value of a WLD may be an inappropriate way to estimate the
opportunity cost of a day spent in the hospital for someone who is not employed (including the
young and the elderly), this opportunity cost is positive and should not be ignored. As a rough
approximation, we assumed that, for the young, the elderly, and any other unemployed
individuals, the opportunity cost of a day spent in the hospital is one-half what it is for
individuals in the workforce, or $51.
Finally, for all hospital admissions included in this analysis, we based estimates of
hospital charges on discharge statistics provided by Elixhauser et al. (1993). The resulting Cost
of Illness values for hospital admissions are shown in Table VII-11.
zzThis is approximately the same as the ratio of employed to total population in the United States. In 1994,
for example, this ratio was (123 million)/(260 million), or 47 percent.
VII-41
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table VII-11. Derivation of Cost of Illness (COI) and Total WTP Estimates for Hospital Admissions Endpoints (1997$a)
Hospital Admissions For:
Ischemic Heart Disease, age >
65 (ICD codes 4 10-4 14)
Congestive Heart Failure, age
> 65 (ICD code 428)
PM-Related "all respiratory
illnesses," all ages (ICD codes
466, 480-482, 485, 490-493)
Ozone-Related "all respiratory
illnesses," all ages (ICD codes
466,480-486, 490-493)
Hospital
Charge
(1)
$13,996
$10,854
$8,414
$8,607
Physician
Charge
(2)
$438
$539
$488
$438
Opportunity Cost
Opportunity
Cost per day
$50.96
$50.96
$25.48
$76.44
Avg Length of
Stay (days)
1
9
8
7
Total Opportunity Cost
(3)
$357
$459
$204
$535
Total Cost of Illness (COI)
(1) + (2) + (3)
(Standard Deviation)
$14,791 ($126)
$11,852 ($166)
$9,106 ($115)
$9,580 ($93)
a Note: Two different escalation factors were used in the adjustment to 1997$. Hospital and physician charges both used escalation factors based upon the
CPI-U for medical care. The opportunity cost adjustment used an escalation factor base upon the CPI-U for "all items." The standard deviation in the Total
Cost of Illness column is based upon a weighted average of each of the three COI components.
VII-42
-------
Chapter VII: Benefit-Cost Analysis
d. Acute Bronchitis
Dockery et al. (1989) examined the relationship between PM and other pollutants on the
reported rates of chronic cough, bronchitis and chest illness, in a study of 5,422 children aged ten
to twelve. Bronchitis and chronic cough were both found to be significantly related to PM
concentrations.
Estimating WTP to avoid a case of acute bronchitis is difficult for several reasons. First,
WTP to avoid acute bronchitis itself has not been estimated. Estimation of WTP to avoid this
health endpoint therefore must be based on estimates of WTP to avoid symptoms that occur with
this illness. Second, a case of acute bronchitis may last more than one day, whereas it is a day of
avoided symptoms that is typically valued. Finally, the C-R function used in the benefit analysis
for acute bronchitis was estimated for children, whereas WTP estimates for those symptoms
associated with acute bronchitis were obtained from adults.
With these caveats in mind, we estimate WTP to avoid a case of acute bronchitis as the
midpoint between a low estimate and a high estimate. The low estimate ($16.32) is the sum of
the midrange values recommended by lEc (1994) for two symptoms believed to be associated
with acute bronchitis: coughing ($7.72) and chest tightness ($8.60). The high estimate was taken
to be twice the value of a minor respiratory restricted activity day ($47.12), or $94.24. The
midpoint between the low and high estimates is $55.26.
e. PM-related Upper Respiratory Symptoms
The benefits analysis used the C-R function for PM-related Upper Respiratory Symptoms
(URS) from Pope et al. (1991). Pope et al. describe URS as consisting of one or more of the
following symptoms: runny or stuffy nose; wet cough; and burning, aching, or red eyes. The
children in the Pope et al. study were asked to record respiratory symptoms in a daily diary, and
the daily occurrences of URS and LRS, as defined above, were related to daily PM10
concentrations. Estimates of WTP to avoid a day of symptoms are therefore appropriate
measures of benefit.
Willingness to pay to avoid a day of URS is based on symptom-specific WTPs to avoid
those symptoms identified by Pope et al. as part of the URS complex of symptoms. Three
contingent valuation (CV) studies have estimated WTP to avoid various morbidity symptoms
that are either within the URS symptom complex defined by Pope et al. (1991) or are similar to
those symptoms identified by Pope et al. In each CV study, participants were asked their WTP
to avoid a day of each of several symptoms. The three individual symptoms that were identified
as most closely matching those listed by Pope et al. for URS are cough, head/sinus congestion,
and eye irritation. A day of URS could consist of any one of seven possible "symptom
complexes" consisting of at least one of these symptoms. It is assumed that each of the seven
VII-43
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
types of URS is equally likely. The mean WTP to avoid a day of URS is therefore the average of
the mean WTPs to avoid each type of URS, or $22.96. This is the point estimate for the dollar
value for PM-related URS used in the benefit analysis. Finally, it is worth emphasizing that
what is being valued here is URS as defined by Pope et al. While other definitions of URS are
certainly possible, we used this definition of URS in the benefits analysis because it is the
incidence of this specific definition of URS that has been related to PM exposure by Pope et
f. PM-related Lower Respiratory Symptoms
Schwartz et al. (1994) estimated the relationship between Lower Respiratory Symptoms
(LRS) and PM-10 concentrations. The method for deriving a point estimate of mean WTP to
avoid a day of LRS is the same as for URS. Schwartz et al. define LRS as at least two of the
following symptoms: cough, chest pain, phlegm, and wheeze. The symptoms for which WTP
estimates are available that reasonably match those listed by Schwartz et al. for LRS are cough
(C), chest tightness (CT), coughing up phlegm (CP), and wheeze (W). A day of LRS, as defined
by Schwartz et al., could consist of any one of the 11 combinations of at least two of these four
symptoms.
We assumed that each of the eleven types of LRS is equally likely. The mean WTP to
avoid a day of LRS as defined by Schwartz et al. (1994) is therefore the average of the mean
WTPs to avoid each type of LRS, or $14.51. This is the point estimate used in the benefit
analysis for the dollar value for LRS as defined by Schwartz et al. The WTP estimates are based
on studies which considered the value of a day of avoided symptoms, whereas the Schwartz et al.
study used as its measure a case of LRS. Because a case of LRS usually lasts at least one day,
and often more, WTP to avoid a day of LRS should be a conservative estimate of WTP to avoid
a case of LRS.
Finally, as with URS, it is worth emphasizing that what is being valued here is LRS as
defined by Schwartz et al. (1994). While other definitions of LRS are certainly possible, this
definition of LRS is used in this benefit analysis because it is the incidence of this specific
definition of LRS that has been related to PM exposure by Schwartz et al.
The point estimates derived for mean WTP to avoid a day of URS and a case of LRS are
based on the assumption that WTPs are additive. For example, if WTP to avoid a day of cough
is $8.60, and WTP to avoid a day of shortness of breath is $6.14, then WTP to avoid a day of
both cough and shortness of breath is $14.74. If there are no synergistic effects among
symptoms, then it is likely that the marginal utility of avoiding symptoms decreases with the
number of symptoms being avoided. If this is the case, adding WTPs would tend to overestimate
WTP for avoidance of multiple symptoms. However, there may be synergistic effects- that is,
the discomfort from two or more simultaneous symptoms may exceed the sum of the discomforts
associated with each of the individual symptoms. If this is the case, adding WTPs would tend to
VII-44
-------
Chapter VII: Benefit-Cost Analysis
underestimate WTP for avoidance of multiple symptoms. It is also possible that people may
experience additional symptoms for which WTPs are not available, again leading to an
underestimate of the correct WTP. However, for small numbers of symptoms, the assumption of
additivity of WTPs is unlikely to result in substantive bias.
There are three sources of uncertainty in the valuation of both URS and LRS: (1) an
occurrence of URS or of LRS may be comprised of one or more of a variety of symptoms (i.e.,
URS and LRS are each potentially a "complex of symptoms"), so that what is being valued may
vary from one occurrence to another; (2) for a given symptom, there is uncertainty about the
mean WTP to avoid the symptom; and (3) the WTP to avoid an occurrence of multiple
symptoms may be greater or less than the sum of the WTPs to avoid the individual symptoms.
g. Ozone-related Any of 19 Respiratory Symptoms
The presence of "any of 19 acute respiratory symptoms" is a somewhat subjective health
effect used by Krupnick et al. (1990). Moreover, not all 19 symptoms are listed in the Krupnick
et al. study. It is therefore not clear exactly what symptoms were included in the study. Even if
all 19 symptoms were known, it is unlikely that WTP estimates could be obtained for all of the
symptoms. Finally, even if all 19 symptoms were known and WTP estimates could be obtained
for all 19 symptoms, the assumption of additivity of WTPs becomes tenuous with such a large
number of symptoms. The likelihood that all 19 symptoms would occur simultaneously,
moreover, is very small.
Acute respiratory symptoms must be either upper respiratory symptoms or lower
respiratory symptoms. In the absence of further knowledge about which of the two types of
symptoms is more likely to occur among the "any of 19 acute respiratory symptoms," we
assumed that they occur with equal probability. Because this health endpoint may also consist of
combinations of symptoms, it was also assumed that there is some (smaller) probability that
upper and lower respiratory symptoms occur together.
To value avoidance of a day of "the presence of any of 19 acute respiratory symptoms"
we therefore assumed that this health endpoint consists either of URS, or LRS, or both. We also
assumed that it is as likely to be URS as LRS and that it is half as likely to be both together.
That is, it was assumed that "the presence of any of 19 acute respiratory symptoms" is a day of
URS with 40 percent probability, a day of LRS with 40 percent probability, and a day of both
URS and LRS with 20 percent probability. Using the point estimates of WTP to avoid a day of
URS and LRS derived above, the point estimate of WTP to avoid a day of "the presence of any
of 19 acute respiratory symptoms" is:
(0.40)($22.96) + (0.40)($14.51) + (0.20)($22.96 + $14.51) = $22.48
Because this health endpoint is only vaguely defined, and because of the lack of information on
VII-45
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
the relative frequencies of the different combinations of acute respiratory symptoms that might
qualify as "any of 19 acute respiratory symptoms," the unit dollar value derived for this health
endpoint must be considered only a rough approximation.
h. Work Loss Days
Ostro (1987) estimated the impact of PM on the incidence of work-loss days (WLD) in a
national sample of the adult working population, ages 18 to 65, living in metropolitan areas.
Separate coefficients were developed for each year in the analysis (1976-1981); we then
combined these coefficients for use in this analysis.
Willingness to pay to avoid the loss of one day of work was estimated by dividing the
median weekly wage for 1990 (U.S. Bureau of the Census, 1992) by five (to get the median daily
wage). This values the loss of a day of work at the median wage for the day lost. Valuing the
loss of a day's work at the wages lost is consistent with economic theory, which assumes that an
individual is paid exactly the value of his labor.
The use of the median rather than the mean, however, requires some comment. If all
individuals in society were equally likely to be affected by air pollution to the extent that they
lose a day of work because of it, then the appropriate measure of the value of a work loss day
would be the mean daily wage. It is highly likely, however, that the loss of work days due to
pollution exposure does not occur with equal probability among all individuals, but instead is
more likely to occur among lower income individuals than among high income individuals. It is
probable, for example, that individuals who are vulnerable enough to the negative effects of air
pollution to lose a day of work as a result of exposure tend to be those with generally poorer
health care. Individuals with poorer health care have, on average, lower incomes. To estimate
the average lost wages of individuals who lose a day of work because of exposure to PM
pollution, then, would require a weighted average of all daily wages, with higher weights on the
low end of the wage scale and lower weights on the high end of the wage scale. Because the
appropriate weights are not known, however, the median wage was used rather than the mean
wage. The median is more likely to approximate the correct value than the mean because means
are highly susceptible to the influence of large values in the tail of a distribution (in this case, the
small percentage of very large incomes in the United States), whereas the median is not
susceptible to these large values. The median daily wage in 1990 was $101.92 (adjusted to 1997
$). This is the value that was used to represent work loss days (WLD).
i. Minor Restricted Activity Days
Ostro and Rothschild (1989) estimated the impact of PM25 on the incidence of minor
restricted activity days (MRAD) in a national sample of the adult working population, ages 18 to
65, living in metropolitan areas. We developed separate coefficients for each year in the analysis
VII-46
-------
Chapter VII: Benefit-Cost Analysis
(1976-1981), which were then combined for use in this analysis.
No studies are reported to have estimated WTP to avoid a minor restricted activity day
(MRAD). However, lEc (1993) has derived an estimate of WTP to avoid a minor respiratory
restricted activity day (MRRAD), using WTP estimates from Tolley et al. (1986) for avoiding a
three-symptom combination of coughing, throat congestion, and sinusitis. This estimate of WTP
to avoid a MRRAD, so defined, is $47.12. Although Ostro and Rothschild (1989) estimated the
relationship between PM25 and MRADs, rather than MRRADs (a component of MRADs), it is
likely that most of the MRADs associated with exposure to PM25 are in fact MRRADs. For the
purpose of valuing this health endpoint, then, we assumed that MRADs associated with PM
exposure may be more specifically defined as MRRADs, and therefore used the estimate of
mean WTP to avoid a MRRAD.
Any estimate of mean WTP to avoid a MRRAD (or any other type of restricted activity
day other than WLD) will be somewhat arbitrary because the endpoint itself is not precisely
defined. Many different combinations of symptoms could presumably result in some minor or
less minor restriction in activity. Krupnick and Kopp (1988) argued that mild symptoms will not
be sufficient to result in a MRRAD, so that WTP to avoid a MRRAD should exceed WTP to
avoid any single mild symptom. A single severe symptom or a combination of symptoms could,
however, be sufficient to restrict activity. Therefore WTP to avoid a MRRAD should, these
authors argue, not necessarily exceed WTP to avoid a single severe symptom or a combination
of symptoms. The "severity" of a symptom, however, is similarly not precisely defined;
moreover, one level of severity of a symptom could induce restriction of activity for one
individual while not doing so for another. The same is true for any particular combination of
symptoms.
Given that there is inherently a substantial degree of arbitrariness in any point estimate of
WTP to avoid a MRRAD (or other kinds of restricted activity days), the reasonable bounds on
such an estimate must be considered. By definition, a MRRAD does not result in loss of work.
WTP to avoid a MRRAD should therefore be less than WTP to avoid a WLD. At the other
extreme, WTP to avoid a MRRAD should exceed WTP to avoid a single mild symptom. The
highest lEc midrange estimate of WTP to avoid a single symptom is $19.30, for eye irritation.
The point estimate of WTP to avoid a WLD in the benefit analysis is $101.92. If all the single
symptoms evaluated by the studies are not severe, then the estimate of WTP to avoid a MRRAD
should be somewhere between $19.30 and $101.92. Because the lEc estimate of $47.12 falls
within this range (and acknowledging the degree of arbitrariness associated with any estimate
within this range), we used the lEc estimate as the point estimate of mean WTP to avoid a
MRRAD.
j. Worker Productivity
The benefits analysis based the valuation used to monetize benefits associated with
VII-47
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
increased worker productivity resulting from improved ozone air quality on information reported
in Crocker and Horst (1981) and summarized in EPA (1994). Crocker and Horst (1981)
examined the impacts of ozone exposure on the productivity of outdoor citrus workers. The
study measured productivity impacts as the change in income associated with a change in ozone
exposure, given as the elasticity of income with respect to ozone concentration (-0.1427). The
reported elasticity translates a ten percent reduction in ozone to a 1.4 percent increase in income.
Given the average daily income for outdoor workers engaged in strenuous activity reported by
the 1990 U.S. Census, $89.64 per day (adjusted to 1997 $), a ten percent reduction in ozone
yields approximately $1 in increased daily wages.
4.
Ozone- and PM-Related Welfare Effects
In addition to the effects on human health described above, emission reductions
attributed to the Tier 2 rule will also produce welfare (i.e., non-health) benefits. Welfare effects
cover a potentially broad range of adverse effects, including adverse impacts on plants, animals,
structural materials, visibility, and ecosystem functions. Like health effects, in order to be
included in a quantified monetary benefits analysis, all of the analytical links between changes in
emissions and the monetary value of the effects must be available. While the required analytical
components are available for certain welfare endpoints, our analysis omits many other likely or
possible welfare categories. The availability of information on each analytical step limits the
total coverage of the welfare effects. All of the welfare benefits that are quantified and included
in the benefits analysis were included in the NOX SIP call. Table VII-12 lists the welfare
categories that are included in the benefits analysis; the technical support document for this RIA
provides further detail on these endpoints (Abt Associates, 1999). Each of these categories will
be discussed separately below.
Table VII-12. Quantified Welfare Effects Included in the Benefits Analysis
Welfare Effect
Agriculture - commodity crops
Nitrogen deposition in estuarine and
coastal waters
Visibility -recreational
Visibility -residential
Household soiling
Pollutant
Ozone
NOX
PM and gases
PM and gases
PM
Study
Taylor (1993)
EPA(1998a)
Chestnut et al. (1997)
McClelland etal. (1991)
ESEERCO(1994)
VII-48
-------
Chapter VII: Benefit-Cost Analysis
a. Commodity Agricultural Crops
The economic value associated with varying levels of yield loss for ozone-sensitive
commodity crops is analyzed using the AGSIM© agricultural benefits model (Taylor et al.,
VII-49
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
1993). AGSIM© is an econometric-simulation model that is based on a large set of statistically
estimated demand and supply equations for agricultural commodities produced in the United
States. The model is capable of analyzing the effects of changes in policies (in this case, the
implementation of the Tier 2 rule) that affect commodity crop yields or production costs. The
technical support document for this RIA provides further details on AGSIM© (Abt Associates,
1999).
The measure of benefits calculated by the model is the net change in consumers' and
producers' surplus from baseline ozone concentrations to the ozone concentrations resulting from
attainment of particular standards. Using the baseline and post-control equilibria, the model
calculates the change in net consumers' and producers' surplus on a crop-by-crop basisaaa. Dollar
values are aggregated across crops for each standard. The total dollar value represents a measure
of the change in social welfare associated with the Tier 2 rule. Although the model calculates
benefits under three alternative welfare measures (perfect competition, price supports, and
modified agricultural policy), results presented here are based on the "perfect competition"
measure to reflect recent changes in agricultural subsidy programs. Under the recently revised
1996 Farm Bill, most eligible farmers have enrolled in the program to phase out government
crop price supports for the AGSIMO-relevant crops: wheat, corn, sorghum, and cotton.
For the purpose of our analysis, the model analyzed the six most economically significant
crops: corn, cotton, peanuts, sorghum, soybean, and winter wheat.bbb The model employs
biological exposure-response information derived from controlled experiments conducted by the
National Crop Loss Assessment Network (NCLAN) (1996).
b. Nitrogen Deposition
Excess nutrient loads, especially that of nitrogen, cause a variety of adverse
consequences to the health of estuarine and coastal waters. These effects include toxic and/or
noxious algal blooms such as brown and red tides, low (hypoxic) or zero (anoxic) concentrations
of dissolved oxygen in bottom waters, the loss of submerged aquatic vegetation due to the light-
filtering effect of thick algal mats, and fundamental shifts in phytoplankton community structure.
Direct C-R functions relating deposited nitrogen and reductions in estuarine benefits are not
available. The preferred willingness-to-pay based measure of benefits depends on the
availability of these C-R functions and on estimates of the value of environmental responses.
Because neither appropriate C-R functions nor sufficient information to estimate the marginal
aaa Agricultural benefits differ from other health and welfare endpoints in the length of the assumed ozone
season. For agriculture, the ozone season is assumed to extend from April to September. This assumption is made
to ensure proper calculation of the ozone statistic used in the exposure-response functions. The only crop affected
by changes in ozone during April is winter wheat.
bbb The total value for these crops in 1997 was $57 billion.
VII-50
-------
Chapter VII: Benefit-Cost Analysis
value of changes in water quality exist at present, this analysis used an avoided cost approach
instead of willingness-to-pay to generate estuary-related benefits. The use of the avoided cost
approach to establish the value of a reduction in nitrogen deposition is problematic, because
there is not a direct link between implementation of the air pollution regulation and the
abandonment of a separate costly regulatory program by some other agency, (i.e. a state
environmental agency). However, there are currently no readily available alternatives to this
approach.ccc
The avoided costs to surrounding communities of reduced nitrogen loadings were
calculated for three case study estuaries.ddd These costs are used to estimate the avoided costs for
ten East Coast estuaries, and two Gulf Coast case study estuaries for which reduced nitrogen
loadings were modeled.eee The avoided cost estimates for the ten East Coast case study estuaries,
which represent approximately half of the estuarine watershed area in square miles along the
East Coast, are then used to extrapolate avoided costs to all East Coast estuaries. The three case
study estuaries are chosen because they have agreed upon nitrogen reduction goals and the
necessary nitrogen control cost data. The remaining estuaries in this analysis are chosen based
on their potential representativeness and our ability to estimate the direct and indirect nitrogen
load from atmospheric deposition.
Our analysis values atmospheric nitrogen reductions on the basis of avoided costs
associated with agreed upon controls of nonpoint water pollution sources. We estimated benefits
using a weighted-average, locally-based cost for nitrogen removal from water pollution (U.S.
EPA, 1998a). Valuation reflects water pollution control cost avoidance based on the weighted
average cost/pound of current non-point source water pollution controls for nitrogen in the three
case study estuaries. Taking the weighted cost/pound of these available controls assumes States
will combine low cost and high cost controls, which could inflate avoided cost estimates.
Reductions in nitrogen deposition from the Tier 2 rule should impact estuaries all along
the eastern seaboard and the Gulf Coast. Nitrogen reduction programs are currently targeting
many of the estuaries in these areas due to current impairment of estuarine water quality by
excess nutrients. Some of the largest of these estuaries, including the Chesapeake Bay, have
established goals for nitrogen reduction and target dates by which these goals should be
ccc Avoided cost is only a proxy for benefits, and should be viewed as inferior to willingness-to-pay based
measures. Current research is underway to develop other approaches for valuing estuarine benefits, including
contingent valuation and hedonic property studies. However, this research is still sparse, and does not contain
sufficient information on the marginal willingness-to-pay for changes in concentrations of nitrogen (or changes in
water quality or water resources as a result of changes in nitrogen concentrations).
ddd The case study estuaries are Albemarle-Pamlico Sounds, Chesapeake Bay, and Tampa Bay.
eee The ten East Coast estuaries are Albemarle-Pamlico Sounds, Cape Cod Bay, Chesapeake Bay, Delaware
Bay, Delaware Inland Bays, Gardiners Bay, Hudson River/Raritan Bay, Long Island Sound, Massachusetts Bays,
and Narragansett Bays. The Gulf Coast estuaries are Sarasota Bay and Tampa Bay.
VII-51
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
achieved. Using the best and most easily implemented existing technologies, many of the
estuaries will not be able to achieve the stated goals by the target dates. Meeting these additional
reductions will require development of new technologies, implementation of costly existing
technologies (such as stormwater controls), or use of technologies with significant
implementation difficulties, such as agricultural best management practices (BMPs). Reductions
in nitrogen deposition from the atmosphere will directly reduce the need for these additional
costly controls. Thus, while the Tier 2 rule does not totally eliminate the need for nutrient
management programs already in place, it may substitute for some of the incremental costs and
programs (such as an agricultural BMP program) necessary to meet the nutrient reduction goals
for each estuary.
The fixed capital costs for non-point controls in the case study estuaries ranged from
$0.75 to $55.59 per pound for agricultural and other rural best management practices and from
$42.98 to $175.16 per pound for urban nonpoint source controls (stormwater controls, reservoir
management, onsite disposal system changes, onsite BMPs).^ Using these as a base, we
calculated the total fixed capital cost per pound (weighted on the basis of fractional relationship
of nitrogen load controlled for the estuary goal) for each of the case-study estuaries and applied
in the valuation of their avoided nitrogen load controlled. The weighted capital costs per pound
for the case-study estuaries are $40.95 for Albemarle-Pamlico Sounds, $26.79 for Chesapeake
Bay, and $108.36 for Tampa Bayggg. For the purposes of our analysis, EPA assumes that
estuaries that have not yet established nutrient reduction goals will utilize the same types of
nutrient management programs as projected for the case study estuaries. For the other nine
estuaries, an average capital cost per pound of nitrogen (from the three case-estuaries) of
$58.70/lb is calculated and applied; it is unclear whether this cost understates or overstates the
costs associated with reductions in these other estuaries. The other nine estuaries generally
represent smaller, more urban estuaries (like Tampa Bay), which typically have fewer technical
and financial options available to control nitrogen loadings from nonpoint sources. This may
result in higher control costs more similar to the Tampa Bay case. On the other hand, these
estuaries may have opportunities to achieve additional point source controls at a lower costs.
Also, increased public awareness of nutrification issues and technological innovation may, in the
future, result in States finding lower cost solutions to nitrogen removal.
The benefits analysis assumed that the ten included East Coast estuaries are highly or
moderately nutrient sensitive, and they represent approximately 45.46 percent of all estuarine
watershed area along the East Coast.1*11 Because NOAA data indicate that approximately 92.6
figures in the original work have been updated to 1997 $ using an all-good CPI index.
8gg The value for Tampa Bay is not a true weighted cost per pound, but a midpoint of a range of $71.89 to
$144.47 developed by Apogee Research for the control possibilities (mostly urban BMPs) in the Tampa Bay
estuary.
m There are 43 East Coast estuaries of which ten were in the sample, and 31 Gulf of Mexico estuaries of
which two are in the sample.
VII-52
-------
Chapter VII: Benefit-Cost Analysis
percent of the watershed and surface area of East Coast estuaries are highly or moderately
nutrient sensitive, it is reasonable to expect that East Coast estuaries not included in this analysis
would also benefit from reduced deposition of atmospheric nitrogen. Therefore, we scaled-up
total benefits from the ten representative East Coast estuaries to include the remainder of the
nutrient sensitive estuaries along the East Coast on the basis of estuary watershed plus water
surface area. Since the ten estuaries are assumed to be nutrient sensitive and account for 48
percent of total eastern estuarine area, we scaled-up estimates by multiplying the estimate for the
ten East Coast estuaries by 2.037 (equal to 92.6 percent divided by 45.46 percent). We then
added this figure to the benefits estimated for the two Gulf Coast estuaries for a total benefits
estimate for nitrogen deposition.
We then annualized all capital cost estimates based on a seven percent discount rate and a
typical implementation horizon for control strategies. Based on information from the three case
study estuaries, this typically ranges from five to ten years. EPA has used the midpoint of 7.5
years for annualization, which yields an annualization factor of 0.1759. Non-capital installation
costs and annual operating and maintenance costs are not included in these annual cost estimates.
Depending upon the control strategy, these costs can be significant. Reports on the
Albemarle-Pamlico Sounds indicate, for instance, that planning costs associated with control
measures comprises approximately 15 percent of capital costs. Information received from the
Association of National Estuary Programs indicates that operating and maintenance costs are
about 30 percent of capital costs, and that permitting, monitoring, and inspections costs are about
one to two percent of capital costs. For these reasons, the annual cost estimates may be
understated.
c. Household Soiling Damage
Welfare benefits also accrue from avoided air pollution damage, both aesthetic and
structural, to architectural materials and to culturally important articles. At this time, data
limitations preclude the ability to quantify benefits for all materials whose deterioration may be
promoted and accelerated by air pollution exposure. However, our analysis addresses one small
effect in this category, the soiling of households by particulate matter.
Assumptions regarding the air quality indicator are necessary to evaluate the C-R
function. PM10 and PM2 5 are both components of TSP. However, it is not clear which
components of TSP cause household soiling damage. The Criteria Document cites some
evidence that smaller particles may be primarily responsible, in which case these estimates are
conservative.
Several studies have provided estimates of the cost to households of PM soiling. The
study that is cited by ESEERCO (1994) as one of the most sophisticated and is relied upon by
EPA in its 1988 Regulatory Impact Analysis for SO2 is Manuel et al. (1982). Using a household
production function approach and household expenditure data from the 1972-73 Bureau of Labor
VII-53
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Statistics Consumer Expenditure Survey for over twenty cities in the United States, Manuel et al.
estimate the annual cost of cleaning per |ig/m3 PM per household as $1.55 ($0.59 per person
times 2.63 persons per household). This estimate is low compared with others (e.g., estimates
provided by Cummings et al. (1981) and Watson and Jaksch (1982) are about eight times and
five times greater, respectively). The ESEERCO report notes, however, that the Manuel et al.
estimate is probably downward biased because it does not include the time cost of do-it-
yourselfers. Estimating that these costs may comprise at least half the cost of PM-related
cleaning costs, they double the Manuel et al. estimate to obtain a point estimate of $3.09
(reported by ESEERCO in 1992 dollars as $2.70).
d. Visibility
Visibility effects reported earlier in this chapter are described in terms of changes in
deciview, a unitless measure useful for comparing the effects of air quality on visibility. This
measure is used in the WTP function for visibility and is directly related to two other common
visibility measures: visual range (measured in km) and light extinction (measured in km"1).
Modeled changes in visibility are measured in terms of changes in light extinction, which are
then transformed into deciviews. A change of one deciview represents a change of
approximately 10 percent in the light extinction budget, "which is a small but perceptible scenic
change under many circumstances." (Sisler, 1996) A change of less than 10 percent in the light
extinction budget represents a measurable improvement in visibility, but may not be perceptible
to the eye in many cases. All of the average regional changes in visibility are substantially less
than one deciview (i.e. less than 10 percent of the light extinction budget), and thus less than
perceptible. However, this does not mean that these changes are not real or significant. Our
assumption is then that individuals can place values on changes in visibility that may not be
perceptible. This is quite plausible if individuals are aware that many regulations lead to small
improvements in visibility which when considered together amount to perceptible changes in
visibility.
The analysis derives the residential visibility valuation estimate from the results of an
visibility study (McClelland et al., 1991). We derive a household WTP value by dividing the
value reported in McClelland et al. by the corresponding hypothesized change in deciview,
yielding an estimate of $17 per unit change in deciview. Due to the somewhat dated methods
used in the McClelland study and inconsistencies of the study with current best practices for
conducting contingent valuation studies, the reliability of the results of the McClelland is
uncertain. EPA recognizes these uncertainties, but believes a non-zero value exists for
residential visibility improvements. Without alternative studies to verify the reliability of the
WTP estimate from McClelland, the low-end estimate in this analysis does not value residential
visibility while the high-end estimate uses the $17 per unit change in deciview obtained from the
study. This value is applied to all households - including any households living in or around
national parks- in any area estimated to experience a change in visibility.
VII-54
-------
Chapter VII: Benefit-Cost Analysis
A separate valuation component is needed for valuing improvements in visibility in
national parks and other areas (collectively known as "Class I areas"). Chestnut (1997)
developed a method for estimating the value to the U.S. public of visibility improvements in
Class I visibility areas. The approach was based on the results of a 1990 Cooperative Agreement
project jointly funded by the EPA and the National Park Service, "Preservation Values For
Visibility Protection at the National Parks." Based on that contingent valuation study of
visibility improvements, Chestnut calculates a household WTP for visibility improvements in
Class I-area National Parks, capturing both use and non-use recreational values, and accounts for
geographic variations in the willingness to pay. The PM and ozone NAAQS RIA (U.S. EPA,
1997b) analysis used this method. Similar to the McClelland study, the reliability of the results
of the Chestnut study are uncertain because of inconsistency with certain elements of best
practices for conducting contingent valuation. Contingent valuation is a rapidly developing field
and new methodologies for study design are continually evolving. As such, studies developed
during the late 1980's and early 1990's may differ in some elements of study design from more
recent studies. EPA recognizes that there are some important aspects of the Chestnut study that
are still useful for providing valuations associated with recreational visibility improvements. In
the author's judgement, the WTP value derived in the Chestnut study "may be indicative of an
accuracy no better than ± 50percent (Chestnut and Rowe, 1990)." Due to these uncertainties,
the low-end estimate presents a conservative estimate of WTP for recreational visibility
improvements that reflects the lower-bound of the variation (-50 percent).
More specifically, the Preservation Values study examined the demand for visibility in
Class I-area National Parks in three broad regions of the country, California, Southwest, and
Southeast. Because the Tier 2 rule has an impact on ambient pollution in all states - even in
California, due to drift from neighboring states - all three regions are relevant to the visibility
analysis. For a given region, the Preservation Values study asked respondents in Arizona,
California, Missouri, New York and Virginia for their willingness to pay to protect visibility at
National Parks in that region. Table VII-13 lists the parks included in the study in the study
regions, as well as the parks in other regions specifically mentioned in the Preservation Values
study. These other parks are used in estimating the visibility benefits in the "transfer regions",
as described below.
VII-55
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table VII-13. Class I Areas Included in Visibility Study By Region
Visibility Region
California & Nevada
Southwest
Southeast
Transfer Region
Northwest
(transfer from
California & Nevada)
North Central
Northeast
National Parks
Yosemite, Sequoia/Kings Canyon, Redwoods, Pinnacles, Lava Beds, Death Valley, Lassen
Volcanic, Joshua Tree, Point Reyes
Grand Canyon, Mesa Verde, Arches, Bandelier, Capitol Reef, Carlsbad Caverns, Bryce
Canyon, Chiricahua, Zion, Saguaro, Canyonlands, Petrified Forest, Rocky Mountain
Shenandoah, Great Smoky Mountains, Mammoth Cave, Everglades
National Parks
Crater Lake, Mount Rainier, North Cascades, Olympic
Yellowstone/Grand Tetons, Badlands, Craters of the Moon, Glacier, Theodore Roosevelt,
Wind Cave
Acadia, Big Bend, Guadalupe Mountains, Isle Royale, Voyageurs
Note: The "indicator" park (where identified) is shown in bold for each regions. In each case the indicator park is a
well-known park in that region. Source: Chestnut (1997).
Photos from each region's "indicator park" were provided as part of the survey
instrument. After a number of preparatory questions, respondents reached the WTP section of
the survey. Respondents were first instructed that their answer to the WTP question applied only
to the region in their survey, and that they did not have to worry about other regions of the
country. After furnishing their WTP, respondents were asked what portion of their stated total
value was for visibility at the indicator park alone. To avoid including benefits outside of the
region, the reported answers were appropriately adjusted. All of these safeguards make it less
likely that there will be overlap between urban (i.e., "residential") and National Park (i.e.,
"recreational") visibility benefits.111
When estimating the benefits attributable to visibility improvement at specific Class I
recreational parks, adjustments can be made to account for the location of parks, whether the
people valuing the park live "in-region" or "out-of-region," and whether or not the park is an
"indicator park." These issues are discussed below.
First, because the regional distribution of national parks throughout the U.S. is so varied,
the estimated WTP coefficient per change in deciview changes in value depending upon the
location of the Class I area. Based on the National Parks Visibility Valuation Study (Chestnut
ulThere are a number of Class I areas in each region that are not National Parks (e.g., Florida's Okefenokee
Wilderness Area), and are thus not included in the estimated value for visibility. We do not attempt to estimate
WTP for these other areas, and simply note that they are omitted.
VII-56
-------
Chapter VII: Benefit-Cost Analysis
and Rowe, 1990), Chestnut (1997, p. 10) estimated coefficients for the three study visibility
regions: California, Southwest, and Southeast. To account for national parks in the rest of the
contiguous U.S., however, the same coefficients are transferred to value visibility changes in
parks located in adjacent regions, termed here as "transfer regions." Table VII-14 displays the
"in-region" and "out-of-region" coefficients used in each of the different visibility regions.
Table VII-14. Estimated Coefficients Used in the Valuation of
WTP for Improved Visibility
Study Visibility Regions and Transfer
Visibility Regions *
California & Northwest
Southwest & Centralwest
Southeast & Northeast
Estimated /? for Out-of-Region
Households
$8.96
$13.51
$4.91
Estimated flfor In-Region
Households
$12.89
$16.82
$7.98
Transfer regions are groups of states adjacent to the study region from which WTP values are assigned.
The in-region coefficient estimates the WTP of residents within a given visibility region
for visibility improvements at all parks located within that same region. The out-of-region
coefficient estimates the WTP of residents living outside a given visibility region for visibility
improvements at all parks located within that region. The results of the survey suggest that in-
region residents are likely to value visibility improvements at their parks more than out-of-region
residents. This is consistent with expectations, as in-region households are more likely to visit,
know about, and care for these parks.
Because the WTP coefficients are for visibility improvements at more than one park
within a given visibility region, the WTP values must be apportioned between parks within a
given visibility region. Our analysis assumes that WTP for visibility is related to a park's
number of visitors. This is clearly a very crude approximation, since the WTP that we are
attempting to estimate includes both use and non-use values, and a visitation rate is a better
measure of use value and is not clearly linked to non-use values. On the other hand, short of
conducting a survey for individual parks, it is difficult to estimate the relative importance of
visibility at each park, and using a visitation rate to weight seems more appropriate than taking a
simple average or using some other weighting metric, such as the size of the park.
For each study visibility region, we sum 1997 visitor-days at each Class I park. We then
divide this total visitation figure into a WTP coefficient (in- or out-of-region, as appropriate) to
create a WTP per visitor-days for the entire study region. Multiplying this new value by each
park's own number of 1997 visitor-days yields an apportioned per-park WTP coefficient for each
park present in the study visibility region. Thus, we apply a visibility valuation function from a
VII-57
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
study region to an extrapolated, transfer region.
For aggregate benefits, the low-end estimate does not value residential visibility and uses
the lower-bound estimate for recreational visibility for each region. In the high set of valuation
assumptions, total visibility benefits consist of residential visibility benefits, as well as in- and
out-of-region recreational visibility benefits (using the WTP estimates reported by Chestnut
without adjustments to reflect the upper-bound of variation).
e.
Ozone- and PM-related Welfare Effect Benefits Estimation
Table VII-15 presents estimates of the monetary benefits arising from each of the welfare
endpoints associated with the air quality changes attributed to the Tier 2 rule.
Table VII-15. Welfare Endpoint Monetary Benefits
Endpoint
Agricultural crop damage
Nitrogen deposition
Household soiling damage
Pollutant
Ozone
NOX
PM
Monetary Benefits (millions 1997$)
Low
-1
200
60.1
High
301
200
60.1
Visibility
Out-of-region recreational
In-region recreational
Residential
PM and gases
PM and gases
PM and gases
266.33
64.10
not valued
266.33
64.10
371.02
5. Total Aggregated Benefits
In our analysis, we aggregated dollar benefits associated with each of the effects
examined, such as hospital admissions, into a total benefits estimate assuming that none of the
included health and welfare effects overlap. The point estimate of the total benefits associated
with the health and welfare effects in each set (low and high) is just the sum of the separate
effects estimates. The estimate of total benefits may be thought of as the end result of a
sequential process in which, at each step, the estimate of benefits from an additional source is
added. Each time an estimate of dollar benefits from a new source (e.g., a new health effect) is
added to the previous estimate of total dollar benefits, the estimated total dollar benefits
increases. The uncertainty surrounding the estimate of total dollar benefits, however, also
VII-58
-------
Chapter VII: Benefit-Cost Analysis
increases.
A significant portion of the uncertainty in the benefit estimate derives from uncertainty
about the true value of the coefficient in the C-R functions and the true dollar value of the
effects. The analysis relies on estimates of these parameters, but the true values being estimated
are unknown. This type of uncertainty can often be probabilistically quantified. For example,
the uncertainty about pollutant coefficients is typically quantified by reported standard errors of
the estimates of the coefficients in the C-R functions estimated by epidemiological studies. The
Technical Support Document for this analysis quantifies the uncertainty associated with each
health and welfare endpoint. Another important source of uncertainty derives from the discrete
set of assumptions used to select endpoints and concentration-response functions and to
determine inputs to the concentration-response functions. This type of uncertainty can be
quantified through the use of sensitivity analyses, but is not easily conveyed in probabilistic
terms.
6. Sensitivity Analyses
A portion of the uncertainty associated with benefits analysis involves discrete choices
between assumptions. We can not easily assign non-arbitrary probabilities to the alternative
assumptions, and instead we use a reasonable range of assumptions. Our analysis uses two sets
of assumptions that incorporate the following key assumptions:
(1) the choice of the PM threshold (15 |ig/m3,or background);
(2) the value placed on reduced mortality associated with PM (the value of a statistical life, or
the value of statistical life adjusted to reflect age-distributions of the affected population);
(3) the value placed on reduced incidence of chronic bronchitis;
(4) whether PM is associated with infant mortality;
(5) whether ozone is associated with the mortality of someone at any age;
(6) whether plantings of commodity crop cultivars are sensitive or insensitive to ozone; and
(7) the value placed on visibility benefits (both residential and recreational visibility).
Table VII-16 presents the estimates for the impacts and the associated economic value for
each set of assumptions. The results shown in the table demonstrate that selected alternative
assumptions drastically changes the total benefits that can be assumed for this rule. Actual
benefits are likely to be between the Low and High estimates provided.
VII-59
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table VII-16. Avoided Incidence and Monetized Benefits Associated with the
Tier 2 Rule for a Range of Assumption Sets
Endpoint
Avoided Incidence
(cases/year)
Low"
Highc
Monetary
(millions
Low
Benefits
1997$)
High
PM
Mortality (long-term exp. - ages 30+)
Mortality (long-term exp. - infants)
Chronic bronchitis
Hosp. Admissions - all respiratory (all ages)
Hosp. Admissions - congestive heart failure
Hosp. Admissions - ischemic heart disease
Acute bronchitis
Lower respiratory symptoms (LRS)
Upper respiratory symptoms (URS)
Work loss days (WLD)
Minor restricted activity days (MRAD)
Household soiling damage
832
-
3,885
504
127
146
984
19,782
3,093
233,000
1,856,000
2,416
10
3,914
836
138
159
4,072
37,437
3,387
415,000
3,370,000
-
2,275
-
281
4.6
1.5
2.2
0.1
0.3
0.1
23.8
87.7
60.1
14,256
56
1,354
7.6
1.7
2.4
0.2
0.5
0.1
42.3
159.3
60.1
Ozone
Mortality (short-term; four U.S. studies)
Hospital admissions - all respiratory (all ages)
Any of 19 acute symptoms
Decreased worker productivity
Agricultural crop damage
Visibility
Nitrogen Deposition
Total (PM + ozone + visibility + N deposition)
-
549
54,101
-
-
388
736
71,545
-
-
-
-
-
-
5.3
1.3
43.0
-1
330
200
3,315
2,312
7.1
1.7
60.4
301
701
200
19,525
" The low assumption set assumes effects from PM do not occur below concentrations of 15 ug/m3, that all mortality and chronic bronchitis
effects occur within the same year of the PM reduction (see section 7.a for a discussion of this uncertainty), utilizes the value of statistical life
year lost approach, ozone-related mortality and PM-related infant mortality are not included in the benefits estimate, chronic bronchitis valued
with the cost of illness approach, plantings of commodity crop cultivars are assumed to be insensitive to ozone, and does not value residential
visibility benefits.
0 The high assumption set assumes a PM threshold of background, utilizes the value of a statistical life approach, both ozone-related mortality
and PM-related mortality are included in the estimation of benefits, chronic bronchitis valued with a willingness-to-pay approach, plantings of
commodity crop cultivars are assumed to be sensitive to ozone, and full accounting for recreational and residential visibility benefits.
VII-60
-------
Chapter VII: Benefit-Cost Analysis
7. Limitations of the Analysis
Given incomplete information, this national benefits analysis yields approximate results
because of the uncertainty associated with any estimate. Potentially important sources of
uncertainty exist and many of these are summarized in Table VII-17. These uncertainties can
cause the total benefits estimate to be understated or overstated. Where possible, we state the
direction of the bias presented by the uncertainty. However, in most cases the effect of the
uncertainty on total benefits is unknown (i.e., it could increase or decrease benefits depending on
specific conditions). The remainder of this Section provides a discussion of four broad areas of
uncertainty.
VII-61
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table VII-17. Sources of Uncertainty in the Benefit Analysis
1. Uncertainties Associated With Concentration-Response Functions
-The value of the ozone- or PM-coefficient in each C-R function.
-Application of a single C-R function to pollutant changes and populations in all locations.
-Similarity of future year C-R relationships to current C-R relationships.
-Correct functional form of each C-R relationship, (e.g., It is uncertain whether there are thresholds and, if so, what
they are.)
-Extrapolation of C-R relationships beyond the range of ozone or PM concentrations observed in the study.
2. Uncertainties Associated With Ozone and PM Concentrations
-Estimating future-year baseline and hourly ozone and daily PM concentrations.
-Estimating the change in ozone and PM resulting from the control policy.
3. Uncertainties Associated with PM Mortality Risk
-No scientific basis supporting a plausible biological mechanism.
-Potential causal agents within the complex mixture of PM responsible for the reported adverse health effects have not
been identified.
-While there were a great number of studies associated with PM10, there were a limited number of studies that directly
measured PM2 5.
-The extent to which adverse health effects are associated with low level exposures that occur many times in the year
versus peak exposures.
-Estimated health effects levels associated with PM2 5 exposure were small.
-Possible confounding in the epidemiological studies of PM25, effects with other factors (e.g., other air pollutants,
weather, indoor/outdoor air, etc.).
-The extent to which effects reported in the long-term studies are associated with historically higher levels of PM rather
than the levels occurring during the period of study.
-Reliability of the limited ambient PM25 monitoring data in reflecting actual PM25 exposures.
4. Uncertainties Associated With Possible Lagged Effects
-What portion of the PM-related long-term exposure mortality effects associated with changes in annual PM levels
would occur in a single year, and what portion might occur in subsequent years. Ignoring lags may lead to an
overestimate of benefits.
5. Uncertainties Associated With Baseline Incidence Rates
-Some baseline incidence rates are not location-specific (e.g., those taken from studies) and may therefore not
accurately represent the actual location-specific rates.
-Current baseline incidence rates may not well approximate what baseline incidence rates will be in the year 2007.
-Projected population and demographics — used to derive incidences - may not well approximate future-year
nonulation and demoeranhics.
J. J. \J\ V^^/LV^\_4. iJ\J ij LJ.J.C41~L\J J. L UL1\J. \-4-V^J.J.
population and demographics.
6. Uncertainties Associated With Economic Valuation
-Unit dollar values associated with health and welfare endpoints are only estimates of mean WTP and therefore have
uncertainty surrounding them.
-Mean WTP (in constant dollars) for each type of risk reduction may differ from current estimates due to differences in
income or other factors.
7. Uncertainties Associated With Aggregation of Monetized Benefits
-Health and welfare benefits estimates are limited to the available C-R functions. Thus, unquantified benefit categories
VII-62
-------
Chapter VII: Benefit-Cost Analysis
will cause total benefits to be under
a. PM Mortality Risk and Health Effects
Table VII-20 summarizes a number of the uncertainties associated with estimating
lortality risk associated with particulate matter (PM). Most of these uncertainties can serve to
ncrease or decrease the estimated benefits relative to a hypothetical "true" prediction. Some
ncertainties may inflate estimates, while others - such as exclusion of effects categories - can
esult in understatement. The fundamental concentration-response relationships used to estimate
enefits are derived from epidemiological studies of community health. Based on these studies
nd other available information, the EPA Criteria Document concluded that the observed
ssociations between paniculate matter and mortality and other serious health effects were
likely causal." The Criteria Document also noted that, as yet, the scientific information did not
rovide a basis for determining what biological mechanisms might account for such effects. To
extent that some chance remains that no causal mechanisms are found for some PM
omponents or for the PM mix taken as a whole, the benefit estimates derived from the
pidemiological studies would be overstated.
Similarly, the evaluation of the epidemiological evidence included an extensive
ssessment of a number of potential pollutant and weather confounders or effects modifiers. The
Criteria Document concluded that these factors could not fully account for the observed
M/effects associations, but it is possible that some portion of the quantitative relationships are
ffected by the presence of other pollutants. While multiple pollutant effects may be additive, it
3 also possible that the PM related effects association may be overstated for some studies, whict
light inflate the benefits estimates derived from such studies.
In addition, following the recommendation of the Advisory Council on Clean Air
Compliance analyses (an SAB advisory committee established to review methodology for the
12 study), the PM mortality benefit estimates have been derived from a single study that likely
ncompasses both short-and long-term mortality effects (Pope et al. 1995). Similarly, the
gency has used a single study (Schwartz 1993) in its estimates of the benefits of reduced cases
f chronic bronchitis. The approach used in both cases assumes that the benefits of the PM
eductions will occur within a year of the reductions. Because some fraction of the estimated
lortality or chronic bronchitis effects may well be associated with multi-year exposures, the
enefits of a given reduction in concentrations in one year will not all be realized in that year. Tc
ate, however, the available studies have not developed any estimates of the relative proportion
f near term as compared to the potential "lagged" consequences of PM reductions (HEES,
999).
Some analysts believe, however, that this analysis should provide an estimate that reflect!
le potential effect of considering such lagged effects in presenting the range of estimated
onofits.—For oxamplo, if ono woro to assume that realization of tho full health benefits from—
VII-63
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
reductions in particulate matter resulting from this rule might take up to 5 years, the estimated
monetized benefits for reductions in premature mortality and chronic bronchitis would be
reduced by $204 million at the low end of the range of total benefits (see Table VII-18 below).
Table VII-18. PM Health Effects and Benefits (No Lag and Lag of up to Five Years)
Health Effects
Chronic Bronchitis
Mortality
Total
Benefits (No Lag)
(millions 1997$)
$281
$2,278
$2,559
Benefits (Lag of up to 5 years)"-"
(millions 1997$)
$259
$2,096
$2,355
As discussed above, SAB has concluded that selection of a value for such a lag at this
time would be arbitrary and inclusion of pollutant-related time lags in mortality is premature
(HEES, 1999). For this reason, we have not incorporated lags into this analysis. The Agency is
committed to working with the SAB and others during the development of the final rule to look
at how to address this issue in the benefits range for both the Tier 2 final rule and RIA and in
future regulatory analyses.
b. Unquantifiable Benefits
In considering the monetized benefits estimates, the reader should be aware that many
limitations for conducting these analyses are mentioned throughout this RIA. One significant
limitation of both the health and welfare benefits analyses is the inability to quantify many PM
and ozone-induced adverse effects listed in Table VII-6. In general, if it were possible to include
the unquantified benefits categories in the total monetized benefits, the benefits estimates
presented in this RIA would increase. Specific examples of unquantified benefits explored in
more detail below include other human health effects, urban ornamental plants, aesthetic injury
to forests, nitrogen in drinking water, and brown clouds.
The benefits of reductions in a number of ozone- and PM-induced health effects have not
been quantified due to the paucity of C-R and/or economic valuation data. These effects include:
reduced pulmonary function, morphological changes, altered host defense mechanisms, cancer,
other chronic respiratory diseases, infant mortality, airway responsiveness, increased
susceptibility to respiratory infection, pulmonary inflammation, acute inflammation and
•'•"This approach assumes that 25 percent of the reductions in health effects reduction are realized in year 1,
25 percent in year 2, 16.67 percent in year 3, 16.67 in year 4 and 16.67 in year 5. This is an illustrative example
only and does not represent any known lag structure for these health effects.
VII-64
-------
Chapter VII: Benefit-Cost Analysis
respiratory cell damage, and premature aging of the lungs.
In addition to the above non-monetized health benefits, there are a number of non-
monetized welfare benefits including: reduced adverse effects on vegetation, forests, and other
natural ecosystems. The CAA and other statutes, through requirements to protect natural and
ecological systems, indicate that these are scarce and highly valued resources. Lack of
comprehensive information, insufficient valuation tools, and significant uncertainties therefore
result in understated welfare benefits estimates in this RIA. However, a number of expert
biologists, ecologists, and economists (Costanza et al., 1997) argue that the benefits of protecting
natural resources are enormous and increasing as ecosystems become more stressed and scarce in
the future. Additionally, agricultural, forest and ecological scientists (Heck and Cowling, 1997)
believe that vegetation appears to be more sensitive to ozone than are humans and consequently,
that damage is occurring to vegetation and natural resources at concentrations below the ozone
NAAQS. Experts also believe that the effect of ozone on plants is both cumulative and long-
term. The specific non-monetized benefits from reductions in ambient ozone concentrations
would accrue from: decreased foliar injury; averted growth reduction of trees in natural forests;
maintained integrity of forest ecosystems (including habitat for native animal species); and the
aesthetics and utility of urban ornamentals (e.g., grass, flowers, shrubs and trees). Other welfare
categories for which there is incomplete information to estimate the economic value of reduced
adverse effects include: materials damage; and reduced sulfate deposition to aquatic and
terrestrial ecosystems.
Other Human Health Effects
Human exposure to PM and ozone is known to cause health effects such as: impaired
airway responsiveness, increased susceptibility to respiratory infection, acute inflammation and
respiratory cell damage, premature aging of the lungs and chronic respiratory damage. An
improvement in ambient PM and ozone air quality is expected to reduce the number of
incidences within each effect category that the U.S. population would experience. Although
these health effects are known to be PM or ozone-induced, C-R data is not available for
quantifying the benefits associated with reducing these effects. The inability to quantify these
effects leads to an underestimation of the monetized benefits presented in this analysis.
Urban Ornamentals
Urban ornamentals represent an additional vegetation category likely to experience some
degree of effects associated with exposure to ambient ozone levels and likely to impact large
economic sectors. In the absence of adequate exposure-response functions and economic
damage functions for the potential range of effects relevant to these types of vegetation, no direct
quantitative economic benefits analysis has been conducted. It is estimated that more than $20
billion (1990 dollars) are spent annually on landscaping using ornamentals (Abt Associates,
1995), both by private property owners/tenants and by governmental units responsible for public
areas, making this a potentially important welfare effects category. However, information and
VII-65
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
valuation methods are not available to allow for plausible estimates of the percentage of these
expenditures that may be related to impacts associated with ozone exposure.
Commercial Forests
Any attempt to estimate economic benefits for commercial forests associated with
reductions in ozone arising from implementation of the Tier 2 rule is constrained by a lack of
exposure-response functions for the commercially important mature trees. Although exposure-
response functions have been developed for seedlings for a number of important tree species,
these seedling functions cannot be extrapolated to mature trees based on current knowledge.
Recognizing this limitation, a study (de Steiger et al., 1990; Pye et al., 1988) involving expert
judgment about the effect of ozone levels on percent growth change has been used to develop
estimates of ozone-related economic losses for commercial forest products. Our analysis,
however, did not quantify benefits from improved production within commercial forests.
Aesthetic Injury to Forests
Ozone is a regionally dispersed air pollutant that has been shown conclusively to cause
discernible injury to forest trees (Fox and Mickler, 1996). One of the welfare benefits expected
to accrue as a result of reductions in ambient ozone concentrations in the United States is the
economic value the public receives from reduced aesthetic injury to forests. There is sufficient
scientific information available that ambient ozone levels cause visible injury to foliage and
impair the growth of some sensitive plant species (U.S. EPA, 1996c, p. 5-521). However,
present analytic tools and resources preclude EPA from quantifying the benefits of improved
forest aesthetics.
Nitrates in Drinking Water
Nitrates in drinking water are currently regulated by a maximum contaminant level
(MCL) of 10 mg/L on the basis of the risk to infants of methemoglobinemia, a condition which
adversely affects the blood's oxygen carrying capacity. In an analysis of pre-1991 data, Raucher
et al.(1993) found that approximately 2 million people were consuming public drinking water
supplies which exceed the MCL. Supplementing these findings, the National Research Council
concluded that 42 percent of the public drinking water users in the U.S. (approximately 105
million people) are either not exposed to nitrates or are exposed to concentrations below 1.3
mg/L (National Research Council 1995).
In a recent epidemiological study by the National Cancer Institute, a statistically
significant relationship between nitrates in drinking water and incidence of non-Hodgkin's
lymphoma were reported (Ward et al., 1996). Though it is generally acknowledged that
traditional water pollution sources such as agricultural runoff are mostly responsible for
violations of the MCL, other more diffuse sources of nitrate to drinking water supplies, such as
that from atmospheric deposition, may also become an important health concern should the
VII-66
-------
Chapter VII: Benefit-Cost Analysis
cancer link to nitrates be found valid upon further study.
Other Unqualified Benefits Categories
There are other welfare benefits categories for which there is incomplete information to
permit a quantitative assessment for this analysis. For some endpoints, gaps exist in the
scientific literature or key analytical components and thus do not support an estimation of
incidence. In other cases, there is insufficient economic information to allow estimation of the
economic value of adverse effects. Potentially significant, but unquantified welfare benefits
categories include: existence and user values related to the protection of Class I areas (e.g.,
Shenendoah National Park), damage to tree seedlings of more than 10 sensitive species (e.g.,
black cherry, aspen, ponderosa pine), non-commercial forests, ecosystems, materials damage, and
reduced sulfate deposition to aquatic and terrestrial ecosystems. Although scientific and
economic data are not available to allow quantification of the effect of ozone in these categories,
the expectation is that, if quantified, each of these categories would lead to an increase in the
monetized benefits presented in this RIA.
c. Potential Disbenefits
In this discussion of unquantified benefits, a discussion of potential disbenefits must also
be mentioned. Several of these disbenefit categories are related to nitrogen deposition, while one
category is related to the issue of ultraviolet light. Because EPA is not able to quantify these
disbenefit categories, total benefits will be overstated.
Passive Fertilization
Several disbenefit categories are related to nitrogen deposition. Nutrients deposited on
crops from atmospheric sources are often referred to as passive fertilization. Nitrogen is a
fundamental nutrient for primary production in both managed and un-managed ecosystems.
Most productive agricultural systems require external sources of nitrogen in order to satisfy
nutrient requirements. Nitrogen uptake by crops varies, but typical requirements for wheat and
corn are approximately 150 kg/ha/yr and 300 kg/ha/yr, respectively (NAPAP, 1990). These rates
compare to estimated rates of passive nitrogen fertilization in the range of 0 to 5.5 kg/ha/yr
(NAPAP, 1991). So, for these crops, deposited nitrogen could account for as much as two to
four percent of nitrogen needs. Holding all other factors constant, farmers' use of purchased
fertilizers or manure may increase as deposited nitrogen is reduced. EPA has not estimated the
potential value of this possible increase in the use of purchased fertilizers, but it is likely that the
overall value is very small relative to the value of other health and welfare endpoints presented in
this analysis. First, reductions in NOX emissions affect only a fraction of total nitrogen
deposition. Approximately 70 to 80 percent of nitrogen deposition is in the form of nitrates (and
thus can be traced to NOX emissions) while most of the remainder is due to ammonia emissions
(Dennis, 1997). The annual average change in nitrogen deposition attributable to the Tier 2 rule
VII-67
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
is about 11 percent of baseline levels, suggesting a relatively small potential change in passive
fertilization. Second, some sources of nitrogen, such as animal manure, are available at no cost
or at a much lower cost than purchased nitrogen. In addition, in certain areas nitrogen is
currently applied at rates which exceed crop uptake rates, usually due to an overabundance of
available nutrients from animal waste. Small reductions in passive fertilization in these areas is
not likely to have any consequence to fertilizer application. The combination of these factors
suggests that the cost associated with compensating for reductions in passive fertilization is
relatively minor.
Information on the effects of changes in passive nitrogen deposition on forests and other
terrestrial ecosystems is very limited. The multiplicity of factors affecting forests, including other
potential stressors such as ozone, and limiting factors such as moisture and other nutrients,
confound assessments of marginal changes in any one stressor or nutrient in forest ecosystems.
However, reductions in deposition of nitrogen could have negative effects on forest and
vegetation growth in ecosystems where nitrogen is a limiting factor (U.S. EPA, 1993).
On the other hand, there is evidence that forest ecosystems in some areas of the United
States are nitrogen saturated (U.S. EPA, 1993). Once saturation is reached, adverse effects of
additional nitrogen begin to occur such as soil acidification which can lead to leaching of
nutrients needed for plant growth and mobilization of harmful elements such as aluminum.
Increased soil acidification is also linked to higher amounts of acidic runoff to streams and lakes
and leaching of harmful elements into aquatic ecosystems.
Ultraviolet Light
A reduction of tropospheric ozone is likely to increase the penetration of ultraviolet light,
specifically UV-b, to ground level. UV-b is an issue of concern because depletion of the
stratospheric ozone layer (i.e., ozone in the upper atmosphere) due to chlorofluorocarbons and
other ozone-depleting chemicals is associated with increased skin cancer and cataract rates.
Currently, EPA is not able to adequately quantify these effects for the purpose of valuing benefits
for this policy.
Other EPA programs exist to address the risks posed by changes in UV-b associated with
changes in total column ozone. As presented in the Stratospheric Ozone RIA (U.S. EPA, 1992),
stratospheric ozone levels are expected to significantly improve over the next century as the
major ozone depleting substances are phased out globally. This expected improvement in
stratospheric ozone levels is estimated to reduce the number of non-melanoma skin cancers by
millions of cases in the U.S. by 2075.
d. Projected Income Growth
Our analysis does not attempt to adjust benefits estimates to reflect expected growth in
VII-68
-------
Chapter VII: Benefit-Cost Analysis
real income. Economic theory argues, however, that WTP for most goods (such as
environmental protection) will increase if real incomes increase. The degree to which WTP may
increase for the specific health and welfare benefits provided by the Tier 2 rule cannot be
estimated due to insufficient income elasticity information.
D. Cost
Since the benefits assessment has been performed on the basis of a fully turned over fleet
of tier 2 vehicles, consistent costs were developed by using the same basis. Costs to be compared
to the monetized value of the benefits were therefore developed for a fleet the size of the year
2010 fleet. For this purpose we used the long term cost once the capital costs have been
recovered and the manufacturing learning curve reductions have been realized, since this most
closely represents the makeup of a fully turned-over fleet.
This analysis also made adjustments in the costs to account for the fact that there is a time
difference between when some of the costs are expended and when the benefits are realized. The
vehicle costs are expended when the vehicle is sold, while the fuel related costs and the benefits
are distributed over the life of the vehicle.
We resolved this difference by using costs distributed over time such that there is a
constant cost per ton of emissions reduction and such that the net present value of these
distributed costs corresponds to the net present value of the actual costs. A constant ratio of cost
to emission reduction over the life of the vehicle would also reflect itself in the ratio of the net
present value of the costs and net present value of the emission reductions. This, of course, is
how EPA determined the cost effectiveness estimates for the proposed rule. Thus, the simplest
way to develop this distributed cost number is simply to multiply the cost effectiveness ratio
(dollars per ton) times the emission reduction estimates for the benefits assessment.
The resulting adjusted costs are somewhat greater than the actual annual cost of the
program, reflecting the time value adjustment. Thus, both because of the assumption of a fully
turned over fleet and because of the time value adjustment, the costs presented in this section do
not represent actual annual costs of the Tier 2/gasoline sulfur program for 2010. Rather, they
represent an approximation of the steady-state cost per ton that would likely prevail in 2015 and
beyond. The benefit cost ratio for the earlier years of the program would be expected to be lower
than that based on these costs, since the fleet-adjusted costs are larger in the early years of the
program while the benefits are smaller.
Since the long term costs are not representative of the per vehicle costs in the early phases
of the program, we also estimated an adjusted cost based on the near term cost effectiveness
value. Using the near term cost effectiveness value of $2134/per ton, the adjusted cost would be
$4.3 billion. While no actual in-use fleet could consist entirely of vehicles experiencing this near
term cost, this value does present an upper bound on the cost figure.
VII-69
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
The resulting adjusted cost values are given in Table VII-19.
Table VII -19. Adjusted Cost for Comparison to Benefits
Cost Basis
Long term
Cost per ton ratio
1748
Tons ofNOx + NMHC
2,003,761
Adjusted Cost
(billions of dollars)
3.5
VII-70
-------
Chapter VII: Benefit-Cost Analysis
Chapter VII References
Abbey, D.E., F. Petersen, P.K. Mills and W.L. Beeson, 1993. "Long-Term Ambient
Concentrations of Total Suspended Particulates, Ozone, and Sulfur Dioxide and Respiratory
Symptoms in a Nonsmoking Population." Archives of Environmental Health, 48(1): 33-46.
Abbey, D.E., B.E. Ostro, F. Petersen and RJ. Burchette, 1995. "Chronic Respiratory Symptoms
Associated with Estimated Long-Term Ambient Concentrations of Fine Parti culates Less Than
2.5 Microns in Aerodynamic Diameter (PM2.5) and Other Air Pollutants." Journal of Exposure
Analysis and Environmental Epidemiology, 5(2): 137-159.
Abt Associates Inc., 1992. The Medical Costs of five Illnesses Related to Exposure to
Pollutants. Prepared for U.S. EPA, Office of Pollution Prevention and Toxics. Washington,
DC. See EPA Air Docket A-96-56, Document No. VI-B-09-(n).
Abt Associates Inc., 1995. Urban Ornamental Plants: Sensitivity to Ozone and Potential
Economic Losses. Prepared for the U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards. Research Triangle Park, NC. July.
Abt Associates Inc., 1998. Air Quality Estimation for the NOx SIP Call RIA. Prepared for U.S.
EPA, Office of Air Quality Planning and Standards, under contract no. 68-D-98-001. Research
Triangle Park, NC. September. See EPA Air Docket A-96-56, Document No. VI-B-09-(gggg)
Abt Associates Inc., 1999. Tier 2 Proposed Rule: Air Quality Estimation, Selected Health and
Welfare Benefits Methods, and Benefit Analysis Results. Prepared for U.S. EPA, Office of Air
Quality Planning and Standards. Research Triangle Park, NC. February.
Butler, R.J., 1983. "Wage and Injury Rate Responses to Shifting Levels of Workers'
Compensation." In: Safety and the Work Force. Worrall, J.D., Ed. Cornell University, ILR
Press, Ithaca, NY.
Chang, J.S., R.A. Brost, I.S.A. Isaksen, S. Madronich, P. Middleton, W.R. Stockwell and CJ.
Walcek, 1987. "A Three-Dimensional Eulerian Acid Deposition Model: Physical Concepts and
Formulation." Journal of Geophysical Research, 92, 14,681-14,700.
Chang, J.S., P.B. Middleton, W.R. Stockwell, CJ. Walcek, I.E. Pleim, H.H. Lansford, F.S.
Binkowski, S. Madronich, N.L. Seaman and D.R. Stauffer, 1990. "The Regional Acid
Deposition Model and Engineering Model. NAPAP SOS/T Report 4." In: Acidic Deposition:
State of Science and Technology, Volume 1. National Acid Precipitation Assessment Program,
Washington DC. December.
Chestnut, L.G., 1997. "Draft Memorandum: Methodology for Estimating Values for Changes in
Visibility at National Parks." April 15. See EPA Air Docket A-96-56, Document No. VI-B-09-(ooo).
VII-71
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Chestnut, L.G. and R.D. Rowe, 1990. Preservation Values for Visibility Protection at the
National Parks: Draft Final Report. Prepared for U.S. Environmental Protection Agency, Office
of Air Quality Planning and Standards, Economic Analysis Branch. Research Triangle, NC.
February, 16.
Costanza, R., R. d'Arge, R. de Groot, S. Farber, M. Grasso, B. Hannon, K. Limburg, S. Naeem,
R.V. O'Neill, J. Paruelo, R.G. Raskin, P. Sutton and M. van den Belt, 1997. "The Value of the
World's Ecosystem Services and Natural Capital." Nature, 387, 253-259.
Cousineau, J., R. Lacroix and A. Girard, 1988. Occupational Hazard and Wage Compensating
Differentials. University of Montreal Working Paper.
Crocker, T.D. and R.L. Horst, Jr, 1981. "Hours of Work, Labor Productivity, and
Environmental Conditions: A Case Study." The Review of Economics and Statistics, 63, 361-368.
Cummings, R., H. Burness and R. Norton, 1981. Methods Development for Environmental
Control Benefits Assessment, Volume V. Measuring Household Soiling Damages from
Suspended Air Particulates, A Methodological Inquiry. Prepared for U. S. Environmental
Protection Agency. Washington, DC.
de Steiger, I.E., J.M. Pye and C.S. Love, 1990. " Air Pollution Damage to U.S. Forests."
Journal of Forestry, 88(8), 17-22.
Dennis, R. 1997. Personal communication. NOAA Atmospheric Research Lab, Research
Triangle Park, NC.
Dennis, R.L., 1997. "Using the Regional Acid Deposition Model to Determine the Nitrogen
Deposition Airshed of the Chesapeake Bay Watershed." In: Atmospheric Deposition to the Great
Lakes and Coastal Waters. Baker, I.E., Ed. Society of Environmental Toxicology and
Chemistry, Pensacola, FL. pp. 393-413. See EPA Air Docket A-96-56, Document No. VI-B-
09-(ccc).
Dennis, R.L., W.R. Barchet, T.L. Clark and S.K. Seilkop, 1990. "Evaluation of Regional Acid
Deposition Models (Part I). NAPAP SOS/T Report 5." In: Acidic Deposition: State of Science
and Technology, Volume 1. National Acid Precipitation Assessment Program, Washington, DC.
September.
Dillingham, A., 1985. "The Influence of Risk Variable Definition on Value of Life Estimates."
Economic Inquiry, 24, 277-294.
Dockery, D.W., C.A. Pope, X.P. Xu, J.D. Spengler, J.H. Ware, M.E. Fay, E.G. Ferris and F.E.
Speizer, 1993. "An Association Between Air Pollution and Mortality in 6 United-States Cities."
New England Journal of Medicine, 329(24), 1753-1759.
VII-72
-------
Chapter VII: Benefit-Cost Analysis
Dockery, D.W., F.E. Speizer, D.O. Stram, J.H. Ware, J.D. Spengler and E.G. Ferris, Jr., 1989.
"Effects of Inhalable Particles on Respiratory Health of Children." American Review of
Respiratory Disease, 139, 587-594.
Douglas, S.G. and R.K. Iwamiya, 1999. Estimating the Effects of the Tier 2 Motor-Vehicle
Standards on Air Quality: Ozone. Prepared for Abt Associates Inc. Prepared by Systems
Applications International, Inc. SYSAPP-99-98/50. January.
E.H. Pechan & Associates, I, 1996. Regional Paniculate Control Strategies Phase II.
Prepared for the U.S. Environmental Protection Agency, Office of Policy, Planning, and
Evaluation. Washington, DC. September.
E.H. Pechan, 1999. Emissions and Air Quality Impacts of Proposed Motor Vehicle Tier 2 and
Fuel Sulfur Standards. Prepared by The Pechan-Avanti Group under EPA Contract No. 68-D9-
8052. Prepared for U.S. EPA, Office of Air Quality Planning and Standards, Innovative
Strategies and Economics Group. Research Triangle Park, NC. January.
Empire State Electric Energy Research Corporation (ESEERCO), 1994. New York State
Environmental Externalities Cost Study. Report 2: Methodology. Prepared by RCG/Hagler,
Bailly, Inc. November.
Fox, S. and R.A. Mickler, Eds. 1996. Impact of Air Pollutants on Southern Pine Forests.
Ecological Studies. Vol.118. Springer-Verlag, New York.
Gar en, J., 1988. "Compensating Wage Differentials and the Endogeneity of Job Riskiness." The
Review of Economics and Statistics, 70(1), 9-16.
Gegax, D., S. Gerking and W. Shulze, 1985. Perceived Risk and the Marginal Value of Safety.
Working paper prepared for the U. S. Environmental Protection Agency.
Gerking, S., M. DeHaan and W. Schulze, 1988. "The Marginal Value of Job Safety: A
Contingent Valuation Study." Journal of Risk and Uncertainty, 1, 185-199.
Heck, W.W. and E.B. Cowling, 1997. "The Need for a Long Term Cumulative Secondary
Ozone Standard—An Ecological Perspective." EM, January, 23-33.
Herzog, H.W., Jr., and A.M. Schlottmann, 1987. Valuing Risk in the Workplace: Market Price,
Willingness to Pay, and the Optimal Provision of Safety. University of Tennessee Working
Paper.
Industrial Economics Incorporated (lEc), 1993. Memorandum to Jim DeMocker, U.S.
Environmental Protection Agency, Office of Air and Radiation, Office of Policy Analysis and
Review. September 30. See EPA Air Docket A-96-56, Document No. VI-B-09-(111).
VII-73
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Industrial Economics Incorporated (lEc), 1994. Memorandum to Jim DeMocker, U.S.
Environmental Protection Agency, Office of Air and Radiation, Office of Policy Analysis and
Review. March 31. See EPA Air Docket A-96-56, Document No. VI-B-09-(nnn).
Ito, K. and G.D. Thurston, 1996. "Daily Pm(10)/Mortality Associations - an Investigation of
At-Risk Subpopulations." Journal of Exposure Analysis and Environmental Epidemiology, 6(1),
79-95.
Jones-Lee, M.W. 1989. The Economics of Safety and Physical Risk. Basil Blackwell, Oxford.
Kinney, P.L., K. Ito and G.D. Thurston, 1995. "A Sensitivity Analysis of Mortality Pm-10
Associations in Los Angeles." Inhalation Toxicology, 7(1), 59-69.
Kniesner, T.J. and J.D. Leeth, 1991. "Compensating Wage Differentials for Fatal Injury Risk in
Australia, Japan, and the United States." Journal of Risk and Uncertainty, 4(1), 75-90.
Korotney, D. 1998. "Tons reduced values for 812 inventory adjustment." Email communication,
U.S. Environmental Protection Agency, Office of Mobile Sources, Ann Arbor, MI. November 9.
Krupnick, A.J. and M.L. Cropper, 1992. "The Effect of Information On Health Risk
Valuations." Journal of Risk and Uncertainty, 5(1), 29-48.
Krupnick, A.J., W. Harrington and B. Ostro, 1990. "Ambient Ozone and Acute Health Effects -
Evidence From Daily Data." Journal of Environmental Economics and Management, 18(1), 1-18.
Krupnick, A.J. and R.J. Kopp, 1988. The Health and Agricultural Benefits of Reductions in
Ambient Ozone in the United States. Resources for the Future. Washington, DC. Discussion
Paper QE88-10. August.
Latimer and Associates, 1996. Paniculate Matter Source - Receptor Relationships Between All
Point and Area Sources in the United States and PSD Class/Area Receptors. Prepared for Bruce
Polkowsky, Office of Air Quality Planning and Standards, U.S. EPA. Research Triangle Park,
NC. September.
Lee, E.H. and W.E. Hogsett, 1996. Methodology for Calculating Inputs for Ozone Secondary
Standard Benefits Analysis: Part II. Prepared for U.S. EPA, Office of Air Quality Planning and
Standards. March. See EPA Air Docket A-96-56, Document No. VI-B-09-(f).
Leigh, J.P., 1987. "Gender, Firm Size, Industry and Estimates of the Value-of-Life." Journal of
Health Economics, 6,255-273.
Leigh, J.P. and R.N. Folsom, 1984. "Estimates of the Value of Accident Avoidance at the Job
Depend on Concavity of the Equalizing Differences Curve." The Quarterly Review of
VII-74
-------
Chapter VII: Benefit-Cost Analysis
Economics and Business, 24(1), 56-66.
Manuel, E.H., R.L. Horst, K.M. Brennan, W.N. Lanen, M.C. Duff and J.K. Tapiero, 1982.
Benefits Analysis of Alternative Secondary National Ambient Air Quality Standards for Sulfur
Dioxide and Total SuspendedParticulates, Volumes I-IV. Prepared for U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards. Research Triangle Park.
Marin, A. and G. Psacharopoulos, 1982. "The Reward for Risk in the Labor Market: Evidence
from the United Kingdom and a Reconciliation with Other Studies." Journal of Political
Economy, 90(4), 827-853.
McClelland, G., W. Schulze, D. Waldman, J. Irwin, D. Schenk, T. Stewart, L. Deck and M.
Thayer, 1991. Valuing Eastern Visibility: A Field Test of the Contingent Valuation Method.
Prepared for U.S. Environmental Protection Agency, Office of Policy, Planning and Evaluation.
June. See EPA Air Docket A-96-56, Document No. VI-B-09-(ppp).
Miller, T. and J. Guria, 1991. The Value of Statistical Life in New Zealand. Report to the New
Zealand Ministry of Transport, Land Transport Division.
Moolgavkar, S.H., E.G. Luebeck, T.A. Hall and E.L. Anderson, 1995. "Air Pollution and Daily
Mortality in Philadelphia." Epidemiology., 6(5), 476-484.
Moore, MJ. and W.K. Viscusi, 1988. "Doubling the Estimated Value of Life: Results Using
New Occupational Fatality Data." Journal of Policy Analysis and Management, 7(3), 476-490.
NAPAP, 1990. Acidic Deposition: State of Science and Technology, Report 18: Response of
Vegetation to Atmospheric Deposition and Air Pollution. National Acid Precipitation
Assessment Program, Office of Director. Washington, DC. NTIS document No. PB92-
100544INZ (order at 1-800-553-6847).
NAPAP, 1991. National Acid Precipitation Assessment Program: 1990 Integrated Assessment
Report. National Acid Precipitation Assessment Program, Office of Director. Washington, DC.
NTIS document No. PB92-100346INZ (order at 1-800-553-6847).
National Research Council. Subcommittee on Nitrate and Nitrite in Drinking Water. 1995.
Nitrate and Nitrite in Drinking Water. National Academy Press, Washington, DC. NTIS
document No. PB95-267092INZ (order at 1-800-553-6847).
Olson, C.A., 1981. "An Analysis of Wage Differentials Received by Workers on Dangerous
Jobs." Journal of Human Resources, 16, 167-185.
Ostro, B.D., 1987. "Air Pollution and Morbidity Revisited: A Specification Test." Journal of
Environmental Economics and Management, 14, 87-98.
VII-75
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Ostro, B.D. and S. Rothschild, 1989. "Air Pollution and Acute Respiratory Morbidity - an
Observational Study of Multiple Pollutants." Environmental Research, 50(2), 238-247.
Pope, C.A., D.W. Dockery, J.D. Spengler and M.E. Raizenne, 1991. "Respiratory Health and
PmlO Pollution - a Daily Time Series Analysis." American Review of Respiratory Disease,
144(3), 668-674.
Pope, C.A., MJ. Thun, M.M. Namboodiri, D.W. Dockery, J.S. Evans, F.E. Speizer and C.W.
Heath, 1995. "Particulate Air Pollution As a Predictor of Mortality in a Prospective Study of Us
Adults." American Journal of Respiratory and Critical Care Medicine, 151 (3), 669-674.
Pye, J.M., I.E. de Steiguer and C. Love, 1988. Expert Opinion Survey on the Impact of Air
Pollutants on Forests of the USA. Proceedings of Air Pollution and Forest Decline, Interlaken,
Switzerland. . October. See EPA Air Docket A-96-56, Document No. VI-B-09-(e).
Raucher, R.S., J.A. Drago, E. Trabka, A. Dixon, A. Patterson, C. Lang, L. Bird and S. Ragland,
1993. An Evaluation of the Federal Drinking Water Regulatory Program Under the Safe
Drinking Water Act as Amended in 1986. Appendix A: Contaminant-Specific Summaries.
Prepared for the American Water Works Association.
Samet, J.M., 1997. Paniculate Air Pollution and Daily Mortality: Analysis of the Effects of
Weather and Multiple Air Pollutants. The Phase IB Report of the Particle Epidemiology
Evaluation Project. Health Effects Institute. March. Available through
http://www.healtheffects.org/pubform.htm
Schwartz, J., 1993. "Particulate Air Pollution and Chronic Respiratory Disease."
Environmental Research, 62, 7-13.
Schwartz, J. and R. Morris, 1995. "Air Pollution and Hospital Admissions For Cardiovascular
Disease in Detroit, Michigan." American Journal of Epidemiology, 142(1): 23-35.
Schwartz, J., D.W. Dockery and L.M. Neas, 1996. "Is Daily Mortality Associated Specifically
With Fine Particles." Journal of the Air & Waste Management Association, 46(10), 927-939.
Schwartz, J., D.W. Dockery, L.M. Neas, D. Wypij, J.H. Ware, J.D. Spengler, P. Koutrakis, F.E.
Speizer and E.G. Ferris, 1994. "Acute Effects of Summer Air Pollution On Respiratory
Symptom Reporting in Children." American Journal of Respiratory and Critical Care Medicine,
150(5), 1234-1242.
Schwartz, J. and R. Morris, 1995. "Air Pollution and Hospital Admissions For Cardiovascular
Disease in Detroit, Michigan." American Journal of Epidemiology, 142(1), 23-35.
Sisler, J.F., 1996. Spatial and Seasonal Patterns and Long Term Variability of the Composition
VII-76
-------
Chapter VII: Benefit-Cost Analysis
of the Haze in the United States: An Analysis of Data from the IMPROVE Network. Colorado
State University, Cooperative Institute for Research in the Atmosphere. Fort Collins, CO. July.
See EPA Air Docket A-96-56, Document No. VI-B-09-(ee).
Smith, R.S., 1974. "The Feasibility of an 'Injury Tax' Approach to Occupational Safety." Law
and Contemporary Problems, 38(4), 730-744.
Smith, R.S., 1976. The Occupational Safety and Health Act: Its Goals and Achievements.
American Enterprise Institute. Washington, DC.
Smith, V.K., 1983. "The Role of Site and Job Characteristics in Hedonic Wage Models."
Journal of Urban Economics, 13, 296-321.
Smith, V.K. and C. Gilbert, 1984. "The Implicit Risks to Life: A Comparative Analysis."
Economics Letters, 16, 393-399.
Taylor, C.R., K.H. Reichelderfer and S.R. Johnson. 1993. Agricultural Sector Models for the
United States: Descriptions and Selected Policy Applications. Iowa State University Press,
Ames, IA.
Thurston, G.D., K. Ito, C.G. Hayes, D.V. Bates and M. Lippmann, 1994. "Respiratory Hospital
Admissions and Summertime Haze Air Pollution in Toronto, Ontario - Consideration of the Role
of Acid Aerosols." Environmental Research, 65(2), 271-290.
Thurston, G.D., K. Ito, P.L. Kinney and M. Lippmann, 1992. "A Multi-Year Study of Air
Pollution and Respiratory Hospital Admissions in 3 New-York State Metropolitan Areas -
Results For 1988 and 1989 Summers." Journal of Exposure Analysis and Environmental
Epidemiology, 2(4), 429-450.
Tolley, G.S. and et al., 1986. Valuation of Reductions in Human Health Symptoms and Risks.
Prepared for U.S. Environmental Protection Agency. January. See EPA Air Docket A-96-56,
Document No. VI-B-09-(mmm).
U.S. Bureau of the Census. 1992. Statistical Abstract of the United States: 1992. 112 ed.
Washington, DC.
U.S. EPA, 1992. Regulatory Impact Analysis: Protection of Stratospheric Ozone. Volume II,
Part I, Appendix F. U.S. EPA, Stratospheric Protection Program. Washington, DC.
U.S. EPA, 1993. Air Quality Criteria for Oxides of Nitrogen. Volume II. U.S. EPA, Office of
Research and Development. Washington, DC. EPA-/600/8-91/049bF. August. NTIS
document No. PB92-176379INZ (order at 1-800-553-6847).
VII-77
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
U.S. EPA, 1994. Documentation for Oz-One Computer Model (Version 2.0). Prepared by
Mathtech, Inc., under Contract No. 68D30030, WA 1-29. Prepared for U.S. EPA, Office of Air
Quality Planning and Standards. Research Triangle Park, NC. August.
U.S. EPA, 1995. Acid Deposition Standard Feasibility Study Report to Congress. U.S. EPA,
Office of Air and Radiation, Acid Rain Division. Washington, DC. EPA 430-R-95-001a. May.
See EPA Air Docket A-96-56, Document No. VI-B-09-(11).
U.S. EPA, 1996a. ProposedMethodology for PredictingPM2.5 from PM10 Values to Assess
the Impact of Alternative Forms and Levels of the PMNAAQS. Prepared by Terence Fitz-
Simons, David Mintz and Miki Wayland (U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standard, Air Quality Trends Analysis Group). June 26. See EPA Air
Docket A-96-56, Document No. VI-B-09-(u).
U.S. EPA, 1996b. Review of National Ambient Air Quality Standards for Ozone: Assessment of
Scientific and Technical Information. OAQPS Staff Paper. U.S. EPA, Office of Air Quality
Planning and Standards. Research Triangle Park, NC. EPA-452\R-96-007. June. NTIS
document No. PB92-190446INZ (order at 1-800-553-6847).
U.S. EPA, 1996c. Air Quality Criteria for Ozone and Related Photochemical Oxidants. U.S.
EPA, Office of Research and Development. Washington, DC. EPA-/600/P-93/004cF. July.
NTIS document No. PC E99/MF E99, PB94-173119INZ (order at 1-800-553-6847).
U.S. EPA, 1996d. Regulatory Impact Analysis for Proposed Particulate Matter National
Ambient Air Quality Standard. Prepared by: Innovative Strategies and Economics Group, Office
of Air Quality Planning and Standards. Research Triangle Park, NC. December. Available at
http://www.epa.gov/ttn/oarpg/tlria.html
U.S. EPA, 1997a. Deposition of Air Pollutants to the Great Waters: Second Report to
Congress. U.S. EPA, Office of Air Quality Planning and Standards. Research Triangle Park,
NC. EPA-453/R-97-011. June. See EPA Air Docket A-96-56, Document No. VI-B-09-(ff).
U.S. EPA, 1997b. The Effects ofSOx and NOx Emission Reductions on Sulfate and Nitrate
Particulate Concentrations. Thomas Braverman, Air Quality Modeling Group, Office of Air
Quality Planning and Standards. Research Triangle Park, NC. May. See EPA Air Docket A-
96-56, Document No. VI-B-09-(cc).
U.S. EPA, 1997c. "Methodology Used to Create PM10 and PM2.5 Air Quality Databases for
RIA Work." Memorandum from David Mintz, Air Quality Trends Analysis Group, Office of Air
Quality Planning and Standards to Allyson Siwik Innovative Strategies and Economics Group,
Office of Air Quality Planning and Standards. July 15. See EPA Air Docket A-96-56,
Document No. VI-B-09-(kk).
VII-78
-------
Chapter VII: Benefit-Cost Analysis
U.S. EPA, 1997d. National Air Pollutant Emission Trends Procedures Document 1990-1996,
Section 4:0: National Criteria Pollutant Estimates, 1985-1996. Draft Document. Office of Air
Quality Planning and Standards. Research Triangle Park, NC. June. See EPA Air Docket A-
96-56, Document No. VI-B-09-(hhh).
U. S. EPA, 1997e. Regulatory Impact Analyses for the Paniculate Matter and Ozone National
Ambient Air Quality Standards and Proposed Regional Haze Rule. U.S. EPA, Office of Air
Quality Planning and Standards. Research Triangle Park, NC. July. See EPA Air Docket A-96-
56, Document No. VI-B-09-(r).
U.S. EPA, 1997f. "Response to Comments Made by AISI on EPA Methodology for Predicting
PM2.5 from PM10." Memorandum to the docket from Terence Fitz-Simons (Office of Air
Quality Planning and Standards, Air Quality Trends Analysis Group). February 6. See EPA Air
Docket A-96-56, Document No. VI-B-09-(w).
U.S. EPA, 1997g. The Benefits and Costs of the Clean Air Act: 1970 to 1990. U.S. EPA, Office
of Air and Radiation, Office of Policy, Planning and Evaluation. Washington, DC. October.
Available at http://www.epa.gov/airprogm/oar/sect812/copy.html
U. S. EPA, 1997h. Technical Support Document for the Regulatory Impact Analyses for the
Particulate Matter and Ozone National Ambient Air Quality Standards and Proposed Regional
Haze Rule. U.S. EPA, Office of Air Quality Planning and Standards. Research Triangle Park,
NC. July 17.
U. S. EPA, 1998a. The Regional NOx SIP Call & Reduced Atmospheric Deposition of Nitrogen:
Benefits to Selected Estuaries. September 22.
U.S. EPA, 1998b. Regulatory Impact Analysis for the NOx SIP Call, FIP, and Section 126
Petitions. U.S. EPA, Office of Air and Radiation. Washington, DC. EPA-452/R-98-003.
December. See EPA Air Docket A-96-56, VI-B-09.
Viscusi, W.K., 1978. "Labor Market Valuations of Life and Limb: Empirical Estimates and
Policy Implications." Public Policy, 26(3), 359-386.
Viscusi, W.K. 1979. Employment Hazards: An Investigation of Market Performance. Harvard
University Press, Cambridge.
Viscusi, W.K., 1981. "Occupational Safety and Health Regulation: Its Impact and Policy
Alternatives." In: Research in Public Policy Analysis and Management. Crecine, J., Ed. JAI
Press, Greenwich, CT. pp. 281-299.
Viscusi, W.K. 1992. Fatal Tradeoffs: Public and Private Responsibilities for Risk. Oxford
University Press, New York.
VII-79
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Viscusi, W.K., W.A. Magat and J. Huber, 1991. "Pricing Environmental Health Risks - Survey
Assessments of Risk - Risk and Risk - Dollar Trade-Offs For Chronic Bronchitis." Journal of
Environmental Economics and Management, 21(1), 32-51.
Ward, M.H., S.D. Mark, K.P. Cantor, D.D. Weisenburger, A. Correa-Villasenor and S.H. Zahm,
1996. "Drinking Water Nitrate and the Risk of Non-Hodgkin's Lymphoma." Epidemiology,
7(September), 465-471.
Watson, W. and J. Jaksch, 1982. "Air Pollution: Household Soiling and Consumer Welfare
Losses." Journal of Environmental Economics and Management, 9, 248-262.
Woodruff, T.J., J. Grille and K.C. Schoendorf, 1997. "The relationship between selected causes
of postneonatal infant mortality and particulate air pollution in the United States."
Environmental Health Perspectives, 105(6), 608-612.
Woolfolk, M.E., R.K. Iwamiya, C. Van Landingham, T.C. Myers, G.E. Mansell, H.H. Tunggal
and S.G. Douglas, 1998. A Prospective Analysis of Air Quality in the U.S.: Air Quality
Modeling. Systems Applications International, Inc. San Rafael, California. SYSAPP-97/69.
VII-80
-------
Chapter VIM: Regulatory Flexibility Analysis
Chapter VIII: Regulatory Flexibility
This chapter presents our Initial Regulatory Flexibility Analysis (IRFA) which evaluates
the impacts of the proposed Tier 2 and gasoline sulfur standards on small businesses. This
analysis has the following objectives: 1) to specify an appropriate definition for "small business"
for entities subject to the final rule, 2) to characterize small businesses in the petroleum refining
and motor vehicle manufacturing industries (described in more detail below in Table VIII-1), 3)
to assess the impact of the proposed standards on these businesses, and 4) to evaluate the relief
provided by regulatory alternatives.
A. Requirements of the Regulatory Flexibility Act
When proposing and promulgating rules subject to notice and comment under the Clean
Air Act, we are generally required under the Regulatory Flexibility Act (RFA) to conduct a
regulatory flexibility analysis unless we certify that the requirements of a regulation will not
cause a significant impact on a substantial number of small entities. The key elements of the
IRFA include:
• the number of affected small entities;
• the projected reporting, record keeping, and other compliance requirements of the
proposed rule, including the classes of small entities that would be affected and
the type of professional skills necessary for preparation of the report or record;
other federal rules that may duplicate, overlap, or conflict with the proposed rule;
and,
any significant alternatives to the proposed rule that accomplish the stated
objectives of applicable statutes and which minimize significant economic
impacts of the proposed rule on small entities.
The Regulatory Flexibility was amended by the Small Business Regulatory Enforcement
Fairness Act of 1996 (SBREFA), to ensure that concerns regarding small entities are adequately
considered during the development of new regulations that affect them.
In developing the NPRM, we concluded that the proposed Tier 2 and gasoline sulfur
standards would likely have a significant impact on a substantial number of small entities. To
comply with the requirements of the RFA, we were required to quantify these economic impacts.
The methodology used to calculate the per-refinery costs for desulfurizing gasoline is located
above in Chapter 5.B.; the cost for an average small refiner to comply with the 30 ppm standard
is described below in section C.
Based on the results of our economic analyses, we convened a Small Business Advocacy
Review Panel (the Panel), as required by SBREFA. The purpose of the Panel was to collect the
VIII-1
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
advice and recommendations of small entity representatives (SERs) that would be affected by the
proposed Tier 2 and gasoline sulfur standards. The report of the Panel has been placed in the
rulemaking record.1
B. Description of Affected Entities
A Tier 2 program establishing stringent vehicle emission standards and requiring
reductions in gasoline sulfur content would primarily affect manufacturers of LDVs, LDTs,
HDGVs, and oil refiners that produce gasoline. Most companies in these industries do not meet
the small business definitions provided in the U.S. Small Business Administration (SBA)
regulations (13 CFRPart 121). However, we have identified several companies within these
industries that are small businesses as defined by SB A. These businesses may be subject to the
Tier 2 vehicle and gasoline sulfur standards and could be significantly impacted by the new
standards. Table VIII-1, below, describes the affected industries, including the small business
size standards SBA has established for each type of economic activity under the Standard
Industrial Classification (SIC) and North American Industrial Classification systems.
VIII-2
-------
Chapter VIM: Regulatory Flexibility Analysis
Table VIII-1. Industries Containing Small Businesses
Potentially Affected by Today's Proposed Rule
Industry
Petroleum Refiners
Petroleum Marketers and
Distributors
Independent Commercial
Importers of Vehicles and
Vehicle Components
Alternative Fuel Vehicle
Converters
Motor Vehicle Manufacturers
NAICSCaks
324110
422710
422720
811112
811198
541514
336311
541690
336312
422720
454312
811198
541514
336111
336112
336120
SIC2 Codes
2911
5171
5172
7533
7549
8742
3592
8931
3714
5172
5984
7549
8742
3711
Defined by SBA as a
Small Business If:3
< 1500 employees
< 100 employees
< $5 million annual sales
< 500 employees
< 750 employees
< 100 employees
< $5 million annual sales
< 1000 employees
1) North American Industry Classification System
2) Standard Industrial Classification system
3) According to SBA's regulations (13 CFR 121), businesses with no more than the listed number of employees or
dollars in annual receipts are considered "small entities" for purposes of a regulatory flexibility analysis.
1.
Small Refiners
Of the approximately 160 petroleum refineries that currently produce gasoline in the U.S.,
about 15 meet SBA's definition of a small business. SBA's SIC code for petroleum refining is
2911. According to this code, a petroleum refining company must have fewer than 1500
employees to qualify as a SBA small business. In the event that we propose gasoline sulfur
control, some small refiners could have greater difficulty than larger refiners in complying with
the standard(s), due to such factors as limited operational flexibility, lack of access to alternate
crude oil feedstocks, limited availability of new sulfur reduction equipment, or difficulty in
raising capital to finance projects.
VIII-3
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
2. Small Petroleum Marketers
While refiners would be the primary affected parties in a gasoline sulfur control program,
some marketers of gasoline, many of which are small by SB A definitions, may be directly subject
to the rule and could be adversely impacted by it. This impact appears to be limited to new or
expanded requirements for reporting the sulfur content of gasoline samples.
SBA defines small businesses in this category (SIC codes 5171 and 5172) as those with
fewer than 100 employees. There are several hundred small gasoline marketers participating at
various points in the national gasoline distribution system.
3. Small Certifiers of Covered Vehicles
In addition to the major vehicle manufacturers, three distinct categories of businesses
relating to LDV, LDTs, and HDGVs exist that would be covered by Tier 2 emission standards.
Some companies in each of these categories are small businesses according to SBA regulations.
Small Independent Commercial Importers
Independent Commercial Importers are companies that hold a Certificate (or Certificates)
of Conformity which permits them to alter imported vehicles to meet U.S. emission standards.
As with alternative fuel vehicle converters described below, these businesses could face greater
technical challenges if emission standards are tightened. We have identified five businesses in
this category that are currently active and that appear to be small entities under SBA regulations.
Alternative Fuel Vehicle Converters
Under certain circumstances, our current policy permits the conversion of gasoline or
diesel vehicles to operate on an alternative fuel without applying for and receiving the EPA
Certificate of Conformity (also known as the "certification" process) that is required of
conventional manufacturers. However, certification can provide certain benefits to a converter,
and a few businesses have completed certification or have expressed interest in certifying
alternative fueled vehicle models. Beginning in model year 2000, converters must seek a
certificate for all of their vehicle models, although there will be some aspects of the certification
process that will be simplified for small volume manufacturers (SVMs), including these
converters. To the extent that companies are involved in this business when Tier 2 emission
standards become effective, they would be subject to such standards and could face greater
technical challenges in achieving the new standards with the vehicles they convert.
Small Volume Vehicle Manufacturers
We permit vehicle manufacturers selling 10,000 or fewer vehicles per year to be
VIII-4
-------
Chapter VIM: Regulatory Flexibility Analysis
designated as SVMs. This status allows vehicle models to be certified under a slightly simpler
certification process. More stringent Tier 2 standards could be relatively more difficult for small
manufacturers to achieve than larger manufacturers to the extent that research and development
resources are more limited. Less than five current SVMs meet the SB A guidelines for vehicle
manufacturers of 1000 or fewer employees.
C. Projected Costs of the Proposed Gasoline Sulfur Standards
The costs for an average-size small refinery (19,000 bbls gasoline/day) to produce
gasoline with a sulfur level of 30 ppm are described below in Table VIII-2. A more detailed
discussion of our refinery cost analysis, in general, can be found above in Chapter 5.
Table VIII-2. Costs for a 19,000 bbls gasoline/day
Refinery to Produce 30 ppm Gasoline
Location
PADD III
PADDIV
Per-Gallon Cost
(cents/gallon)
2.9
3.4
Operating Cost
($million/year)
8
9
Capital Cost
($million/year)
16
22
Costs for a small refinery located in PADD II to produce 30 ppm gasoline would fall
between the costs for a refinery in PADD III and a refinery in PADD IV.
In comparison, the average annual sales of small refiners in the U.S. were approximately
$385 million for 1997 based on data obtained from Dun & Bradstreet.
D. The Types and Number of Small Entities to Which the Proposed Rule Would
Apply
The types and number of small entities to which the proposed rule would apply are
described in Table VIII-3, below.
VIII-5
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table VIII-3. Types and Number of Small Entities to
Which the Proposed Tier 2/Gasoline Sulfur Rule Would Apply
Type of Small Entity
Small Refiners
Small Gasoline Marketers
Small Certifiers of
Covered Vehicles
Number of Companies
Affected by Today 's Rule
Approximately 17
Several Hundred
Approximately 15
Using our 1990 refinery baseline data, established for the purposes of the RFG and
anti-dumping programs, we have estimated that small refiners produce approximately 3.5 percent
of all gasoline in the U.S. Furthermore, of the 17 refineries that we have identified as meeting
SBA's definition of small business, nine already have gasoline sulfur levels less than 90 ppm.
Therefore, approximately eight small refineries (out of 160 refineries in the U.S.) will need to
significantly reduce their gasoline sulfur levels to comply with the proposed gasoline sulfur
standards.
We are also aware that there are several hundred gasoline distributors/marketers in the
U.S. The proposed rule may include a new requirement for them to add sulfur content to the set
of gasoline quality parameters they currently report or record. However, this requirement should
not be burdensome since sulfur content is generally measured along with other parameters and
the results would simply need to be recorded and reported.
E. Projected Reporting, Recordkeeping, and Other Compliance Requirements
of the Proposed Rule
We are proposing to require that refiners and importers keep and make available to us
certain records which demonstrate compliance with the sulfur program requirements. These
records include information about each batch of gasoline produced or imported, including batch
volume, sulfur test results and calculations used to determine compliance. We believe that the
proposed recordkeeping requirements for refiners and importers are necessary to allow
independent auditors and our inspectors to determine if the gasoline produced or imported, in
fact, met the applicable sulfur standards when it left the refinery or import facility. A similar
record retention requirement is included in the RFG and anti-dumping regulations.
Because the information required to be reported under today's rule in many cases is not
included in the RFG and anti-dumping compliance reports, and because we believe it would be
difficult to adapt the present RFG and anti-dumping reports to include the information required
VIII-6
-------
Chapter VIM: Regulatory Flexibility Analysis
under today's proposed rule, we are proposing to require refiners and importers to submit a
separate annual sulfur compliance report along with the refiner's or importer's RFG and/or anti-
dumping compliance reports. The sulfur report form would be relatively short and would require
only the minimum information necessary to demonstrate compliance with the applicable sulfur
standards. Parties would be required to include the refiner and refinery facility registration
numbers or importer registration number issued under the RFG regulations, the total volume of
gasoline (RFG and conventional gasoline) produced at the refinery (or refineries, if aggregated
prior to 2006) or imported by the importer during the averaging period, and the annual average
sulfur content of the gasoline produced or imported. Small refiners who have EPA-approved
individual baselines also be required to include the sulfur standards applicable to the refinery.
F. Other Relevant Federal Rules Which May Duplicate, Overlap, or Conflict
with the Proposed Rule
The Tier 2 emission standards and gasoline sulfur control regulations that we are
proposing are similar in many respects to existing regulations; in some cases, these regulations
are replacing earlier requirements with more stringent requirements for refiners and vehicle
manufacturers. However, the Panel is not aware of any area where the new regulations would
duplicate, overlap, or conflict with the existing federal, state, or local regulations.
G. Regulatory Alternatives
The Panel considered a wide range of options and regulatory alternatives for providing
small businesses with flexibility in complying with potential Tier 2 vehicle emission and
gasoline sulfur standards. As a part of the process, the Panel requested and received comment on
several early ideas for compliance flexibility that were suggested by the SERs and Panel
members. Taking into consideration the comments received on these ideas as well as additional
business and technical information gathered about the affected small entities, the Panel
recommended that we solicit comment on several of them. As described below, the Panel
recommended some of these concepts individually and, in the case of small refiners,
recommended a comprehensive option that incorporates several ideas. The Panel took
considerable time in addressing the concerns of the small refiners, who indicated their belief that
their businesses may have to close if relief is not considered for their industry. Taken together,
the Panel believed that these options would provide meaningful relief to small businesses in each
of the industry sectors potentially affected by a Tier 2/gasoline sulfur control program while
protecting the environmental goals of the program.
1. Small Refiners
The Panel recommended that small refiners be provided a four- to six-year period during
VIII-7
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
which less stringent gasoline sulfur requirements would apply. Each refinery's gasoline sulfur
limit would be based on its individual average sulfur level as reported in its most recent batch
report (submitted under the reformulated gasoline program, e.g., for 1997) available at the time
of the proposed rule. This four- to six-year period of relief would begin at the time that final
standards become effective for the refining industry as a whole. Following this period of relief,
small refiners would be required to meet the industry-wide standard, although temporary
hardship relief would be available on a case-by-case basis. The Panel believed that the additional
time that this approach would provide would 1) allow larger refiners to demonstrate new sulfur-
reduction technologies, 2) permit the costs of advanced technology units to drop as the volume of
their sales increases, 3) free-up industry engineering and construction resources, and 4) provide
additional time for small refiners to raise capital for infrastructure changes.
Although during the Panel process we had not yet decided on an approach for a proposed
sulfur control program, several small refiner options were discussed which made assumptions
about the program that might be in place. Among the program designs that we were considering
during the Panel process, the "worst case" scenario for small refiners was a national, year-round
sulfur requirement of 30 ppm on average with an 80 ppm per-gallon cap beginning in 2004. The
following discussion of the specific small refiner relief provisions assumed the existence of the
"worst case" scenario and a scenario where the gasoline sulfur standards would be higher than 30
and 80 ppm. The Panel emphasized that we had not yet made decisions regarding the level and
scope of sulfur controls that we were intending to propose.
a. Interim Sulfur Standards
In the Panel's recommended approach, small refiners covered by this special provision
would be assigned interim sulfur standards based on their individual refinery gasoline sulfur
levels today, according to Table VIII-4 below.
VIII-8
-------
Chapter VIM: Regulatory Flexibility Analysis
Table VIII-4. Federal Gasoline Sulfur Program with Sulfur
Standards of 30 ppm on Average and an 80 ppm Per-Gallon Cap
Average Refinery Sulfur Level (ppm)
Oto30
31 to 80
81 to 200
201 and above
Interim Sulfur Standards (average/cap,
ppm) *
30/80
80 (Cap only)
Average: Maintain current average level
Cap: Factor of 2.0 above the average
Average: One-half current average level
minimum and 300 ppm maximum
Cap: Factor of 1.5 above average level
200 ppm
* Note that if the federal program were to include a phase-in of sulfur standards, and if a refiner's current average
sulfur level was below the phase-in level, the phase-in level would become the refiner's compliance level for the
period of the phase-in.
More generally, if standards higher than 30/80 ppm were promulgated, the recommended
interim standards for small refiners would be at the levels described in Table VIII-5, below.
Table VIII-5. Federal Gasoline Sulfur Program with Sulfur
Standards Above 30 ppm on Average and an 80 ppm Per-Gallon Cap
Average Refinery Sulfur Level (ppm)
0-200
201-400
401-600
601 and above
Interim Sulfur Standards (average/cap, ppm) *
Average: Maintain federal standard or current
average level
Cap: Factor of 2 times the average
Average: 200 ppm or federal standard
Cap: Factor of 1 .5 times the average
Average: One-half of current average level
Cap: Factor of 1 .5 times the average
300/450
* Note that if the federal program were to include a phase-in of sulfur standards, and if a refiner's current average
sulfur level was below the phase-in level, the phase-in level would become the refiner's compliance level for the
period of the phase-in.
/'. Duration of Interim Standards
In addition to recommending that we propose a duration of four to six years during which
the interim standards would apply, beginning from the effective date of the sulfur standard, the
VIII-9
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Panel also recommends that we specifically request comment on an alternative duration of 10
years.
b. Hardship Relief
/'. Small Refiners
The Panel believed that it would be impossible to predict what the nature of the refining
industry would be in the latter part of the next decade, when small refiners will need to comply
with the final gasoline sulfur standard(s). Given this uncertainty, the Panel recommended that
we propose provisions for small refiners that would allow us on a case-by-case basis to extend
some form of relief from the standards for an additional period of time in cases of severe
hardship. The Panel recommended that we design such a proposed hardship relief provision to
include, at a minimum, the following characteristics:
• Criteria for granting of hardship relief that are sufficiently specific to help assure
fairness among recipients of such relief while allowing a degree of flexibility for
EPA to address special problems that may face individual refiners. Such criteria
should be designed to require a demonstration that the refiner faces extreme
economic consequences absent the relief and has exhausted other channels that
could limit the consequences. EPA should consider including in proposed
hardship relief provisions criteria such as, demonstrated inability on the part of the
small refiner to develop sufficient capital, the temporary unavailability of new
lower-cost sulfur removal technology, or the temporary unavailability of
engineering or construction resources necessary for the design and installation of
the new equipment.
A provision for a small refiner to propose an appropriate time period for this
additional relief. The Panel believes that the refiner should be expected to
carefully document the need for a specific period of additional relief. The Panel
also believes that such a period should be a minimum of two years so that the
refiner can demonstrate a degree of stability into the future when seeking capital
or credit.
The Panel was hopeful that the time provided by the interim standards for small refiners
(perhaps added to any time provided by a phase-in of the industry-wide program) would allow
for industry technology prove-out and cost reductions and for individual refiner planning such
that hardship relief would be seldom or never needed. The Panel was also satisfied that current
OMS management is committed to providing hardship relief if and when the need is
demonstrated and we encourage future OMS management to be similarly open to small refiners
facing dire economic impacts due to gasoline sulfur reduction standards.
VIII-10
-------
Chapter VIM: Regulatory Flexibility Analysis
Finally while the Panel recommended a refinery-based compliance option for small
refiners, as discussed above, OMB noted that the Panel received comments from small refiners
and small gasoline distributors supporting a geographically-limited sulfur program proposed by
API and NPRA. In light of these comments, OMB recommended that we evaluate the
API/NPRA proposal.
/'/'. Small Marketers of Gasoline
The Panel believed that adding gasoline sulfur to the fuel parameters already being
sampled and tested by gasoline marketers will likely result in little, if any, additional burden.
The gasoline marketer SERs that commented to the Panel did not address this issue. The Panel
did not recommend any special provisions for gasoline marketers. (These parties raised concerns
about indirect effects of a sulfur control program on marketers, especially if some refiners go out
of business and reduce the number of gasoline suppliers. However, the focus of the RFA and
SBREFA is on direct effects of a potential rule on small entities, which in this case do not appear
to be problematic.)
2. Small Certifiers of Covered Vehicles
The Panel recommended that EPA solicit comment on several ideas suggested by small
companies that certify LDVs, LDTs, and HDGVs, as discussed further below. However, several
other concerns that these businesses raised to the Panel do not appear to be affected by potential
new Tier 2 emission standards but rather involve existing regulations. While the Panel did not
believe that these "non-Tier 2" issues would be appropriately addressed in a Tier 2 rulemaking,
the Panel encouraged EPA to meet with small certifiers designated as ICIs to discuss those
issues.
The Panel recommended that EPA solicit comment on the following potential regulatory
options:
1) For small certifiers that convert imported vehicles to U.S. standards or that convert
vehicles to operate on alternative fuels, provide a delay in required compliance of two
years after Tier 2 standards apply to the model (engine family) involved.
2) If the Tier 2 program involves a phase-in of standards, allow small certifiers to comply at
the end of such a phase-in.
3) If the Tier 2 program does not involve a phase-in of standards, delay compliance for small
certifiers until 2007 (or three years after the program begins industry-wide).
4) Establish a credit program as a part of the Tier 2 program, and provide incentives for
large manufacturers to make credits available to small certifiers. In addition, develop a
VIII-11
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
program to provide credits to small certifiers for taking older vehicles off the road
(scrappage).
5) Design a case-by-case hardship relief provision that would delay required compliance for
small certifiers that demonstrate that they would face a severe economic impact from
meeting the Tier 2 standards.
VIII-12
-------
Chapter VIM: Regulatory Flexibility Analysis
Chapter VIII. References
1. Report of the Small Business Advocacy Panel on Tier 2 Light-Duty Vehicle and Light-
Duty Truck Emission Standards, Heavy-Duty Gasoline Engine Standards, and Gasoline
Sulfur Standards, October 1998.
VIII-13
-------
Appendix A: 47-State and Four-Cities Analyses
Appendix A: 47-State and Four-Cities Analyses
Double-click on the icon below to open Appendix A:
Acrobat Document
A-l
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
A-2
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Appendix B: Evidence Supporting the Irreversibility of
Sulfur's Emission Impact
Fuel sulfur impacts vehicle emissions in two basic ways. One is an immediate impact,
which occurs within a few miles of driving. The other is a more lasting impact, ranging from 20
or more miles to potentially permanent. This lasting effect of sulfur on emissions is termed
irreversibility, referring to the fact that the emission impact of high sulfur fuel does not reverse
when low sulfur fuel is used.
The immediate impact of sulfur on emissions is summarized in an EPA technical report.1
There, it was shown that operation on typical conventional gasoline containing 330 ppm sulfur
increases exhaust VOC and NOx emissions from LEV and Tier 2 vehicles (on average) by 40
percent and 150 percent, respectively, relative to their emissions with certification fuel containing
roughly 30 ppm sulfur. All of the data supporting these impacts were generated with very short
exposures to high sulfur gasoline, essentially a few miles of preconditioning and a few miles of
actual emission testing. When the vehicles were tested using low sulfur fuel after being operated
on high sulfur fuel, special preconditioning was performed to ensure that any residual effect of
the high sulfur fuel was removed. This preconditioning would not normally occur through
normal vehicle operation, so the emission impacts described in Section III only strictly apply to
situations where the vehicle operated on fuel with a single sulfur level over its entire life.
In this section, we are concerned with the impact of sulfur under more realistic
conditions. In particular, we are interested in vehicles' emission response following exposure to
low sulfur fuel after exposure to high sulfur fuel. We are also concerned with the potential that
long term exposure to high sulfur fuel may increase emissions to a greater degree than the short
term exposures simulated in most emission testing.
This section is divided into five parts. The first section describes the sensitivity of
vehicle exhaust emissions to gasoline sulfur content. The second discusses the theory of how
sulfur affects catalytic activity and the conditions conducive for its removal (sulfur
reversibility/irreversibility). The third describes the vehicle testing programs which have
attempted to measure the reversibility of the sulfur impact. This part also compares the relative
impact of long term sulfur exposure versus short term exposure. The fourth presents criteria for
evaluating the wide range of sulfur reversibility data which are available. Finally, the fifth
describes EPA's projections of the degree of sulfur reversibility for various vehicle types (e.g.,
Tier 1 vehicle, LEVs, and Tier 2 vehicles).
A. Exhaust Emission Sensitivity to Sulfur Content
The sulfur in gasoline increases exhaust emissions of HC, CO, and NOx by decreasing
the efficiency of the three-way catalyst used in current and advanced emission control systems.
B-l
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
For the purpose of this document, we will refer to this phenomenon as "sulfur sensitivity." Sulfur
sensitivity has been demonstrated through numerous laboratory and vehicle fleet studies. These
studies have demonstrated that significant reductions in HC, CO, and in particular, NOx
emissions can be realized by reducing fuel sulfur levels. Sulfur sensitivity for Tier 0 and Tier 1
vehicles is marginal, with NOx emissions decreasing between 11 percent to 16 percent when
sulfur is reduced from 330 ppm to 40 ppm. Sulfur sensitivity for LEV and ULEV vehicles,
however, is much more significant. When sulfur is increased from 40 ppm to 330 ppm, we
project that emissions increase by the following percentages:
Vehicle Type NMHC NOx
LEV and ULEV LDV 40% 134%
LEV and ULEV LOT 24% 42%
These percentages apply to "normal emitting" vehicles, which generally are those in-use vehicles
with emissions at or below twice their applicable emission standards. Higher emitting vehicles
are projected to be less sensitive to sulfur, because the catalyst is not operating at peak efficiency
in-use and should therefore be less affected on a percentage basis by higher sulfur levels.
We anticipate that Tier 2 vehicles will be at least as sensitive to sulfur as LEV and ULEV
LDVs and possibly even more so, due to the greater stringency of the proposed Tier 2 emission
standards, especially for NOx. At present, however, we have only projected that Tier 2 vehicles
will be just as sensitive as LEV and ULEV LDVs and not more so.
More detailed discussions of sulfur sensitivity can be found in the "EPA Staff Paper on
Gasoline Sulfur Issues,"2 published May 1, 1998, and the EPA report which developed sulfur
sensitivity estimates for a range of vehicle classes for incorporation in the draft version of EPA's
fleet-wide emissions model, MOBILE6. This report is titled "Fuel Sulfur Effects on Exhaust
Emissions"3 and is dated January 5, 1999.
Sulfur sensitivity has been shown to be variable and to depend upon both catalyst
formulation and vehicle operating conditions, which are discussed in detail in both reports.
Another variable, which was not discussed in either report, is the effect of real world vehicle
aging with sulfur. Sulfur sensitivity is temperature dependent. Sulfur adheres to the catalyst
surface more thoroughly at lower catalyst temperatures (approximately 450°C to 500°C) than
higher temperatures. Several vehicle manufacturers have suggested that the sulfur sensitivity
results from the numerous fleet studies actually underestimate the sensitivity of sulfur on exhaust
emissions, because the test cycles (FTP or LA4 cycles) used to saturate the catalyst with sulfur
result in catalyst temperatures that are too high. Specifically, the argument is that most vehicles
achieve catalyst temperatures over the FTP that exceed 450°C, thus not allowing complete
adsorption of sulfur to the catalyst surface, whereas real-world vehicle operation in metropolitan
non-attainment areas quite frequently result in catalyst temperatures at or below 450°C.
B-2
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
We have a sulfur test program currently underway to assess the validity of this concern.
Although testing is still ongoing, very preliminary results from a single vehicle seems to support
the fact that emissions are even more sensitive to sulfur when the catalyst is saturated with sulfur
via on-road operation. We plan to have the results of this test program available prior to the
development of technical analyses for the final rule. If the remainder of the testing supports the
early results and manufacturer's contentions, real world levels of sulfur sensitivity would be even
greater than the levels discussed above, further supporting the need for sulfur reduction in
gasoline.
A second concern about the current estimates of sulfur sensitivity is that all of the
vehicles in the test programs used to develop to above projections of sulfur sensitivities were
only exposed to high sulfur fuel for a few miles of driving prior to emission testing. In addition
to adsorbing onto the surface of the catalyst, sulfur can also penetrate into the precious metal
layer, especially into palladium, and into the oxygen storage material. This penetration may not
have fully occurred during the very few miles of operation prior to emission testing on high
sulfur fuel. In an API sulfur reversibility test program discussed further below in this appendix,
vehicles' sulfur sensitivity were measured after both short-term exposure to high sulfur fuel and
after 1,000-2,000 miles of driving with high sulfur fuel. For the five vehicles tested, NMHC
emission sensitivity was the same with both short-term and longer-term exposure to high sulfur
fuel. However, NOx emission sensitivity was 25-50% higher after longer-term exposure to high
sulfur fuel when compared to short-term exposure. Thus, the above sulfur sensitivities could
significantly underestimate the impact of sulfur on NOx emissions for LEVs, ULEVs and Tier 2
vehicles. We plan to investigate this issue further in the future.
B. Theory Supporting the Reversibility and Irreversibility of Sulfur's Emission
Impact
Sulfur impacts emissions from modern vehicles primarily by reducing the efficiency of
the three-way catalyst. Molecules of sulfur (either in the form of sulfur dioxide or hydrogen
sulfide) adsorb on the catalyst surface and basically take up space so that molecules of HC, CO
and NOx cannot adsorb and react to form water, nitrogen, oxygen and carbon dioxide. With
palladium catalysts, it appears that sulfur also penetrates into the metal itself, forming a reservoir
of sulfur within the catalyst. Sulfur dioxide also penetrates into the oxygen storage medium of
the catalyst and reduces the ability of the catalyst to manage the level of oxygen on the catalyst
surface. This oxygen management function is a key component of the 98 percent plus
efficiencies of today's three-way catalysts, particularly for controlling NOx emissions.
EPA summarized the basic chemical and thermodynamic mechanisms involved in
sulfur's two types of interference in it staff paper on gasoline sulfur in May of 1998.4 This paper
also summarized the conditions required to remove sulfur from the catalyst once the vehicle had
been exposed to high sulfur fuel. The results of a number of studies showed that generally high
temperatures (in excess of 700° F) are required to remove sulfur from both the surface of the
catalyst and from the washcoat matrix. In addition to high temperature, a rich exhaust (absence
B-3
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
of oxygen coupled with presence of HC and CO, or a low air-fuel ratio) or an alternating
sequence of rich and lean (presence of more oxygen in the exhaust than is needed to oxidize the
HC and CO present, or a high air fuel ratio) exhaust was often needed to fully regenerate the
catalyst. Larger degrees of lean and rich exhaust appear to be much more conducive to sulfur
removal than small changes in air fuel ratio. When these rich or alternating rich-lean conditions
were not present, even higher temperatures were required to remove the sulfur from the catalyst,
when such removal was successful. However, when the combination of temperature and
variation in the air-fuel ratio is sufficient, the sulfur accumulated from operation on high sulfur
fuel appears to be essentially eliminated and the emission impact of the high sulfur fuel is fully
reversed.
If sulfur reversibility was the only criteria involved in catalyst design, auto manufacturers
could place their catalysts right up against the engine and design the onboard computer to vary
the air fuel ratio from rich to lean sufficiently to regenerate the catalyst after any temporary
exposure to high sulfur fuel. Engine exhaust temperatures are generally high enough at the
exhaust manifold during typical driving to facilitate sulfur removal. The onboard computer is
certainly capable of varying the air-fuel ratio significantly. However, other critical catalyst design
criteria prevent such the use of such simple measures. First, excessive temperatures can
thermally damage the catalyst and reduce its efficiency. Second, simultaneously high conversion
efficiencies of HC, CO and NOx require very tight air fuel ratio control (minimal swings to either
rich or lean conditions).
Regarding catalyst temperature, auto manufacturers must balance a number of conflicting
criteria. One important criterion for catalyst design is that it light-off quickly. Most of the HC
and CO emissions from LEV vehicles, and significant amounts of NOx emissions, occur prior to
catalyst light-off Achieving this has affected the type and amount of materials used in the
catalyst and resulted in moving the catalyst closer to the engine. Many manufacturers have
switched to catalysts containing palladium, which generally can withstand higher temperatures
than platinum and rhodium catalysts. At the same time, catalyst manufacturers have improved
the design of their platinum and rhodium catalysts so that they can withstand higher
temperatures, as well. Moving the catalyst closer to the engine also increases catalyst
temperature during warmed-up operation, other factors being equal. Despite improvements in
the thermal durability of catalysts, sufficiently high temperatures can still cause a significant loss
of catalyst efficiency.
Engine load also affects exhaust and catalyst temperature. The engine load for a given
vehicle is a function of vehicle speed, rate of acceleration, vehicle weight and road grade, with
higher levels of all of these factors leading to higher engine loads and catalyst temperatures.
Vehicles which carry the most widely varying loads and which are driven the most aggressively
will generally experience the most variation in their catalyst temperature. Manufacturers must
design their catalysts to both light-off quickly and stay warm under light loads while not
sustaining thermal damage under heavy loads. Light trucks and sporty vehicles probably present
the most difficult challenges in this regard. For example, light trucks are most often driven with
one person and minimal cargo. However, they also are used to carry numerous passengers or
B-4
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
carry or pull heavy cargo up steep hills. The catalyst must be designed to withstand the higher
temperatures of these heavier loads.
One additional factor affecting catalyst temperature is the upcoming implementation of
EPA and California SFTP standards. The SFTP standards address emissions generated while the
vehicle is driving aggressively (high speeds and high rates of acceleration) and while the air
conditioning is turned on, both of which generate higher engine loads than exist during EPA's
FTP test cycle. Manufacturers have historically designed their engines to run rich under high
loads. The excess fuel decreases exhaust and catalyst temperature relative to an engine running
at stoichiometry (just the right amount of air to burn the fuel). The SFTP standards will require
that manufacturers reduce much of this high-load enrichment in order to reduce HC and CO
emissions during these high loads. Therefore, all other factors being equal, exhaust and catalyst
temperatures under extreme conditions will increase after implementation of the SFTP standards,
which begin their phase-in in the 2001 model year. Thus, the SFTP standards incrementally
increase the difficulty of quickly lighting-off the catalyst while still protecting it from thermal
damage during extreme driving conditions. While these extreme conditions must be considered
in the catalyst design process, their frequency in-use is not sufficient to rely upon for sulfur
removal. For example, some vehicle owners own and tow trailers up steep hills, while others do
not. Therefore, while the SFTP standards may increase temperatures under some conditions,
they will not necessarily increase sulfur removal capability for the general vehicle population.
Requiring manufacturers to increase the temperature of their catalysts under light loads to
improve sulfur reversibility would therefore increase temperatures under heavy loads even
further. EPA has not assessed the feasibility of manufacturers increasing warmed-up catalyst
temperatures beyond that required by the current standards, as well as the proposed Tier 2
standards, without additional degradation in catalyst efficiency. However, based on the Johnson-
Matthey data presented in the next section, even very high temperatures of 900° F are not
necessarily sufficient to fully reverse the sulfur impact if extensive use of high sulfur fuel has
occurred. Regular operation at such temperatures places the catalyst at risk of thermal damage
from even occasional excursions above this level, which can regularly occur from the types of
high load operation described above, as well as occasional spark plug misfire. Since the vast
majority of the HC, CO and NOx emission control occurring under both the current standards
and the proposed Tier 2 standards relies on the proper operation of the catalyst over the life of the
vehicle, increasing catalyst temperatures to enhance sulfur reversibility risks essentially all of the
benefits of EPA's exhaust emission control program (both current and proposed),. Therefore, it
would be imprudent to require vehicle manufacturers to design catalysts that operate at
temperatures high enough to improve the reversibility of sulfur effects and also meet the
proposed Tier 2 standards in-use.
Moving to the variation in air-fuel ratio, manufacturers have significantly enhanced their
engines' and computers' abilities over the past few years specifically to avoid large swings in
rich and lean operations. This ability to maintain tight control of the air-fuel ratio has increased
catalyst efficiency significantly in the process. Designing the vehicle to have alternating rich-
lean operation may improve the reversibility of sulfur effects, but would reduce catalyst
B-5
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
efficiency and potentially prevent the achievement of both current and proposed Tier 2 exhaust
emissions standards. As was the case with increasing catalyst temperature, it would be counter-
productive to reverse this progress in overall emission control just to enhance the sulfur
reversibility of catalyst systems.
Thus, the two changes in emission control design, hotter catalyst temperatures and
variable air-fuel ratios both run counter to other design criteria aimed at achieving stringent
emission standards in-use. Therefore, EPA believes that sulfur reversibility should be evaluated
with the catalyst temperatures and air-fuel ratio control of today's cleanest vehicles, considering
the impact of the future SFTP standards. The next section will do this by evaluating the available
sulfur reversibility data on such vehicles.
C. Results of Sulfur Reversibility Test Programs
EPA has received data from three test programs which evaluate the reversibility of
sulfur's impact on vehicle emissions. These three programs are summarized in the following
three sections. A fourth section summarizes other test data received, as well as describing some
EPA testing which is underway.
1. Coordinating Research Council (CRC) Sulfur Reversibility Program
The CRC sulfur reversibility program evaluated six 1997 LEV LDV models that were
part of their original sulfur sensitivity program. The following table lists the six vehicles used in
the program.
Table B-l. CRC Test Vehicles
Vehicle
Ford Taurus
Ford Escort
Honda Civic
Toyota Camry
Nissan Sentra
Suzuki Metro
Number of Cylinders
6
4
4
4
4
4
Engine Displacement
3.0L
2.0L
1.6L
2.2L
1.6L
1.3L
All six vehicles were equipped with 100K mile bench aged catalysts and oxygen sensors.
B-6
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Testing was performed in two phases -1 and II. Phase I consisted of three FTP tests (with a
single LA4 cycle run in between) with an initial baseline fuel containing 30 ppm sulfur. Three
additional FTP tests (again with the single LA4 preconditioning) were run using fuel containing
600 ppm sulfur. In order to evaluate the reversibility of the effects of the higher 600 ppm sulfur
from the catalyst surface of the six vehicles, all of the vehicles ran eight FTP tests using an LA4
test just prior to each FTP as a sulfur "purge" cycle. The LA4 cycle was chosen as a purge cycle
because of its general representativeness of city driving. Reversibility was defined as the ratio of
1) the difference between the average of emissions with high sulfur fuel and the average of
emissions from the subsequent eight tests using low sulfur fuel to 2) the difference between the
average of the high sulfur results with the average of the initial baseline low sulfur results. Total
mileage accumulation during purge testing was roughly 250 miles. In other words, after 250
miles of operation, emission performance stabilized and no further purging of sulfur from the
catalyst surface occurred.
Phase II consisted of three FTP tests with fuel containing 600 ppm sulfur followed by two
FTP tests with 30 ppm sulfur fuel with an LA4 purge cycle prior to each FTP. Six FTP tests
were then performed with a US06 cycle prior to each FTP as a sulfur purge cycle. The US06
cycle was chosen as a purge cycle to simulate aggressive high speed and load operation that
would encourage higher catalyst temperatures and rich A/F operation. Reversibility was
determined in the same manner as in phase I (same initial 30 ppm sulfur baseline). Total mileage
accumulation turned out to be roughly 200 miles.
The following table lists the results of the CRC sulfur reversibility test program.
B-7
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table B-2. Sulfur Reversibility: CRC Test Program (%)
Vehicle
Manufac
Ford
Ford
Honda
Nissan
Toyota
Suzuki
Fleet
Estimate
Models
Taurus
Escort
Civic
Sentra
Camry
Metro
NMHC
Purge Cycle
LA4
69.0
137.0
94.0
99.0
112.0
170.0
97.0
US06
83.0
122.0
99.0
111.0
98.0
165.0
108.0
NOx
Purge Cycle
LA4
70.0
95.0
96.0
85.0
50.0
86.0
84.0
US06
95.0
100.0
97.0
88.0
102.0
87.0
95.0
The fleet estimate used for the CRC data was determined by averaging the baseline low
sulfur results, the high sulfur results and the final low sulfur results for all vehicles and
determining reversibility as discussed above. These results indicate that on average, NMHC
emissions are very reversible, regardless of purge cycle used (LA4 or US06). The Ford Taurus,
however, showed only a moderate level of reversibility for NMHC, especially with the LA4
purge cycle (69 percent). The results for NOx indicate that with the LA4 purge cycle, the
average level of reversibility is 84 percent with the Toyota Camry having reversibility as low as
50 percent. When using the US06 purge cycle, NOx emissions were far more reversible with an
average reversibility of 95 percent. The Nissan Sentra and Suzuki Metro showed almost the
exact same level of reversibility with both purge cycles.
2. American Petroleum Institute Sulfur Reversibility Program
The API programkkk evaluated a total of seven vehicles, four were 1998 LEV LDVs, one
was a 1998 ULEV LDV, and the other two were Tier 1 vehicles (LDV and LDT1). All of the
vehicles had been driven for 6,000-10,000 miles, except for the S10 pickup, which had 50,000
kkk
API has completed a third-party review of the results of their test program (as well as the CRC test
program). See "Reversibility of Gasoline Sulfur Effects on Low Emissions Vehicles," T. J. Truex and L.S. Caretto
for API, April?, 1999.
B-8
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
miles on it. API replaced the catalysts of all of the vehicles. Reversibility of the sulfur effect
was measured for all of these vehicles with their new catalysts thermally aged to the equivalent
of 4,000 miles (i.e., low mileage catalysts) and after only a very short exposure to high sulfur
fuel. Four of these vehicles were also tested with 1,000 miles of road aging on high sulfur fuel
(540 ppm) prior to reversibility testing.
The sulfur reversibility of two vehicles was also tested after short term exposure to high
sulfur fuel with their catalysts thermally aged to represent 100,000 miles of driving. (However,
the oxygen sensors were not aged.) Finally, one vehicle was tested after 2,000 miles of driving
using high sulfur fuel with its catalysts thermally aged to represent 100,000 miles of driving.
All of the vehicles were tested in a sequence similar to the one used by CRC. The
program started with testing using low sulfur fuel (40 ppm). This was followed by testing with a
high sulfur fuel (540 ppm). Then, the fuel was switched back to the low sulfur fuel and the
vehicle operated over either an LA4 or US06 cycle, which was used as a sulfur purge cycle.
Following this purge cycle, emissions were again measured with the FTP.
One major difference between the API and CRC programs was that API generally only
performed two tests at each sulfur level, including the purge cycle phase. Thus, statistically
speaking, the API program is weaker than the CRC program. Examination of individual
emission test results shows significant variability occurred.
Table B-3 lists the vehicle tested in the API program.
Table B-3. API Test Vehicles
Vehicle
1998 Ford Taurus (LEV)
1998 Honda Accord (ULEV)
1998 Toyota Aval on (LEV)
1998 Nissan Altima (LEV)
1998 Ford Grand Marquis (LEV)
1998 Ford Town Car (Tierl)
1997 Chevrolet S-10 (Tierl)
Numberqf Cylinders
6
6
6
4
8
8
6
Engine Displacement
3.0L
2.3L
3.0L
2.4L
4.6L
4.6L
4.3L
API screened specific vehicles for this test program by performing emission testing over both
the FTP and the US06 cycle. API believed that these vehicles were nearly in compliance with
future SFTP standards and therefore representative of 2000 and later emission control
B-9
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
technology. This will be discussed further below.
Table B-4 shows the sulfur reversibility emission results for all of the vehicles when
tested with low mileage (4,000 mile) catalysts.
Table B-4. Sulfur Reversibility: API Test Program
Low Mileage Catalysts, Short-Term Exposure to High Sulfur Fuel (%)
Vehicle
Manufac
Ford
Honda
Toyota
Nissan
Ford
Ford
Chevrolet
Models
98 Taurus
98 Accord (ULEV)
98 Avalon
98 Altima
98 Gr. Mar
98 Town Car (Tierl)
97 S-10 (Tierl)
Fleet Estimate
NMHC
Purge Cycle
LA4
100.0
23.1
71.4
800
103.2
46.3
66.7
67.9
US06
n/a *
100.0
42.9
n/a*
80.6
60.0
154.2
45.9
NOx
Purge Cycle
LA4
96.2
78.3
52.1
125.0
84.5
95.0
70.3
83.3
US06
n/a
97.8
106.3
n/a
71.8
104.4
117.4
92.3
* Vehicle not tested with US06 purge cycle.
The most obvious difference between the reversibilities measured by API and those found
by CRC is that API's average NMHC reversibility rate when using the LA4 as a purge cycle is 68
percent, while CRC's average NMHC reversibility rate shows nearly ful reversible at 97 percent.
The measured NOx reversibilities (with the LA4 purge cycle) were almost identical in the two
programs, 83 percent for API compared to 84 percent for CRC.
API found much higher reversibility using the US06 cycle as a purge cycle for NOx (92.3
percent). However, the opposite was true for NMHC (45.9 percent). This 45.9 percent
reversibility is considerably lower than that found in the CRC program, where NMHC emissions
were essentially fully reversible after purging with the US06 cycle.
Another difference between the API and CRC test results is the great deal of disparity
between the reversibilities measured for individual vehicles in the API program. Some vehicles
were highly reversible while others were not. The CRC results appear to be more consistent
B-10
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
from vehicle-to-vehicle. This could be a result of the fact that CRC performed eight purge/FTP
combinations with low sulfur fuel after exposure to high sulfur fuel, compared to API, which
only performed two purge/FTP combinations. The CRC data showed that emissions after the
switch back to low sulfur fuel fluctuated up and down before reaching a more consistent level
during the eight tests. It is also possible that API simply experienced greater test-to-test
variability, or that the vehicles in the API program simply differed more in their inherent
reversibility.
Table B-5 shows measured reversibility for vehicles with low mileage catalysts that were
operated on high sulfur fuel (540 ppm) for 1,000 miles on the road. Four vehicles were evaluated
in this manner. The Taurus was tested with the LA4 purge cycle, but not the US06, while the
Accord, Avalon, and Grand Marquis all were tested with the US06 purge cycle but not the LA4.
As with the low mileage catalyst data, there is a significant amount of disparity between vehicles,
especially for NMHC reversibility with the US06 cycle. Reversibility of NOx emissions with the
US06 cycle, however, are consistent and indicate that the sulfur effect is almost fully reversible
with the US06 cycle. The Taurus with only short term exposure to high sulfur fuel was 100
percent reversible with the LA4 purge cycle for NMHC, but only 67.9 percent reversible with the
LA4 cycle after road aging. Reversibility of NOX emissions from the Taurus was nearly
complete for both short term and longer term exposure to high sulfur fuel.
Table B-5. Sulfur Reversibility: API Test Program
Low Mileage Catalysts, 1,000 Mile Exposure to High Sulfur Fuel (%)
Models
98 Taurus
98 Accord (ULEV)
98 Avalon
98 Grand Marquis
Fleet Estimate
1, 000 Mile Exposure
NMHC
Purge Cycle
LA4
67.5
n/a
n/a
n/a
67.5
US06
102.5
100.0
75.0
45.5
88.0
NOx
Purge Cycle
LA4
97.6
n/a
n/a
n/a
97.6
US06
169.0
94.5
101.4
101.9
106.5
Short-Term Exposure
NMHC
Purge Cycle
LA4
100.0
23.1
71.4
103.2
94.0
US06
n/a *
100.0
42.9
80.6
69.0
NOx
Purge cycle
LA4
96.2
78.3
52.1
84.5
77.7
US06
n/a
97.8
106.3
71.8
88.4
Table B-7 shows measured reversibility for vehicles with catalysts bench aged to
represent 100,000 mile of driving. Only two vehicles were tested with this configuration - the
Taurus and the Altima. Due to problems with the fuel tank on the original Altima used in the
program, a second Altima was procured and tested with a 100K catalyst system. Reversibility of
B-ll
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
the Altima's emissions was measured after both short-term exposure to high sulfur fuel, as well
as after 2,000 miles of highway driving with high sulfur fuel. This was the only vehicle in the
API program that had both a 100,000 mile catalyst and extended road aging with high sulfur
fuel. It was also the only vehicle with 2,000 miles of driving with high sulfur fuel instead of
1,000 like the other four vehicles with more extended use with high sulfur fuel.
Table B-6. Sulfur Reversibility: API with 100K Aged Catalysts Test Program (%)
Models
NMHC
Purge Cycle
LA4
US06
NOx
Purge Cycle
LA4
US06
Short-term Exposure to High Sulfur Fuel
98 Taurus
98 Altima
Fleet estimate
207.0
84.9
120.0
107.0
102.7
104.0
88.7
78.9
87.3
85.4
89.2
87.3
2,000 Mile Exposure to High Sulfur Fuel
98 Altima
n/a
115.1
n/a
93.9
The Taurus showed very similar levels of NMHC emission reversibility (after the LA4
purge cycle) with both low mileage and high mileage catalysts (essentially fully reversible in
both cases). NOx emission reversibility dropped from 96.2 percent with the low mileage catalyst
to 88.7 percent with the 100,000 mile catalyst. NOx emission reversibility did not improve after
purging with US06 cycles.
The first Altima tested, which had a 4000 mile catalyst, was fully reversibility for both
NMHC and NOx emissions with the LA4 purge cycle. The second Altima, which had a 100,000
mile catalyst showed less reversibility, only 84.9 percent for NMHC emissions and 78.9 percent
for NOx emissions. Both NMHC and NOx emission reversibility improved with purging with
the US06 cycle, though NOx emissions were still not fully reversible.
The second Altima showed similar NMHC and NOx reversibility with both short-term
and long-term exposure to high sulfur fuel with the US06 purge cycle. The second Altima was
not tested with the LA4 purge cycle.
While the focus of the API test program was reversibility, the fact both short and longer
term exposures to high sulfur fuel were evaluated also allows the comparison of sulfur sensitivity
under these two conditions. Table B-7 presents the FTP emissions on both low and high sulfur
B-12
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
fuel, the latter after both very short term exposure to high sulfur levels and 1,000-2,000 miles of
driving on high sulfur fuel. All five vehicles so tested showed greater NOx emission increases
after exposure to high sulfur fuel for 1,000-2,000 miles than occurred after short term exposure.
NMHC emissions, on the other hand, showed essentially the same sensitivity to sulfur after either
short or longer term exposure. The increased NOx emission sensitivity with extended mileage is
a concern, as the sulfur sensitivity being projected for current and future vehicles was derived
from testing which included only short term exposure to high sulfur fuel. The API data indicates
that the effect of sulfur on NOx emissions may be 30-50 percent greater than is currently being
projected. EPA plans to investigate this further in the future.
Table B-7. Sulfur Sensitivity: API Test Program
Low Mileage Catalysts, Short-Term Exposure to High Sulfur Fuel (g/mi)
FTP Test
Sulfur Level
Sulfur
Exposure
Vehicle
Taurus
Accord
Aval on
Gr. Marq.
Average
Altima
NMHC
30 ppm
—
540 ppm
Short-term
540 ppm
1,000 Mile
NOx
30 ppm
—
540 ppm
Short-term
540 ppm
1,000 Mile
Low Mileage Catalysts
0.033
0.029
0.040
0.044
0.037
0.051
0.036
0.058
0.075
0.055
0.073
0.041
0.060
0.055
0.057
0.075
0.100
0.068
0.040
0.071
0.101
0.164
0.130
0.143
0.135
0.117
0.245
0.143
0.152
0.164
100,000 Mile Catalysts
0.041
0.059
0.057
0.061
0.112
0.132
3. Johnson Matthey Sulfur Reversibility Program
Johnson Matthey (JM), a catalyst manufacturer, conducted a test program to evaluate if
long term exposure to high sulfur fuel damaged catalysts, whether the damage was reversible
when the system was run on low sulfur fuel, and to determine whether exposure of catalysts to
higher temperatures with low sulfur fuel reversed the damage. Four catalyst designs: Palladium
(Pd), Pd/Rhodium (Rh), Platinum (Pt)/Rh, and Pt/Pd/Rh, were bench aged for 45 hours
approximately equal to 50K miles. There were two sets of catalysts. Set A was aged using 87
ppm sulfur, while set B was aged using 735 ppm. Set A was always used as the baseline test.
B-13
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Set B was used to measure high sulfur results (735 ppm) and the consequent low sulfur results.
A single vehicle was used for all of the testing - a 1990 Tier 0 Mitsubishi Galant. Each catalyst
was installed on the vehicle and then evaluated. Each catalyst was located in a front underbody
position. Sulfur purging with the US06 cycle was not evaluated, only the LA4 purge cycle.
After evaluating reversibility performance of each catalyst to exposure of low sulfur fuel,
JM attempted to demonstrate the effect prolonged exposure to high catalyst temperatures would
have on sulfur reversibility, since most of the scientific literature has suggested that exposing the
catalyst system to temperatures over 700° C should facilitate reversibility. Each catalyst was run
over a steady-state sulfur recovery cycle which involved an A/F ratio (APR) oscillation of
stoichiometry +/- 0.5 APR @ 0.10 Hz (5 seconds rich/5 seconds lean) and a catalyst bed
temperature of 700° C. This procedure was followed by a cycle at 800° C and then a final cycle
at 900° C. What JM found was that according to their results, sulfur is highly irreversible even
when the catalyst is exposed for five straight hours to temperatures of 700° C, 800° C, and 900°
C.
Table B-8 lists the results of JM's sulfur reversibility program. The fleet estimate is also
a simple arithmetic average of the data. It should be noted that all of the catalysts had very poor
reversibility when only switching back to low sulfur fuel. As the catalysts were then exposed to
increasing temperature, the results became mixed - some catalysts improving, while others
deteriorated. For NMHC, the Pt/Rh catalyst appears to be the only catalyst that showed more
reversibility in response to increased temperatures. Curiously, most of the catalysts seemed to
respond best to the 700° C and 800° C temperatures and poorly to the highest temperature (900°
C). For NOx, the results were even poorer. For example, the Pd/Rh catalyst never experienced a
NOx sulfur reversibility rate above 18 percent. In fact, most of the time it had negative
reversibility rates, meaning once the catalyst was re-exposed to low sulfur fuel after having been
operated on high sulfur fuel, the low sulfur emission results were higher than the high sulfur
results.
B-14
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Table B-8. Sulfur Reversibility: Johnson Matthey Test Program (LA4 Purge Cycle) (%)
Catalyst
Pd
Pd/Rh
Pt/Rh
Pt/Pd/Rh
Fleet
Estimate
NMHC
NOx
Sulfur Purging Temperature
None
50.0
30.0
50.0
38.0
42.0
700° C
50.0
30.0
92.0
85.0
64.3
800° C
66.0
20.0
92.0
46.0
56.0
900° C
17.0
10.0
83.0
54.0
41.0
None
44.0
-5.0
32.0
47.0
29.5
700° C
75.0
18.0
32.0
65.0
47.5
800° C
88.0
-50.0
32.0
-35.0
8.8
900° C
69.0
-33.0
32.0
47.0
28.8
4. Other Testing
Honda has suggested that in order for complete sulfur adsorption onto the catalyst
surface, catalyst temperatures must be below 500° C. Honda believes that the cycles that have
been used in the various sulfur test programs to adhere the sulfur to the catalyst surface have
been inadequate. They believe catalyst temperatures over the FTP or LA4 test have generally
been exceeding 500° C at some point or another during the cycle. They proposed a conditioning
cycle that consisted of an extended 35 mph cruise making sure the catalyst temperature does not
exceed 450° C - 500° C. Full sulfur adsorption is determined by monitoring feedgas SO2 and
exhaust SO2 until they are the same. Honda found that when using their conditioning cycle,
NMHC emissions were 20 percent more irreversible than when using the FTP or LA4 for sulfur
conditioning. NOx emissions were 19 percent more irreversible.
As a result of Honda's information, the IK road aging results from the API program, and
the apparent effect of sulfur aging on the JM results, we have undertaken our own EPA sulfur
reversibility evaluation program. The primary focus of our program is to determine the effect
road aging has on sulfur adsorption to the catalyst and the subsequent removal or reversal of the
sulfur. At this time, testing is still underway and we are just about to start the road aging.
Although this data will not be available for the NPRM, it will be completed in time for the final
rule.
D. Criteria for Evaluating Sulfur Reversibility Data
Projecting the degree of sulfur irreversibility for various vehicles types under
representative in-use conditions is difficult due to inadequacies in essentially all of the available
data. As mentioned in the previous section, the sulfur reversibility testing would ideally have
B-15
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
used vehicles designed to meet a range of FTP and SFTP standards, thermally aged catalyst
systems prior to testing, exposed these systems to high sulfur fuel for a few thousand miles of
typical driving, and used representative driving cycles to purge sulfur between emission tests.
While many of the vehicles tested had thermally aged catalyst systems, none were
designed to meet SFTP standards. API tested the vehicles in their test program over the US06
cycle to assess the degree to which they might already be in compliance with future SFTP
standards. The results showed that two out of the eight (including the second Altima) vehicles
were below the 0.14 g/mi US06 NMHC+NOx standard for LEVs and ULEVs, while half of the
vehicles were below the 8.0 g/mi US06 CO standard. However, only one vehicle met the
NMHC+NOx standard with any significant margin of safety. Thus, this screening data does not
support the contention that these vehicles were essentially in compliance with the SFTP
standards. Also, API did not measure emissions over the SC03 cycle, which simulates emissions
with the air conditioning system turned on and which is also a part of the SFTP requirement.
Thus, there is no evidence that these vehicles were designed to meet non-US06 related SFTP
requirements.
In addition to US06 emission data, API also measured each vehicle's air-fuel ratio on a
second-by- second basis over both the FTP and US06 cycles. The standard deviation of the air-
fuel ratio over the FTP averaged 0.35, while that over the US06 cycle averaged 1.03. Thus,
variability in air-fuel mixture control was nearly three time as great over the US06 cycle as over
the FTP. Tight air-fuel mixture control is essential to maintaining low engine-out emissions and
high catalyst over the entire emission test and during in-use driving. The first step a
manufacturer will take in order to comply with the SFTP requirements will be to modify the
engine calibration to achieve the best level of air-fuel mixture control possible. The fact that
these vehicles showed much greater variability over the US06 cycle indicates that manufacturers
had not yet begun the process of making these vehicles SFTP-compliant. Previous studies have
shown that wide swings in air-fuel ratio reduce the impact of sulfur on catalyst efficiency and
emissions relative to minor swings in air-fuel ratio. Thus, this criterion appears to be of critical
importance in projecting the reversibility of SFTP-compliant LEVs and Tier 2 vehicles.
Moving to exposure to high sulfur fuel, while a few vehicles were operated on high sulfur
fuel for a thousand miles or more, only two vehicles had catalysts which were thermally aged to
more than a few thousand miles. Also, the driving cycles used to purge sulfur after switching
from high to low sulfur fuel were not any of the driving cycles developed to be fully
"representative" of recent driving patterns. The result of these shortcomings is that considerable
technical judgment has to be used to project the degree of sulfur irreversibility which would
occur for both current and future vehicles.
EPA established a number of criteria for evaluating the available data in order to project
likely levels of in-use sulfur reversibility. The first criterion is to focus exclusively on testing of
vehicles with thermally aged catalysts. We believe that this is essential, because catalysts prior to
thermal aging contain far more surface area and oxygen storage capacity than is needed to meet
low emission levels. It is possible for sulfur to deactivate a considerable portion of the surface
B-16
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
area and oxygen storage with minor impacts on overall catalyst performance. This would not be
representative of the impact of sulfur on real-world emissions over most of the vehicle's life.
The second criterion is to give priority to testing where the catalyst has been exposed to
high sulfur for a considerable period of time. Long term exposure would be the predominant
mode of exposure under a regional sulfur program such as that proposed by API. With the high
sulfur region being potentially quite large geographically, most vehicles entering it are likely to
be there awhile. Simply crossing the region that would have 300 ppm average sulfur levels under
the regional program proposed by API represents roughly 1000 or more miles.
Under any sulfur control program, sulfur levels will vary from batch to batch of gasoline
produced by refineries. For example, under the proposal, sulfur could easily vary from less than
10 ppm to 80 ppm in-use after 2008. Under this scenario, exposure to high sulfur fuel could be
one tankful at a time, or could continue for several tankfuls, depending on the production
patterns at refineries and the purchasing patterns of individuals. Still, even one tankful of fuel
typically lasts 300 miles or more. This is far more than the 10-20 miles of exposure to high
sulfur fuel which occurred in most of the testing summarized in the previous section.
Development of the subsequent criteria are more complex, because the issues of SFTP
compliance and representative driving cycles are not as easily addressed. None of the vehicles
tested were certified to either the Tier 1 or LEV SFTP emission standards. While some of the
vehicles may have SFTP emissions close to or even below the applicable SFTP standards, it is
still likely that manufacturers would change their engine calibrations to enhance compliance with
these emission standards in-use. This is confirmed by the API measurements of air-fuel ratio
over the FTP and US06 cycles. Thus, none of the test vehicles can be assured of having SFTP-
compliant engine calibrations.
Likewise, only the LA4 and US06 driving cycles were used in the test programs
performed to date. The LA4 cycle was derived from driving patterns in Los Angeles in the early
1970's. However, due to physical limitations in the dynamometers in use at the time, all
accelerations greater than 3.3 mph per second were reduced to this level. This, plus the fact that
driving has become more aggressive over the past 25 years makes the LA4 cycle less aggressive
on average than today's typical driving. However, the LA4 cycle does include driving as fast as
58 mph, so it is also not representative of light, city driving.
The US06 cycle is made up of real-world driving segments.111 However, the concentration
of aggressive driving is much higher than occurs in the real world. Therefore, the length of time
that the catalyst is exposed to both high temperatures and rich conditions is much higher than
would occur in the real world. This could easily remove more sulfur than would be removed in-
All but one of the segments were taken from EPA's REP05 cycle, which represents the aggressive
portion of in-use driving. The remaining segment was taken from ARB's HL07 cycle, which was intended to
represent aggressive in-use driving in California.
B-17
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
use even during aggressive driving.
As mentioned in Section B, meeting the SFTP standards will requiring the tightening of
air-fuel mixture control and reduce the amount of rich operation in-use during aggressive
driving. Both of these changes directionally reduce sulfur removal. This primarily affects the sulfur
reversibility testing after preconditioning with the US06 cycle. In particular, it casts considerable
doubt in the applicability of measured reversibilities using the US06 purge cycle to SFTP-
compliant vehicles. Therefore, the measured levels of sulfur reversibility after operation on
US06 cycles will not be used to project the in-use levels of sulfur reversibility for SFTP-
compliant vehicles.
For pre-SFTP vehicles, the US06 cycle still likely over-estimates the amount of sulfur
reversibility which would occur in-use, due to its unrepresentative concentration of high
temperatures and rich operation. Thus, the measured levels of sulfur reversibility after operation
on both LA4 and US06 cycles will be used to project the in-use levels of sulfur reversibility for
pre-SFTP vehicles.
In summary, the projections developed in the following section will:
1. Only use data from vehicles with aged catalyst systems,
2. Emphasize data from vehicles whose catalysts experienced substantial use with high
sulfur fuel,
3. For projections regarding SFTP-compliant vehicles, only use data where the sulfur was
purged using the LA4 cycle, and
4. For projections regarding pre-SFTP vehicles, use data where the sulfur was purged using
either the LA4 or US06 cycle.
E. Projected Levels of Sulfur Irreversibility In-Use
Applying the first criterion developed in Section D. results in the retention of the CRC
and JM data (Tables B-2 and B-8), as that testing was performed on vehicles with thermally aged
catalysts. It also allows the use of the API data contained in Table B-6. However, the remaining
API data apply to vehicles with low mileage catalysts, which are not sufficiently representative of
in-use operation. Therefore, EPA's current conclusions about reversibility of sulfur effects do
not rely on the API data except that in Table B-6.
Of these data, only the JM data and the second Altima tested by API also involved
extensive use of high sulfur fuel prior to the measurement of reversibility. JM simulated the use
of high sulfur fuel through oven aging, so it may not be fully representative as actual vehicle
driving on the road. However, oven aging to simulate on-road thermal degradation is well
established, so the same for sulfur aging should be equally acceptable.
API actually drove the second Altima on the road for 2,000 miles with high sulfur fuel.
B-18
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
However, the type of driving actually performed is not known, raising some uncertainly about its
representativeness. More importantly, API only measured reversibility with this vehicle after
operating the vehicle with low sulfur fuel over the US06 cycle. This took advantage of the
vehicle's widely varying air-fuel ratio over this test cycle and likely purged more sulfur off of the
catalyst than would occur with an SFTP-compliant vehicle. JM, on the other hand, simulated the
type of air-fuel ratio control which would be indicative of an SFTP-compliant vehicle with it
oven aging and purging. Thus, overall, the JM data should receive the greatest weight in this
analysis. Table B-9 summarizes the results of these three test programs.
Table B-9. Sulfur Reversibility: Summary of Relevant Test Programs (%)
Models
CRC (6 vehicles)
JM (4 catalysts)
API (2-3 vehicles)
NMHC
Purge Cycle
LA4
97%
4-64%
Compete
US06
Complete
41-64%
Complete
NOx
Purge Cycle
LA4
84%
30-48%
87%
US06
95%
9-48%
89%
The JM results are shown as a range, as it is difficult at this time to average the results at
the various sulfur purging temperatures. However, the JM data were placed into the two purge
cycle categories by assuming that the test results with no oven-based sulfur purging and the
purging at 700° C were similar to LA4 driving, while the results with oven purging at 700-900°
C were similar to US06 driving. Another relevant factor is that the results of the API testing are
the most erratic, primarily due to the relatively small number of replicate tests.
As can be seen, there is considerable variation in the measured levels of sulfur
reversibility in the above test data. In particular, the JM data show much less reversibility and
should be given the most weight because it was the only test program thus far to include catalysts
aged both thermally and with high sulfur fuel. Therefore, the overall projection of reversibility
primarily hinges on the relative weight given to the JM data. In any event, there will be
considerable uncertainty in any summary projection developed from these results, because of the
limitations in the test methods described above.
For pre-SFTP vehicles, we decided to utilize reversibility measurements using both the
LA4 and US06 driving cycles. Since the CRC and API vehicles are pre-SFTP vehicles, these
two test programs were given considerable weight along with those of JM. For these vehicles,
we project that NMHC emissions are fully reversible, while NOx emissions are only 85 percent
reversible. For SFTP-compliant vehicles, we decided above to utilize reversibility measurements
using only the LA4 driving cycle. We also gave additional weight to the JM data, as the vehicles
B-19
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
tested in the CRC and API programs were not SFTP-compliant. While the JM test vehicle was
also not SFTP-compliant, the purging of the sulfur from the catalyst was conducted in a way that
was more consistent with that of an SFTP-compliant vehicle. Given this, we project that both
NMHC and NOx emissions from SFTP-compliant vehicles, will be 50 percent reversible. Based
on the average results of the three test programs, we could have projected a higher reversibility
for NMHC emissions and a lower reversibility for NOx emissions. However, examining the
results for the individual vehicles and catalysts and given the fact that none of the vehicles tested
were SFTP-compliant, we decided to project a single level of reversibility for both pollutants.
B-20
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Appendix B. References
1. "Development of Light-Duty Emission Inventory Estimates in the Notice of Proposed
Rulemaking for Tier 2 and Sulfur Standards," U.S. EPA, February 1999.
2. "EPA Staff Paper on Gasoline Sulfur Issues," U.S. EPA, May 1998, EPA420-R-98-005.
3. EPA Report Number M6.FUL.001
4. "EPA Staff Paper on Gasoline Sulfur Issues," U.S. EPA, May 1998, EPA420-R-98-005.
B-21
-------
Appendix C: One-hour and Eight-hour County Design Values
Appendix C: One-Hour and Eight-Hour County Design Values
The tables contained in this appendix are as follows:
Table C-l. Areas formally designated as nonattainment areas for ozone under the 1-hour
NAAQS, outside California, as of August 10, 1998.
Table C-2. Metropolitan areas and rural counties with 1995-1997 measured design values
exceeding the 1-hour standard.
Table C-3. Metropolitan areas and rural counties with 1995-1997 measured design values
exceeding the 8-hour standard.
Table C-4. Metropolitan areas and rural counties with 1995-1997 measured design values
within 15% of the 8-hour standard.
Table C-5. Metropolitan areas and rural counties with design values projected to exceed the
1-hour standard in 2007 with ROTR controls but without Tier 2/Sulfur controls.
Table C-6. Metropolitan areas and rural counties with design values projected to exceed the
8-hour standard in 2007 with ROTR controls but without Tier 2/Sulfur controls.
Table C-7. Metropolitan areas and rural counties with design values projected to be less than
the 8-hour standard but within 15% or it in 2007 with ROTR controls but without
Tier 2/Sulfur controls.
Table C-8. Metropolitan areas and rural counties with design values projected to exceed the
8-hour standard in 2007 with ROTR and Tier 2/Sulfur controls.
Table C-9. Metropolitan areas and rural counties with design values projected to exceed the
8-hour standard in 2010 with ROTR controls but without Tier 2/Sulfur controls.
Table C-10. Metropolitan areas and rural counties with design values projected to exceed the
8-hour standard in 2010 with ROTR controls and Tier 2/Sulfur controls.
Table C-l 1. One-hour county design values (ppb) and population.
Table C-12. Eight-hour county design values (ppb) and population.
Table C-13. Metropolitan areas and rural counties with design values that exceeded the 1-hour
standard prior to ROTR controls, and were projected to meet but remain within
15% of the 1-hour standard after ROTR controls.
C-l
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-l. Areas formally designated as nonattainment areas for ozone under the
1-hourNAAQS, outside California, as of August
Areas Listed Alphabetically
Atlanta, GA
Baltimore, MD
Baton Rouge, LA
Beaumont-Port Arthur, TX
Birmingham, AL
Boston-Lawrence-Worcester (E. MA), MA-NH
Chicago-Gary-Lake County, IL-IN
Cincinnati-Hamilton, OH-KY
Dallas-Fort Worth, TX
Door Co, WI
El Paso, TX
Greater Connecticut
Houston-Galveston-Brazoria, TX
Kent & Queen Anne's Co.s, MD
Lancaster, PA
Louisville, KY-IN
Manitowoc Co, WI
Milwaukee-Racine, WI
Muskegon, MI
New York-N. New Jersey -Long Island, NY-NJ-C
Philadelphia-Wilmington-Trenton, PA-NJ-DE-M
Phoenix, AZ
Pittsburgh-Beaver Valley, PA
Portland, ME
Portsmouth-Dover-Rochester, NH
Providence (All RI), RI
Springfield (Western MA), MA
St Louis, MO-IL
Sunland Park, NM (New Area 1995)
Washington, DC-MD-VA
10, 1998. l
Classification
Serious
Severe- 15
Serious
Moderate
Marginal
Serious
Severe- 17
Moderate
Serious
Marginal
Serious
Serious
Severe- 17
Marginal
Marginal
Moderate
Moderate
Severe- 17
Moderate
Severe- 17
Severe- 15
Serious
Moderate
Moderate
Serious
Serious
Serious
Moderate
Marginal
Serious
1990
Popu-
lation
(1000)
2,654
2,348
559
361
751
5,505
7,886
1,705
3,560
26
592
2,470
3,731
52
423
834
80
1,735
159
17,651
6,010
2,092
2,468
441
183
1,003
812
2,390
8
3,924
Located
in a
CMSA
or MSA?
msa
msa
msa
msa
msa
cmsa
cmsa
cmsa
cmsa
neither
msa
cmsa
cmsa
part
msa
msa
neither
cmsa
msa
cmsa
cmsa
msa
cmsa
msa
msa
cmsa
msa
msa
msa
msa
C-2
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-2. Metropolitan areas and rural counties with 1995-1997 measured design values
exceeding the 1-hour standard.
Name of metropolitan area or rural county
Design 1990 Population
Vdusfab)
Manitowoc WI 126 80,421
DoorWI 127 25,690
Jefferson TN 125 33,016
Mason MI 125 25,537
KentMD 129 17,842
SagadahocME 125 33,535
IbervilleLA 139 31,049
La Porte IN 146 107,066
Atlanta, GA MSA 145 2,959,500
Barnstable-Yarmouth, MA MSA 131 134,954
Baton Rouge, LA MSA 131 528,261
Beaumont-Port Arthur, TX MSA 139 361,218
Birmingham, AL MSA 132 839,942
Boston-Worcester-Lawrence, MA-NH-ME-CT CMSA 138 5,455,403
Chicago-Gary-Kenosha, IL-IN-WI CMSA 129 8,239,820
Cincinnati-Hamilton, OH-KY-IN CMSA 125 1,817,569
Dallas-Fort Worth, TX CMSA 139 4,037,282
Grand Rapids-Muskegon-Holland, MI MSA 137 937,891
Hartford, CT MSA 144 1,157,585
Houma, LAMSA 127 182,842
Houston-Galveston-Brazoria, TX CMSA 189 3,731,029
Kansas City, MO-KS MSA 128 1,582,874
Lancaster, PA MSA 125 422,822
Longview-Marshall, TX MSA 139 193,801
Louisville, KY-IN MSA 125 949,012
Memphis, TN-AR-MS MSA 131 1,007,306
Milwaukee-Racine, WI CMSA 127 1,607,183
New York-Northern New Jersey-Long Island, NY-NJ-CT-PA CMSA 157 19,549,649
Philadelphia-Wilmington-Atlantic City, PA-NJ-DE-MD CMSA 152 5,893,019
Pittsburgh, PA MSA 133 2,394,811
Providence-Fall River-Warwick, RI-MA MSA 133 1,134,350
St. Louis, MO-IL MSA 131 2,492,348
Springfield, MA MSA 132 587,884
Washington-Baltimore, DC-MD-VA-WV CMSA 145 6,726,395
C-2
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-3. Metropolitan areas and rural counties with 1995-1997 measured design values
exceeding the 8-hour standard.
Name of metropolitan area or rural county Design 1990 Population
Vdusfccb)
Manitowoc WI
Kewaunee WI
Door WI
Madison VA
Jefferson TN
Chester SC
Franklin PA
Preble OH
Logan OH
Knox OH
Clinton OH
Northampton NC
Haywood NC
Granville NC
Caswell NC
Jefferson NY
Essex NY
Ste. Genevieve MO
Mason MI
Cass MI
Benzie MI
KentMD
Sagadahoc ME
Knox ME
Iberville LA
Livingston KY
Hancock KY
La Porte IN
Sussex DE
ClayAL
Allentown-Bethlehem-Easton, PA MSA
Altoona, PA MSA
Atlanta, GA MSA
Augusta-Aiken, GA-SC MSA
Barnstable- Yarmouth, MA MSA
Baton Rouge, LA MSA
Beaumont-Port Arthur, TX MSA
Benton Harbor, MI MSA
Biloxi-Gulfport-Pascagoula, MS MSA
95
93
92
86
96
87
86
86
86
91
97
86
85
94
89
88
86
87
98
94
88
96
95
87
96
86
89
104
93
86
95
90
110
87
100
95
93
98
86
80,421
18,878
25,690
11,949
33,016
32,170
121,082
40,113
42,310
47,473
35,415
20,798
46,942
38,345
20,693
110,943
37,152
16,037
25,537
49,477
12,200
17,842
33,535
36,310
31,049
9,062
7,864
107,066
113,229
13,252
595,081
130,542
2,959,500
415,220
134,954
528,261
361,218
161,378
312,368
C-4
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-3. Metropolitan areas and rural counties with 1995-1997 measured design values
exceeding the 8-hour standard.
Name of metropolitan area or rural county Design 1990 Population
Vdusfab)
Birmingham, AL MSA 95 839,942
Boston-Worcester-Lawrence, MA-NH-ME-CT CMSA 97 5,455,403
Buffalo-Niagara Falls, NY MSA 85 1,189,340
Canton-Massillon, OH MSA 88 394,106
Charlotte-Gastonia-Rock Hill, NC-SC MSA 97 1,162,140
Chattanooga, TN-GA MSA 90 424,347
Chicago-Gary-Kenosha, IL-IN-WI CMSA 96 8,239,820
Cincinnati-Hamilton, OH-KY-IN CMSA 99 1,817,569
Cleveland-Akron, OH CMSA 99 2,859,644
Columbus, OH MSA 92 1,345,450
Dallas-Fort Worth, TX CMSA 104 4,037,282
Dayton-Springfield, OH MSA 93 951,270
Detroit-Ann Arbor-Flint, MI CMSA 92 5,187,171
Dover, DE MSA 94 110,993
Elkhart-Goshen, IN MSA 90 156,198
Erie, PA MSA 87 275,572
Evansville-Henderson, IN-KY MSA 93 278,990
Fayetteville, NC MSA 87 274,713
Fort Wayne, IN MSA 90 456,281
Grand Rapids-Muskegon-Holland, MI MSA 99 937,891
Greensboro-Winston-Salem-High Point, NC MSA 89 1,050,304
Greenville, NC MSA 88 108,480
Greenville-Spartanburg-Anderson, SC MSA 88 830,539
Harrisburg-Lebanon-Carlisle, PA MSA 88 587,986
Hartford, CT MSA 101 1,157,585
Houma, LAMSA 85 182,842
Houston-Galveston-Brazoria, TX CMSA 117 3,731,029
Huntington-Ashland, WV-KY-OH MSA 88 312,529
Indianapolis, IN MSA 97 1,380,491
Jamestown, NY MSA 85 141,895
Janesville-Beloit, WI MSA 85 139,510
Johnson City-Kingsport-Bristol, TN-VA MSA 88 436,047
Johnstown, PA MSA 88 241,280
Kalamazoo-Battle Creek, MI MSA 87 429,453
Kansas City, MO-KS MSA 94 1,582,874
Knoxville, TN MSA 95 585,960
Lake Charles, LA MSA 85 168,134
Lancaster, PA MSA 96 422,822
Lexington, KY MSA 85 405,936
Lima, OH MSA 89 154,340
C-5
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-3. Metropolitan areas and rural counties with 1995-1997 measured design values
exceeding the 8-hour standard.
Name of metropolitan area or rural county
Longview-Marshall, TX MSA
Louisville, KY-IN MSA
Memphis, TN-AR-MS MSA
Milwaukee-Racine, WI CMSA
Nashville, TN MSA
New York-Northern New Jersey-Long Island, NY-NJ-CT-PA CMSA
Norfolk- Virginia Beach-Newport News, VA-NC MSA
Owensboro, KY MSA
Parkersburg-Marietta, WV-OH MSA
Philadelphia- Wilmington-Atlantic City, PA-NJ-DE-MD CMSA
Pittsburgh, PA MSA
Portland, ME MSA
Providence-Fall River-Warwick, RI-MA MSA
Raleigh-Durham-Chapel Hill, NC MSA
Reading, PA MSA
Richmond-Petersburg, VA MSA
St. Louis, MO-IL MSA
San Antonio, TX MSA
Scranton—Wilkes-Barre—Hazleton, PA MSA
Sharon, PA MSA
Sheboygan, WI MSA
South Bend, IN MSA
Springfield, MA MSA
Steubenville-Weirton, OH-WV MSA
Terre Haute, IN MSA
Toledo, OH MSA
Tulsa, OK MSA
Tyler, TX MSA
Washington-Baltimore, DC-MD-VA-WV CMSA
Wheeling, WV-OH MSA
York, PA MSA
Youngstown-Warren, OH MSA
Design
Vdefeb)
91
92
95
98
99
108
87
87
89
110
105
94
96
89
92
90
100
87
90
92
92
91
97
85
88
89
88
89
107
86
87
93
1990 Population
193,801
949,012
1,007,306
1,607,183
985,026
19,549,649
1,444,710
87,189
149,169
5,893,019
2,394,811
221,095
1,134,350
858,485
336,523
865,640
2,492,348
1,324,749
638,524
121,003
103,877
247,052
587,884
142,523
147,585
614,128
708,954
151,309
6,726,395
159,301
339,574
600,895
C-6
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-4. Metropolitan areas and rural counties with
within 15% of the 8-hour standard.
Name
Simpson KY
Yancey NC
Wythe VA
Pike KY
Rockingham NC
McLean KY
Person NC
Putnam TN
Monroe MO
Walworth WI
Sharkey MS
Macoupin IL
Ulster NY
St Mary LA
Hardin KY
Henry VA
Union SC
Humphreys TN
Graves KY
Lee MS
Pointe Coupee LA
Haywood TN
Lawrence PA
Frederick VA
Huron MI
Edmonson KY
Jefferson WI
Effingham IL
Grant LA
Hamilton NY
Fond Du Lac WI
Caroline VA
Darlington SC
Camden NC
Dodge WI
Greenbrier WV
Columbia WI
Bell KY
Barnwell SC
1995-1997 measured design
Design Value
(ppb)
83
84
78
74
84
83
84
80
84
83
80
83
82
81
81
82
81
77
74
75
83
82
84
84
82
82
80
81
75
75
79
84
78
83
78
83
81
76
81
values
Pop'n.
15,145
15,419
25,466
72,583
86,064
9,628
30,180
51,373
9,104
75,000
7,066
47,679
165,304
58,086
89,240
56,942
30,337
15,795
33,550
65,581
22,540
19,437
96,246
45,723
34,951
10,357
67,783
31,704
17,526
5,279
90,083
19,217
61,851
5,904
76,559
34,693
45,088
31,506
20,293
C-7
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-4. Metropolitan areas and rural counties with
within 15% of the 8-hour standard.
Name
Hancock ME
Coos NH
Cheshire NH
Bennington VT
Adams MS
Albany-Schenectady-Troy, NY MSA
Appleton-Oshkosh-Neenah, WI MSA
Asheville, NC MSA
Austin-San Marcos, TX MSA
Bangor, ME MSA
Champaign-Urbana, IL MSA
Charleston-North Charleston, SC MSA
Charleston, WV MSA
Clarksville-Hopkinsville, TN-KY MSA
Columbia, SC MSA
Columbus, GA-AL MSA
Corpus Christi, TX MSA
Davenport-Moline-Rock Island, IA-IL MSA
Decatur, AL MSA
Decatur, IL MSA
Elmira, NY MSA
Green Bay, WI MSA
Hickory-Morganton-Lenoir, NC MSA
Huntsville, AL MSA
Jackson, MS MSA
Jacksonville, FL MSA
Lafayette, LA MSA
Lakeland-Winter Haven, FL MSA
Lansing-East Lansing, MI MSA
Little Rock-North Little Rock, AR MSA
Madison, WI MSA
Minneapolis-St. Paul, MN-WI MSA
Mobile, AL MSA
Monroe, LA MSA
Montgomery, AL MSA
New Orleans, LA MSA
Oklahoma City, OK MSA
Orlando, FL MSA
Pensacola, FL MSA
Peoria-Pekin, IL MSA
1995-1997 measured design
Design Value
(ppb)
83
77
74
79
77
81
81
75
81
73
81
76
81
80
83
83
83
78
81
83
73
82
79
84
77
81
84
74
83
81
79
81
79
77
77
84
83
78
84
80
values
Pop'n.
46,948
34,828
70,121
35,845
35,356
861,623
315,121
191,772
846,227
91,629
173,025
506,877
250,454
169,439
453,932
260,862
349,894
350,855
131,556
117,206
95,195
194,594
292,405
293,047
395,396
906,727
345,053
405,382
432,684
513,026
367,085
2,538,776
476,923
142,191
292,517
1,285,262
958,839
1,224,844
344,406
339,172
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-4. Metropolitan areas and rural counties with 1995-1997 measured design values
within 15% of the 8-hour standard.
Name
Pittsfield, MA MSA
Roanoke, VA MSA
Rochester, NY MSA
Rockford, IL MSA
Rocky Mount, NC MSA
Sarasota-Bradenton, FL MSA
Savannah, GA MSA
Shreveport-Bossier City, LA MSA
Springfield, IL MSA
Springfield, MO MSA
Syracuse, NY MSA
Tampa-St. Petersburg-Clearwater, FL MSA
Utica-Rome, NY MSA
Victoria, TX MSA
Wichita, KS MSA
Williamsport, PA MSA
Wilmington, NC MSA
Design Value
(ppb)
77
78
83
79
84
76
72
82
76
78
79
82
72
78
74
73
79
sum
MA count
MA sum
Cnty count
Cnty sum
Pop'n.
r
88,695
224,592
1,062,470
329,676
133,369
489,483
257,899
376,330
189,550
264,346
742,237
2,067,959
316,645
74,361
485,270
118,710
171,269
26,226,237
52
24,306,857
44
1,919,380
C-9
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-5. Metropolitan areas and rural counties with design values projected to exceed the 1-
hour standard in 2007 with ROTR controls but without Tier 2/Sulfur controls.
TVT Design Value _ ,
Name / , N Pop n.
(ppb) *__
IbervilleLA 132 31,049
La Porte IN 131 107,066
Beaumont-Port Arthur, TX MSA 129 361,218
Hartford, CT MSA 125 1,157,585
Houston-Galveston-Brazoria, TX CMSA 175 3,731,029
Longview-Marshall, TX MSA 129 193,801
Memphis, TN-AR-MS MSA 125 1,007,306
New York-Northern New Jersey-Long Island, NY-NJ-CT-PA 136 19,549,649
CMSA
Philadelphia-Wilmington-Atlantic City, PA-NJ-DE-MD CMSA 126 5,893,019
Washington-Baltimore, DC-MD-VA-WV CMSA 126 6,726,395
sum 38,758,117
MA count 8
MA sum 38,620,002
Cnty count 2
Cntysum 138,115
C-10
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-6. Metropolitan areas and rural counties with design values projected to exceed the 8-
hour standard in 2007 with ROTR controls but without Tier 2/Sulfur controls.
Name
Design Value
(ppb)
Pop'n.
Manitowoc WI
Mason MI
Iberville LA
La Porte IN
Atlanta, GA MSA
Barnstable-Yarmouth, MA MSA
Baton Rouge, LA MSA
Beaumont-Port Arthur, TX MSA
Benton Harbor, MI MSA
Biloxi-Gulfport-Pascagoula, MS MSA
Boston-Worcester-Lawrence, MA-NH-ME-CT CMSA
Charlotte-Gastonia-Rock Hill, NC-SC MSA
Chicago-Gary-Kenosha, IL-IN-WI CMSA
Cincinnati-Hamilton, OH-KY-IN CMSA
Cleveland-Akron, OH CMSA
Dallas-Fort Worth, TX CMSA
Grand Rapids-Muskegon-Holland, MI MSA
Hartford, CT MSA
Houston-Galveston-Brazoria, TX CMSA
Kansas City, MO-KS MSA
Knoxville, TN MSA
Memphis, TN-AR-MS MSA
Milwaukee-Racine, WI CMSA
Nashville, TN MSA
New York-Northern New Jersey-Long Island, NY-NJ-CT-PA CMSA
Philadelphia-Wilmington-Atlantic City, PA-NJ-DE-MD CMSA
Pittsburgh, PA MSA
St. Louis, MO-IL MSA
San Antonio, TX MSA
Springfield, MA MSA
Tulsa, OK MSA
Washington-Baltimore, DC-MD-VA-WV CMSA
85
88
91
93
95
89
88
86
86
88
86
94
90
87
86
94
89
88
109
90
85
90
94
88
97
94
95
90
85
87
87
93
80,421
25,537
31,049
107,066
2,959,500
134,954
528,261
361,218
161,378
312,368
5,455,403
1,162,140
8,239,820
1,817,569
2,859,644
4,037,282
937,891
1,157,585
3,731,029
1,582,874
585,960
1,007,306
1,607,183
985,026
19,549,649
5,893,019
2,394,811
2,492,348
1,324,749
587,884
708,954
6,726,395
sum
MA count
MA sum
Cnty count
Cnty sum
79,546,273
28
79,302,200
4
244,073
C-ll
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-7. Metropolitan areas and
the 8-hour standard but within 15%
controls.
Name
Kewaunee WI
Door WI
Jefferson TN
Chester SC
Logan OH
Knox OH
Clinton OH
Haywood NC
Granville NC
Caswell NC
Jefferson NY
Essex NY
Ste. Genevieve MO
Cass MI
Benzie MI
KentMD
Sagadahoc ME
Knox ME
Livingston KY
Sussex DE
ClayAL
Simpson KY
Rockingham NC
Monroe MO
Walworth WI
Sharkey MS
Ulster NY
St Mary LA
Union SC
Pointe Coupee LA
Lawrence PA
Huron MI
Jefferson WI
Fond Du Lac WI
Camden NC
Dodge WI
Columbia WI
Hancock ME
Adams MS
rural counties with design values projected to be less than
or it in 2007 with ROTR controls but without Tier 2/Sulfur
Design Value
(ppb)
84
83
80
78
73
77
80
75
83
74
79
78
77
83
79
82
82
76
72
81
72
72
73
78
77
74
72
76
73
77
74
73
73
73
74
72
76
72
74
Pop'n.
18878
25690
33016
32170
42310
47473
35415
46942
38345
20693
110943
37152
16037
49477
12200
17842
33535
36310
9062
113229
13252
15145
86064
9104
75000
7066
165304
58086
30337
22540
96246
34951
67783
90083
5904
76559
45088
46948
35356
C-12
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-7. Metropolitan areas and rural counties with design values
the 8-hour standard but within 15% or it in 2007 with ROTR controls
controls.
Name
Albany-Schenectady-Troy, NY MSA
Allentown-Bethlehem-Easton, PA MSA
Altoona, PA MSA
Appleton-Oshkosh-Neenah, WI MSA
Augusta-Aiken, GA-SC MSA
Austin-San Marcos, TX MSA
Birmingham, AL MSA
Buffalo-Niagara Falls, NY MSA
Canton-Massillon, OH MSA
Chattanooga, TN-GA MSA
Columbia, SC MSA
Columbus, GA-AL MSA
Columbus, OH MSA
Corpus Christi, TX MSA
Davenport-Moline-Rock Island, IA-IL MSA
Dayton-Springfield, OH MSA
Detroit-Ann Arbor-Flint, MI CMSA
Dover, DE MSA
Elkhart-Goshen, IN MSA
Erie, PA MSA
Evansville-Henderson, IN-KY MSA
Fayetteville, NC MSA
Fort Wayne, IN MSA
Green Bay, WI MSA
Greensboro— Winston-Salem— High Point, NC MSA
Greenville, NC MSA
Greenville-Spartanburg-Anderson, SC MSA
Harrisburg-Lebanon-Carlisle, PA MSA
Houma, LA MSA
Huntsville, AL MSA
Indianapolis, IN MSA
Jacksonville, FL MSA
Jamestown, NY MSA
Janesville-Beloit, WI MSA
Johnson City-Kingsport-Bristol, TN-VA MSA
Johnstown, PA MSA
Kalamazoo-Battle Creek, MI MSA
Lafayette, LA MSA
Lake Charles, LA MSA
projected to be less than
but without Tier 2/Sulfur
Design Value
(ppb)
72
80
75
76
77
77
83
77
76
82
74
75
78
81
74
79
84
79
79
75
80
79
81
75
80
77
79
75
81
73
84
79
74
78
77
74
75
79
77
Pop'n.
861623
595081
130542
315121
415220
846227
839942
1189340
394106
424347
453932
260862
1345450
349894
350855
951270
5187171
110993
156198
275572
278990
274713
456281
194594
1050304
108480
830539
587986
182842
293047
1380491
906727
141895
139510
436047
241280
429453
345053
168134
C-13
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-7. Metropolitan areas and rural counties with design values
the 8-hour standard but within 15% or it in 2007 with ROTR controls
controls.
Name
Lakeland-Winter Haven, FL MSA
Lancaster, PA MSA
Lansing-East Lansing, MI MSA
Lexington, KY MSA
Lima, OH MSA
Little Rock-North Little Rock, AR MSA
Longview-Marshall, TX MSA
Louisville, KY-IN MSA
Madison, WI MSA
Minneapolis-St. Paul, MN-WI MSA
Mobile, AL MSA
Monroe, LA MSA
New Orleans, LA MSA
Norfolk- Virginia Beach-Newport News, VA-NC MSA
Oklahoma City, OK MSA
Orlando, FL MSA
Owensboro, KY MSA
Pensacola, FL MSA
Portland, ME MSA
Providence-Fall River-Warwick, RI-MA MSA
Raleigh-Durham-Chapel Hill, NC MSA
Reading, PA MSA
Richmond-Petersburg, VA MSA
Rochester, NY MSA
Rocky Mount, NC MSA
Sarasota-Bradenton, FL MSA
Scranton—Wilkes-Barre—Hazleton, PA MSA
Sharon, PA MSA
Sheboygan, WI MSA
Shreveport-Bossier City, LA MSA
South Bend, IN MSA
Steubenville-Weirton, OH-WV MSA
Tampa-St. Petersburg-Clearwater, FL MSA
Terre Haute, IN MSA
Toledo, OH MSA
Tyler, TX MSA
Victoria, TX MSA
Wichita, KS MSA
Wilmington, NC MSA
projected to be less than
but without Tier 2/Sulfur
Design Value
(ppb)
72
83
74
75
75
79
84
77
74
79
75
73
80
78
83
78
73
82
83
82
82
78
77
75
76
76
75
81
83
77
80
73
83
72
80
82
75
72
72
Pop'n.
405382
422822
432684
405936
154340
513026
193801
949012
367085
2538776
476923
142191
1285262
1444710
958839
1224844
87189
344406
221095
1134350
858485
336523
865640
1062470
133369
489483
638524
121003
103877
376330
247052
142523
2067959
147585
614128
151309
74361
485270
171269
C-14
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-7. Metropolitan areas and rural counties with design values projected to be less than
the 8-hour standard but within 15% or it in 2007 with ROTR controls but without Tier 2/Sulfur
controls.
TVT Design Value _ ,
Name / , N Pop n.
(ppb) *__
York, PA MSA 74 339574
Youngstown-Warren, OH MSA 8J_ 600895
sum 49,387,949
MA count 80
MA sum 47,630,414
Cnty count 39
Cnty sum 1,757,535
C-15
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-8. Metropolitan areas and rural counties with design values projected to exceed the 8-
hour standard in 2007 with ROTR and Tier 2/Sulfur controls.
Name
Design Value
(ppb)
Pop'n.
Mason MI
Iberville LA
La Porte IN
Atlanta, GA MSA
Barnstable-Yarmouth, MA MSA
Baton Rouge, LA MSA
Beaumont-Port Arthur, TX MSA
Benton Harbor, MI MSA
Biloxi-Gulfport-Pascagoula, MS MSA
Charlotte-Gastonia-Rock Hill, NC-SC MSA
Chicago-Gary-Kenosha, IL-IN-WI CMSA
Cincinnati-Hamilton, OH-KY-IN CMSA
Cleveland-Akron, OH CMSA
Dallas-Fort Worth, TX CMSA
Grand Rapids-Muskegon-Holland, MI MSA
Hartford, CT MSA
Houston-Galveston-Brazoria, TX CMSA
Kansas City, MO-KS MSA
Memphis, TN-AR-MS MSA
Milwaukee-Racine, WI CMSA
Nashville, TN MSA
New York-Northern New Jersey-Long Island, NY-NJ-CT-PA CMSA
Philadelphia-Wilmington-Atlantic City, PA-NJ-DE-MD CMSA
Pittsburgh, PA MSA
St. Louis, MO-IL MSA
Springfield, MA MSA
Tulsa, OK MSA
Washington-Baltimore, DC-MD-VA-WV CMSA
87
90
92
91
87
86
85
85
87
90
90
88
85
92
87
86
109
88
88
93
86
97
94
94
87
85
86
91
25,537
31,049
107,066
2,959,500
134,954
528,261
361,218
161,378
312,368
1,162,140
8,239,820
1,817,569
2,859,644
4,037,282
937,891
1,157,585
3,731,029
1,582,874
1,007,306
1,607,183
985,026
19,549,649
5,893,019
2,394,811
2,492,348
587,884
708,954
6,726,395
sum
MA count
MA sum
Cnty count
Cnty sum
72,099,740
25
71,936,088
O
163,652
C-16
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-9. Metropolitan areas and rural counties with design values
hour standard in 2010 with ROTR controls but without Tier 2/Sulfur
projected to exceed the 8-
controls.
Name
Design Value
(ppb)
Pop'n.
Mason MI
Iberville LA
La Porte IN
Atlanta, GA MSA
Barnstable-Yarmouth, MA MSA
Baton Rouge, LA MSA
Beaumont-Port Arthur, TX MSA
Benton Harbor, MI MSA
Biloxi-Gulfport-Pascagoula, MS MSA
Boston-Worcester-Lawrence, MA-NH-ME-CT CMSA
Charlotte-Gastonia-Rock Hill, NC-SC MSA
Chicago-Gary-Kenosha, IL-IN-WI CMSA
Cincinnati-Hamilton, OH-KY-IN CMSA
Cleveland-Akron, OH CMSA
Dallas-Fort Worth, TX CMSA
Grand Rapids-Muskegon-Holland, MI MSA
Hartford, CT MSA
Houston-Galveston-Brazoria, TX CMSA
Kansas City, MO-KS MSA
Memphis, TN-AR-MS MSA
Milwaukee-Racine, WI CMSA
Nashville, TN MSA
New York-Northern New Jersey-Long Island, NY-NJ-CT-PA CMSA
Philadelphia-Wilmington-Atlantic City, PA-NJ-DE-MD CMSA
Pittsburgh, PA MSA
St. Louis, MO-IL MSA
Springfield, MA MSA
Tulsa, OK MSA
Washington-Baltimore, DC-MD-VA-WV CMSA
87
90
92
94
88
87
85
85
87
85
93
90
87
85
93
88
87
109
89
89
93
87
97
94
94
89
86
86
92
25,537
31,049
107,066
2,959,500
134,954
528,261
361,218
161,378
312,368
5,455,403
1,162,140
8,239,820
1,817,569
2,859,644
4,037,282
937,891
1,157,585
3,731,029
1,582,874
1,007,306
1,607,183
985,026
19,549,649
5,893,019
2,394,811
2,492,348
587,884
708,954
6,726,395
sum
MA count
MA sum
Cnty count
Cnty sum
77,555,143
26
77,391,491
3
163,652
C-17
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-10. Metropolitan areas and rural counties with design values
hour standard in 2010 with ROTR controls and Tier 2/Sulfur controls.
Name
Mason MI
Iberville LA
La Porte IN
Atlanta, GA MSA
Barnstable- Yarmouth, MA MSA
Biloxi-Gulfport-Pascagoula, MS MSA
Charlotte-Gastonia-Rock Hill, NC-SC MSA
Chicago-Gary-Kenosha, IL-IN-WI CMSA
Cincinnati-Hamilton, OH-KY-IN CMSA
Dallas-Fort Worth, TX CMSA
Grand Rapids-Muskegon-Holland, MI MSA
Houston-Galveston-Brazoria, TX CMSA
Kansas City, MO-KS MSA
Memphis, TN-AR-MS MSA
Milwaukee-Racine, WI CMSA
New York-Northern New Jersey-Long Island, NY-NJ-CT-PA CMSA
Philadelphia- Wilmington-Atlantic City, PA-NJ-DE-MD CMSA
Pittsburgh, PA MSA
St. Louis, MO-IL MSA
Tulsa, OK MSA
Washington-Baltimore, DC-MD-VA-WV CMSA
projected to exceed the 8-
Design Value
(ppb)
86
89
91
89
85
86
87
91
88
90
86
110
86
86
92
98
93
93
85
85
90
sum
MA count
MA sum
Cnty count
Cnty sum
Pop'n.
25,537
31,049
107,066
2,959,500
134,954
312,368
1,162,140
8,239,820
1,817,569
4,037,282
937,891
3,731,029
1,582,874
1,007,306
1,607,183
19,549,649
5,893,019
2,394,811
2,492,348
708,954
6,726,395
65,458,744
18
65,295,092
3
163,652
Table C-ll. One-hour county design values (ppb) and population
All 1-hour county design values are listed in this table. The design value headings in the
table are defined as follows:
dvl—measured 1-hour design value 1995-1997.
bgl—projected design value in 2007 after ROTR.
ntllO—interpolated design value in 2010 without Tier 2/Sulfur controls.
t!07—projected design values in 2007 after ROTR and Tier 2/Sulfur (OMS4).
tllO—interpolated design values in 2010 after ROTR and Tier 2/Sulfur.
t!20—projected design values in 2020 after ROTR and Tier 2/Sulfur (OMS3).
u!07—interpolated design values in 2007 after ROTR and Tier 2/Sulfur.
C-18
-------
Appendix C: One-hour and Eight-hour County Design Values
ul 10—interpolated design values in 2010 after ROTR and Tier 2/Sulfur.
u!20—interpolated design values in 2020 after ROTR and Tier 2/Sulfur.
The "u" values correspond to today's proposal. The "t" values correspond to OMS3 and
OMS4. For discussing the effects of today's proposal on design values, we projected design
values using the rollback method for three modeling runs: 2007 post ROTR, OMS3, and OMS4.
All other design values have been linearly interpolated based on NOX. See the main text for
further details.
Table C-ll. One-hour county design values (p
State
AL
AL
AL
AL
AL
AL
AL
AL
AL
AL
AL
AL
AR
AR
AR
AR
CT
CT
CT
CT
CT
CT
CT
DC
DE
DE
DE
FL
County
Clay
Colbert
Elmore
Geneva
Jefferson
Lawrence
Madison
Mobile
Montgomery
Morgan
Shelby
Sumter
Crittenden
Montgomery
Newton
Pulaski
Fairfield
Hartford
Litchfield
Middlesex
New Haven
New London
Tolland
District of Columbia
Kent
New Castle
Sussex
Alachua
dvl
110
83
102
84
132
98
102
111
92
114
127
83
122
79
83
108
138
138
120
135
157
144
127
125
124
139
123
101
bgl
90
69
86
77
117
86
87
105
84
101
111
76
111
74
78
106
124
122
104
116
136
125
111
107
104
116
107
97
ntllO
89
68
85
76
116
85
86
104
83
100
109
75
110
73
77
105
123
121
103
115
135
124
110
106
103
115
106
96
)b) and population
t!07
87
67
83
75
114
83
84
103
81
98
106
73
108
72
76
102
123
121
102
113
133
122
108
106
101
113
104
93
tno
85
66
81
73
112
81
82
101
79
96
103
72
106
71
74
100
122
119
100
111
131
120
106
106
99
111
102
91
t!20
82
64
78
71
108
79
79
98
75
93
98
70
103
69
72
97
121
117
97
109
129
117
104
106
95
107
98
86
u!07
87
67
83
75
114
83
84
103
81
98
106
73
108
72
76
102
123
121
102
113
133
122
108
106
101
113
104
93
ullO
85
66
81
73
112
81
82
101
79
96
103
72
106
71
74
100
122
119
100
111
131
120
106
106
99
111
102
91
u!20
82
64
78
71
108
79
79
98
75
93
98
70
103
69
72
97
121
117
97
109
129
117
104
106
95
107
98
86
Pop'n.
75,000
25,617
46,975
34,773
31,679
80,421
18,878
67,783
90,083
4,590
25,690
76,559
45,088
34,693
25,466
11,949
56,942
45,723
19,217
35,845
51,373
35,303
33,016
15,795
19,437
50,480
25,741
34,854
C-19
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-ll. One-hour county design values (p
State
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
IA
IA
IA
IA
County
Brevard
Broward
Duval
Escambia
Hillsborough
Lee
Leon
Manatee
Orange
Osceola
Palm Beach
Pasco
Pinellas
Polk
Sarasota
Seminole
St Johns
St Lucie
Volusia
Bibb
Chatham
DeKalb
Douglas
Fannin
Fulton
Glynn
Gwinnett
Muscogee
Paulding
Richmond
Rockdale
Spalding
Harrison
Linn
Palo Alto
Polk
dvl
86
103
116
113
112
88
96
96
106
96
89
92
93
99
99
95
91
82
89
122
85
136
140
92
143
89
121
108
112
118
145
126
79
75
70
86
b^l
84
103
114
111
113
87
90
97
107
95
83
89
91
96
99
92
88
80
87
107
78
117
122
81
124
84
103
98
97
105
123
109
77
71
68
82
ntllO
83
103
113
110
112
86
89
96
106
94
82
88
90
95
98
91
87
79
86
105
77
115
120
80
122
83
101
96
95
104
121
107
76
70
67
81
pb) and population
t!07
82
103
110
109
111
85
86
95
103
92
80
86
89
93
96
89
85
78
83
102
76
111
116
78
119
82
97
93
92
101
117
103
76
70
67
80
tno
80
103
108
107
109
84
83
93
100
90
78
84
87
91
94
87
83
76
81
99
74
107
112
76
116
80
94
90
89
98
114
100
75
69
66
79
t!20
77
103
103
105
106
82
78
90
94
86
75
81
84
88
92
82
79
73
77
92
72
100
105
72
109
78
87
84
83
93
107
93
74
68
65
77
u!07
82
103
110
109
111
85
86
95
103
92
80
86
89
93
96
89
85
78
83
102
76
111
116
78
119
82
97
93
92
101
117
103
76
70
67
80
ullO
80
103
108
107
109
84
83
93
100
90
78
84
87
91
94
87
83
76
81
99
74
108
113
76
116
80
94
90
89
98
114
100
75
69
66
79
u!20
77
103
103
105
106
82
78
90
94
86
75
81
84
88
92
82
79
73
77
92
72
100
105
72
109
78
87
84
83
93
107
93
74
68
65
77
Pop'n.
40,339
73,712
36,815
30,337
57,494
61,851
32,170
20,293
23,862
96,246
121,082
36,490
10,333
31,969
40,113
42,310
47,473
35,415
15,419
11,268
86,064
30,180
20,798
25,078
46,942
38,345
39,995
20,693
5,904
165,304
110,943
5,279
37,152
38,592
74,929
34,828
C-20
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-ll. One-hour county design values (p
State
IA
IA
IA
IA
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
County
Scott
Story
Van Buren
Warren
Adams
Champaign
Cook
Du Page
Effingham
Hamilton
Jersey
Kane
Lake
Macon
Macoupin
Madison
McHenry
Peoria
Randolph
Rock Island
Sangamon
St Clair
Will
Winnebago
Allen
Clark
De Kalb
Elkhart
Floyd
Hamilton
Hancock
Johnson
Kosciusko
La Porte
Lake
Lawrence
dvl
95
87
82
74
89
94
127
103
97
89
112
116
116
100
102
128
108
95
94
83
98
108
108
93
106
125
82
113
125
116
120
102
100
146
117
100
b^l
90
83
76
70
80
77
124
95
79
70
98
102
108
83
85
108
97
82
82
76
83
97
92
83
95
104
73
100
111
102
105
84
87
131
105
75
ntllO
89
82
75
69
79
76
124
94
78
69
97
101
107
82
84
107
96
81
81
75
82
96
91
82
94
103
72
99
111
101
104
83
86
130
104
74
pb) and population
t!07
88
81
74
69
78
75
125
94
77
69
94
101
107
80
82
105
96
80
80
75
80
95
90
82
93
102
71
98
111
100
104
83
85
129
104
73
tno
87
80
73
68
77
74
125
93
76
68
92
99
106
78
80
103
94
79
79
74
78
93
89
81
92
101
70
97
110
99
103
82
84
127
102
72
t!20
86
78
71
66
75
72
127
92
74
66
87
97
104
76
77
98
92
77
77
73
76
90
87
80
90
99
69
95
110
98
102
80
82
125
100
71
u!07
88
81
74
69
78
75
124
94
77
69
94
101
107
80
82
105
96
80
80
75
80
95
90
82
93
102
71
98
111
100
104
83
85
129
104
73
ullO
87
80
73
68
77
74
125
93
76
68
92
99
106
78
80
103
94
79
79
74
78
93
89
81
92
101
70
97
110
99
103
82
84
127
102
72
u!20
86
78
71
66
75
72
126
92
74
66
87
97
104
76
77
98
92
77
77
73
76
90
87
80
90
99
69
95
110
98
102
80
82
125
100
71
Pop'n.
70,121
35,410
49,216
16,037
9,104
47,880
7,066
65,581
75,555
8,377
9,071
35,356
10,415
19,776
37,308
25,537
34,951
49,477
12,200
17,842
35,308
49,767
33,535
18,653
52,602
36,310
115,904
46,948
58,086
22,540
31,049
17,526
30,083
10,361
15,145
49,489
C-21
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-ll. One-hour county design values (p
State
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
KS
KS
KS
KS
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
County
Madison
Marion
Morgan
Porter
Posey
St Joseph
Tippecanoe
Vanderburgh
Vigo
Warrick
Miami
Pawnee
Sedgwick
Wyandotte
Bell
Boone
Boyd
Bullitt
Campbell
Christian
Daviess
Edmonson
Fayette
Graves
Greenup
Hancock
Hardin
Henderson
Jefferson
Jessamine
Kenton
Lawrence
Livingston
McCracken
McLean
Oldham
dvl
112
115
103
124
99
114
104
114
107
115
114
80
96
113
92
108
122
116
115
101
108
118
101
92
114
114
113
108
120
98
114
95
108
100
103
112
b^l
98
99
83
109
83
100
85
97
88
92
112
80
93
110
74
90
98
94
99
80
89
96
90
73
90
82
86
85
100
81
97
72
91
81
83
97
ntllO
97
98
82
108
82
99
84
96
87
91
111
80
92
109
73
89
97
93
98
79
88
95
89
72
89
81
85
84
100
80
96
71
90
80
82
96
pb) and population
t!07
96
98
82
107
81
98
84
95
86
91
110
80
92
109
72
89
96
92
97
78
88
94
89
72
89
80
84
84
100
79
95
70
89
79
82
95
tno
94
97
81
105
80
97
83
94
85
89
109
80
91
108
70
88
94
90
95
77
87
92
88
71
87
79
82
83
99
78
94
69
88
78
81
94
t!20
92
95
79
103
79
96
81
92
83
87
107
80
89
107
68
86
92
88
93
75
85
90
87
69
85
77
80
81
99
76
92
67
86
76
79
92
u!07
96
98
82
107
81
98
84
95
86
91
110
80
92
109
72
89
96
92
97
78
88
94
89
72
89
80
84
84
100
79
95
70
89
79
82
95
ullO
94
97
81
105
80
97
83
94
85
89
109
80
91
108
70
88
94
90
95
77
87
92
88
71
87
79
82
83
99
78
94
69
88
78
81
94
u!20
92
95
79
103
79
96
81
92
83
87
107
80
89
107
68
86
92
88
93
75
85
90
87
69
85
77
80
81
99
76
92
67
86
76
79
92
Pop'n.
72,583
30,283
9,628
62,879
9,062
13,998
89,240
7,864
33,550
10,357
31,506
7,555
7,676
74,252
10,669
14,730
42,836
107,066
65,294
34,583
47,679
8,499
31,704
66,090
62,496
15,992
113,229
7,666
7,841
16,174
23,647
13,252
292,594
181,276
149,285
247,105
C-22
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-ll. One-hour county design values (p
State
KY
KY
KY
KY
KY
KY
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
MA
MA
MA
MA
MA
MA
MA
MA
MA
County
Perry
Pike
Pulaski
Scott
Simpson
Trigg
Ascension
Beauregard
Bossier
Caddo
Calcasieu
East Baton Rouge
Grant
Iberville
Jefferson
Lafayette
Lafourche
Livingston
Orleans
Ouachita
Pointe Coupee
St Bernard
St Charles
St James
St John The Baptist
StMary
West Baton Rouge
Barnstable
Berkshire
Bristol
Essex
Hampden
Hampshire
Middlesex
Plymouth
Suffolk
dvl
90
98
94
101
101
101
121
117
98
101
116
131
91
139
107
109
127
127
96
95
111
98
115
119
114
104
114
131
99
138
113
126
132
109
102
95
b^l
77
83
76
89
87
85
113
106
92
95
104
122
84
132
103
102
122
118
89
90
103
91
110
112
107
98
106
118
85
119
100
115
118
95
89
86
ntllO
76
82
75
88
86
84
112
105
91
94
103
121
83
131
102
101
121
117
88
89
102
90
109
111
106
97
105
117
84
118
99
114
117
94
88
86
pb) and population
t!07
74
80
74
86
84
83
111
105
90
93
103
118
83
130
102
100
120
115
88
88
101
90
108
110
106
97
103
115
83
116
98
114
115
93
87
86
tno
73
78
73
84
82
81
109
104
89
91
102
116
82
129
101
99
119
113
87
87
99
89
107
109
105
96
101
113
81
114
96
112
113
91
85
85
t!20
71
76
71
82
80
79
107
103
87
89
100
112
80
127
100
97
118
110
85
86
97
88
107
107
103
94
97
110
79
110
93
110
110
89
83
84
u!07
74
80
74
86
84
83
111
105
90
93
103
118
83
130
102
100
120
115
88
88
101
90
108
110
106
97
103
115
83
116
98
114
115
93
87
86
ullO
73
78
73
84
82
81
109
104
89
91
102
116
82
129
101
99
119
113
87
87
99
89
107
109
105
96
101
113
81
114
96
112
113
91
85
85
u!20
71
76
71
82
80
79
107
103
87
89
100
112
80
127
100
97
118
110
85
86
97
88
107
107
103
94
97
110
79
110
93
110
110
89
83
84
Pop'n.
291,130
130,542
140,510
140,320
174,821
54,091
648,951
71,120
545,837
54,457
352,910
41,611
189,719
120,940
18,375
576,407
146,601
186,605
380,105
70,526
58,214
19,419
239,397
80,509
161,378
115,243
31,760
651,525
99,358
506,325
245,845
164,587
456,310
670,080
336,073
1,398,468
C-23
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-ll. One-hour county design values (p
State
MA
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
ME
ME
ME
ME
ME
ME
ME
ME
ME
ME
ME
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
County
Worcester
Anne Arundel
Baltimore
Baltimore City
Calvert
Carroll
Cecil
Charles
Harford
Kent
Montgomery
Prince George's
Cumberland
Hancock
Kennebec
Knox
Oxford
Pen ob scot
Piscataquis
Sagadahoc
Somerset
Washington
York
Allegan
Benzie
Berrien
Cass
Clinton
Genesee
Huron
Ingham
Kalamazoo
Kent
Lenawee
Macomb
Mason
dvl
108
142
130
137
105
115
152
118
145
129
118
132
121
115
98
119
79
95
80
125
92
107
126
137
108
119
115
88
99
110
97
106
124
104
124
125
b^l
93
123
112
122
85
94
126
98
126
110
98
113
107
100
87
104
71
84
72
109
83
93
113
120
97
103
102
77
89
98
87
92
111
94
111
113
ntllO
92
122
111
121
84
93
125
97
125
109
97
112
106
99
86
103
70
83
71
108
82
92
112
119
96
102
101
76
88
97
86
91
110
93
111
112
pb) and population
t!07
90
120
109
120
82
91
123
95
122
107
96
110
105
97
85
101
69
82
70
107
80
90
110
118
95
101
100
75
88
96
85
90
109
93
111
112
tno
88
118
107
118
80
89
121
93
120
105
94
108
103
95
83
99
67
80
68
105
78
88
108
117
93
100
98
74
87
95
84
88
107
91
111
110
t!20
85
114
103
116
78
87
116
91
116
102
91
105
100
91
80
96
64
78
65
101
75
85
106
115
91
98
96
73
86
93
83
86
105
89
111
108
u!07
90
120
109
120
82
91
123
95
122
107
96
110
105
97
85
101
69
82
70
107
80
90
110
118
95
101
100
75
88
96
85
90
109
93
111
112
ullO
88
118
107
118
80
89
121
93
120
105
94
108
103
95
83
99
67
80
68
105
78
88
108
117
93
100
98
74
87
95
84
88
107
91
111
110
u!20
85
114
103
116
78
87
116
91
116
102
91
105
100
91
80
96
64
78
65
101
75
85
106
115
91
98
96
73
86
93
83
86
105
89
111
108
Pop'n.
709,705
435,276
104,233
120,005
663,906
260,120
220,756
968,532
131,761
367,585
168,767
173,025
295,039
128,776
207,619
511,433
110,605
131,497
50,319
285,536
128,181
5,105,067
128,932
475,594
317,471
516,418
183,241
357,313
781,666
291,479
113,909
866,228
150,187
83,866
142,031
57,589
C-24
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-ll. One-hour county design values (p
State
MI
MI
MI
MI
MI
MI
MI
MI
MN
MN
MN
MN
MN
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MS
MS
MS
MS
MS
MS
MS
MS
MS
MS
MS
MS
NC
County
Mecosta
Muskegon
Oakland
Ottawa
Roscommon
St Clair
Washtenaw
Wayne
Anoka
Dakota
Lake
St Louis
Washington
Clay
Greene
Jackson
Jefferson
Monroe
Platte
St Louis
St Louis City
St. Charles
Ste Genevieve
Adams
Choctaw
DeSoto
Franklin
Hancock
Hinds
Jackson
Lauderdale
Lee
Madison
Sharkey
Warren
Alexander
dvl
124
136
117
113
99
119
104
114
106
91
74
80
103
128
101
88
125
96
116
119
108
131
108
97
81
131
94
105
97
109
92
96
89
95
97
94
b^l
108
121
105
101
87
108
94
104
103
88
74
80
100
122
90
85
113
89
112
106
99
117
96
94
73
125
91
100
89
113
83
88
82
89
91
85
ntllO
107
120
104
100
86
107
93
103
103
88
74
80
99
121
89
84
112
88
111
105
98
116
95
93
72
124
90
99
88
112
82
87
81
88
90
84
pb) and population
t!07
105
120
104
100
86
106
92
103
103
88
74
80
99
120
87
83
110
87
110
103
97
113
94
92
71
124
89
98
87
111
81
85
80
88
90
81
tno
103
118
103
99
85
104
90
102
102
87
74
80
98
118
85
82
108
85
109
100
96
111
92
91
70
123
88
97
85
110
79
83
79
87
89
79
t!20
101
116
103
98
83
102
88
101
101
87
74
80
96
115
81
80
106
83
107
95
94
107
90
89
68
121
86
95
83
109
76
81
78
85
87
74
u!07
105
120
104
100
86
106
92
103
103
88
74
80
99
120
87
83
110
87
110
103
97
113
94
92
71
124
89
98
87
111
81
85
80
88
90
81
ullO
103
118
103
99
85
104
90
102
102
87
74
80
98
118
85
82
108
85
109
100
96
111
92
91
70
123
88
97
85
110
79
83
79
87
89
79
u!20
101
116
103
98
83
102
88
101
101
87
74
80
96
115
81
80
106
83
107
95
94
107
90
89
68
121
86
95
83
109
76
81
78
85
87
74
Pop'n.
68,941
215,499
142,585
514,990
81,129
122,354
1,412,140
99,821
271,126
285,720
179,278
128,300
37,068
961,437
66,929
291,145
273,525
1,852,810
1,170,103
264,036
85,167
150,979
148,723
147,548
573,809
136,731
93,182
370,712
100,043
31,513
117,206
327,140
36,033
717,400
145,607
1,083,592
C-25
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-ll. One-hour county design values (p
State
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
ND
NE
NE
NH
NH
NH
NH
NH
County
Buncombe
Caldwell
Camden
Caswell
Chatham
Cumberland
Davie
Duplin
Durham
Edgecombe
Forsyth
Franklin
Granville
Guilford
Haywood
Johnston
Lincoln
Martin
Mecklenburg
New Hanover
Northampton
Person
Pitt
Rockingham
Rowan
Swain
Wake
Yancey
Cass
Douglas
Lancaster
Belknap
Carroll
Cheshire
Coos
Grafton
dvl
86
97
93
111
103
106
105
89
103
102
115
110
116
109
107
107
105
90
123
102
100
100
104
113
122
78
118
108
75
88
69
89
88
91
101
77
b^l
78
87
82
93
84
96
92
80
91
91
105
93
102
98
96
91
93
78
121
93
83
82
90
100
107
66
110
92
75
86
67
80
79
80
91
69
ntllO
77
86
81
92
83
95
91
79
89
90
104
91
100
96
95
90
92
77
119
92
82
81
89
99
106
65
108
91
75
85
66
79
78
79
90
68
pb) and population
t!07
75
83
80
90
80
92
88
77
86
87
101
88
97
93
92
87
89
75
116
91
80
79
87
96
103
64
105
89
75
85
66
79
76
78
88
67
tno
73
81
78
88
78
89
86
75
83
85
98
85
94
90
90
85
87
73
113
89
78
77
85
93
100
62
101
87
75
84
65
77
75
76
86
66
t!20
69
76
75
83
74
84
81
72
78
80
93
80
87
85
85
80
82
70
106
85
74
73
80
88
94
60
94
84
75
82
65
75
73
73
83
64
u!07
75
83
80
90
80
92
88
77
86
87
101
88
97
93
92
87
89
75
116
91
80
79
87
96
103
64
105
89
75
85
66
79
76
78
88
67
ullO
73
81
78
88
78
89
86
75
83
85
98
85
94
90
90
85
87
73
113
89
78
77
85
93
100
62
102
87
75
84
65
77
75
76
86
66
u!20
69
76
75
83
74
84
81
72
78
80
93
80
87
85
85
80
82
70
106
85
74
73
80
88
94
60
94
84
75
82
65
75
73
73
83
64
Pop'n.
2,111,687
91,476
282,937
430,459
110,993
198,213
156,198
95,195
275,572
44,920
165,058
43,044
25,968
102,874
274,566
51,666
335,113
150,171
300,836
35,324
181,596
90,509
158,983
500,631
187,768
194,594
265,878
347,420
27,859
107,924
226,800
145,196
93,894
44,506
237,813
41,172
C-26
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-ll. One-hour county design values (p
State
NH
NH
NH
NH
NH
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
County
Hillsborough
Merrimack
Rockingham
Strafford
Sullivan
Atlantic
Bergen
Camden
Cumberland
Essex
Gloucester
Hudson
Hunterdon
Mercer
Middlesex
Monmouth
Morris
Ocean
Union
Albany
Bronx
Chautauqua
Chemung
Dutchess
Erie
Essex
Hamilton
Herkimer
Jefferson
Kings
Madison
Monroe
New York
Niagara
Oneida
Onondaga
dvl
111
98
130
101
90
124
122
137
115
114
128
120
119
131
139
138
124
149
109
105
123
104
88
113
91
101
97
88
110
124
89
102
121
102
95
102
b^l
98
88
117
91
81
106
108
119
96
98
112
113
102
113
121
121
105
125
94
94
121
90
78
97
79
91
86
78
99
110
79
93
114
93
85
90
ntllO
97
87
116
90
80
105
107
118
95
97
111
112
101
112
120
120
104
124
93
93
121
89
77
96
78
90
85
77
98
109
78
92
113
92
84
89
pb) and population
t!07
95
87
114
89
79
103
106
116
93
97
109
112
100
110
119
118
103
122
92
91
121
88
76
94
78
88
84
76
97
108
77
91
113
92
83
89
tno
93
86
112
87
77
101
105
114
91
96
107
111
98
108
117
116
101
120
91
89
121
86
75
92
77
86
82
74
95
106
75
90
112
91
81
87
t!20
91
84
110
85
75
98
103
111
87
94
104
111
95
106
114
113
98
115
89
87
121
84
73
89
75
84
79
72
92
104
73
88
112
90
79
85
u!07
95
87
114
89
79
103
106
116
93
97
109
112
100
110
119
118
103
122
92
91
121
88
76
94
78
88
84
76
97
108
77
91
113
92
83
89
ullO
93
86
112
87
77
101
105
114
91
96
107
111
98
108
117
116
101
120
91
89
121
86
75
92
77
86
82
74
95
106
75
90
112
91
81
87
u!20
91
84
110
85
75
98
103
111
87
94
104
111
95
106
114
113
98
115
89
87
121
84
73
89
75
84
79
72
92
104
73
88
112
90
79
85
Pop'n.
254,957
851,783
143,196
128,699
174,092
70,709
27,544
85,860
2,818,199
217,399
191,707
51,150
96,827
36,742
61,834
238,912
45,527
108,936
797,159
130,669
55,920
88,109
254,441
53,794
77,982
672,971
83,829
141,895
139,510
143,596
163,029
223,411
153,411
57,867
23,466
161,993
C-27
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-ll. One-hour county design values (p
State
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
County
Orange
Putnam
Queens
Richmond
Saratoga
Schenectady
Suffolk
Ulster
Wayne
Westchester
Allen
Ashtabula
Butler
Clark
Clermont
Clinton
Cuyahoga
Delaware
Franklin
Geauga
Greene
Hamilton
Jefferson
Knox
Lake
Lawrence
Licking
Logan
Lorain
Lucas
Madison
Mahoning
Medina
Miami
Montgomery
Portage
dvl
115
122
125
137
101
94
138
97
102
121
106
105
125
118
116
121
108
99
107
112
111
119
111
113
119
113
115
100
101
111
112
109
110
110
112
114
b^l
99
105
119
121
91
84
120
84
91
104
89
89
104
99
99
100
94
85
88
100
93
112
93
96
103
87
97
84
91
100
93
95
97
93
94
101
ntllO
98
104
119
120
90
83
119
83
90
103
88
88
103
98
98
99
93
84
87
99
92
112
92
95
102
86
96
83
90
99
92
94
96
92
93
100
pb) and population
t!07
97
102
120
119
88
82
118
81
89
102
88
87
102
97
98
98
92
84
86
99
91
112
91
94
101
86
95
82
89
98
91
93
95
91
93
98
tno
95
100
120
117
87
80
116
79
88
100
87
86
100
96
97
96
90
83
85
97
90
112
89
93
99
84
93
80
88
97
89
91
93
90
92
96
t!20
93
97
121
114
85
78
112
77
86
98
85
84
98
94
95
94
88
81
83
95
88
113
87
91
97
82
91
78
86
95
87
88
91
88
90
94
u!07
97
102
119
119
88
82
118
81
89
102
88
87
102
97
98
98
92
84
86
99
91
111
91
94
101
86
95
82
89
98
91
93
95
91
93
98
ullO
95
100
120
117
87
80
116
79
88
100
87
86
100
96
97
96
90
83
85
97
90
112
89
93
99
84
93
80
88
97
89
91
93
90
92
96
u!20
93
97
120
114
85
78
112
77
86
98
85
84
98
94
95
94
88
81
83
95
88
112
87
91
97
82
91
78
86
95
87
88
91
88
90
94
Pop'n.
633,232
85,969
335,749
31,255
51,043
68,250
164,762
130,598
168,134
405,382
422,822
281,912
57,883
133,239
111,486
225,366
23,867
30,508
109,755
213,641
349,660
104,948
87,777
64,404
664,937
47,567
33,263
149,967
367,085
383,545
398,978
67,910
826,330
49,939
1,255,488
72,831
C-28
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-ll. One-hour county design values (p
State
OH
OH
OH
OH
OH
OH
OH
OH
OK
OK
OK
OK
OK
OK
OK
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
County
Preble
Stark
Summit
Trumbull
Union
Warren
Washington
Wood
Cleveland
Comanche
Latimer
Me Clain
Oklahoma
Okmulgee
Tulsa
Allegheny
Beaver
Berks
Blair
Bucks
Cambria
Dauphin
Delaware
Erie
Franklin
Lackawanna
Lancaster
Lawrence
Lehigh
Luzerne
Lycoming
Mercer
Montgomery
Northampton
Perry
Philadelphia
dvl
110
107
113
109
75
124
110
94
102
92
100
95
110
94
121
133
105
118
114
137
102
113
126
105
114
110
125
101
114
110
91
111
122
116
103
130
b^l
88
93
99
95
63
104
84
82
101
86
95
91
110
92
121
121
93
101
95
115
86
97
110
89
92
96
109
90
96
90
75
98
104
98
86
120
ntllO
87
92
98
94
62
103
83
81
100
85
94
90
109
91
120
120
92
100
94
114
85
96
109
88
91
95
108
89
95
89
74
97
103
97
85
119
pb) and population
t!07
86
91
97
92
62
102
82
80
98
84
93
89
108
90
120
120
91
98
92
112
84
93
107
87
90
94
106
87
93
88
74
95
101
96
84
119
tno
85
90
95
90
61
100
80
79
96
83
92
87
106
89
119
118
90
96
90
110
82
91
105
85
89
92
104
85
91
86
73
93
99
94
82
118
t!20
83
88
93
87
59
97
78
78
94
81
90
85
104
87
117
116
89
93
88
106
80
86
101
83
87
89
100
82
88
84
71
90
96
91
79
117
u!07
86
91
97
92
62
102
82
80
98
84
93
89
108
90
120
120
91
98
92
112
84
93
107
87
90
94
106
87
93
88
74
95
101
96
84
119
ullO
85
90
95
90
61
100
80
79
96
83
92
87
106
89
119
118
90
96
90
110
82
91
105
85
89
92
104
85
91
86
73
93
99
94
82
118
u!20
83
88
93
87
59
97
78
78
94
81
90
85
104
87
117
116
89
93
88
106
80
86
101
83
87
89
100
82
88
84
71
90
96
91
79
117
Pop'n.
959,275
175,034
304,715
95,328
243,641
145,896
275,227
50,251
378,643
142,191
49,210
209,085
103,281
35,061
510,784
81,021
67,675
118,570
20,879
42,437
39,996
448,306
66,631
496,938
804,219
433,203
671,780
827,645
553,124
1,321,864
378,977
325,824
1,951,598
421,353
2,300,664
1,203,789
C-29
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-ll. One-hour county design values (p
State
PA
PA
PA
RI
RI
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
County
Washington
Westmoreland
York
Kent
Providence
Abbeville
Aiken
Anderson
Barnwell
Berkeley
Charleston
Cherokee
Chester
Darlington
Edgefield
Oconee
Pickens
Richland
Spartanburg
Union
Williamsburg
York
Anderson
Blount
Bradley
Coffee
Davidson
Dickson
Dyer
Giles
Hamblen
Hamilton
Haywood
Humphreys
Jefferson
Knox
dvl
117
125
109
133
117
93
104
114
99
94
102
106
107
94
93
92
107
107
117
98
85
114
110
124
106
105
110
120
112
104
96
113
97
102
125
120
b^l
90
107
93
114
100
82
92
103
85
84
94
93
97
82
81
83
98
96
106
89
74
103
93
110
92
87
102
101
100
93
80
104
84
86
105
110
ntllO
89
106
92
113
99
81
91
101
84
83
93
92
96
81
80
82
96
95
105
88
73
102
92
109
91
86
101
100
99
92
79
103
83
85
104
109
pb) and population
t!07
88
104
91
110
97
79
89
98
83
82
91
89
93
79
78
79
93
92
102
85
72
99
90
106
88
85
101
99
97
91
77
101
82
84
101
107
tno
86
102
89
108
95
77
87
95
81
80
89
86
91
77
76
76
90
89
99
83
70
96
88
104
86
83
100
97
95
89
75
98
80
82
99
105
t!20
84
99
86
105
92
73
82
90
77
77
87
81
86
73
71
71
85
83
92
79
68
91
85
100
81
80
99
94
92
85
73
93
78
80
95
102
u!07
88
104
91
110
97
79
89
98
83
82
91
89
93
79
78
79
93
92
102
85
72
99
90
106
88
85
101
99
97
91
77
101
82
84
101
107
ullO
86
102
89
108
95
77
87
95
81
80
89
86
91
77
76
76
90
89
99
83
70
96
88
104
86
83
100
97
95
89
75
98
80
82
99
105
u!20
84
99
86
105
92
73
82
90
77
77
87
81
86
73
71
71
85
83
92
79
68
91
85
100
81
80
99
94
92
85
73
93
78
80
95
102
Pop'n.
825,380
83,941
1,487,536
874,866
553,099
107,776
307,647
778,206
259,462
493,819
133,793
52,141
599,611
174,253
22,795
416,444
677,491
107,728
287,529
87,189
86,915
62,254
262,798
182,827
71,347
441,946
502,824
541,174
1,585,577
230,082
547,651
224,327
678,111
138,053
1,336,449
370,321
C-30
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-ll. One-hour county design values (p
State
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
VA
VA
VA
VA
VA
VA
County
Lawrence
Loudon
Madison
Putnam
Rutherford
Sevier
Shelby
Sullivan
Sumner
Williamson
Wilson
Bexar
Brazoria
Cameron
Collin
Dallas
Denton
Ellis
Galveston
Gregg
Harris
Hidalgo
Jefferson
Nueces
Orange
Smith
Tarrant
Travis
Victoria
Webb
Alexandria City
Arlington
Caroline
Charles City
Chesterfield
Fairfax
dvl
93
112
64
99
95
111
128
111
124
110
108
121
148
81
132
134
139
118
182
139
189
78
139
115
121
109
133
112
95
90
124
123
109
119
114
124
b^l
81
94
55
83
81
96
121
99
108
99
97
118
139
81
120
123
124
112
171
129
175
77
129
112
112
101
120
106
91
90
108
107
92
102
97
107
ntllO
80
93
54
82
80
95
120
98
107
98
96
117
138
81
119
122
123
111
170
128
175
77
128
111
111
100
119
105
90
90
108
107
91
101
96
106
pb) and population
t!07
78
91
53
80
78
93
118
97
105
96
94
116
137
81
118
121
121
110
169
127
176
77
127
111
111
98
117
102
90
90
108
107
88
99
94
104
tno
76
89
52
78
76
91
116
95
103
94
92
114
136
81
116
119
119
108
168
125
176
77
126
111
110
96
115
100
89
90
107
106
86
97
92
102
t!20
74
85
50
76
74
87
112
93
100
91
88
111
134
81
112
116
114
106
166
122
177
77
124
111
108
94
112
97
89
90
106
105
81
94
88
100
u!07
78
91
53
80
78
93
118
97
105
96
94
116
137
81
118
121
121
110
169
127
175
77
127
111
111
98
117
102
90
90
108
107
88
99
94
104
ullO
76
89
52
78
76
91
116
95
103
94
92
114
136
81
116
119
119
108
168
125
176
77
126
111
110
96
115
100
89
90
107
106
86
97
92
102
u!20
74
85
50
76
74
87
112
93
100
91
88
111
134
81
112
116
114
106
166
122
176
77
124
111
108
94
112
97
89
90
106
105
81
94
88
100
Pop'n.
204,584
186,093
139,352
243,135
161,135
596,270
423,380
36,414
81,306
38,759
181,835
336,523
63,306
6,282
217,881
209,274
79,332
713,968
89,123
252,913
56,558
212,907
249,238
171,380
993,529
20,539
262,852
396,685
1,185,394
277,776
211,707
216,935
219,039
328,149
121,003
103,877
C-31
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-ll. One-hour county design values (p
State
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
VT
VT
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
County
Fauquier
Frederick
Hampton City
Hanover
Henrico
Henry
Madison
Prince William
Roanoke
Stafford
Suffolk City
Wythe
Bennington
Chittenden
Brown
Columbia
Dane
Dodge
Door
Florence
Fond Du Lac
Jefferson
Kenosha
Kewaunee
Manitowoc
Marathon
Milwaukee
Oneida
Outagamie
Ozaukee
Polk
Racine
Rock
Sauk
Sheboygan
St Croix
dvl
97
102
109
124
115
101
99
110
94
110
108
94
99
85
108
104
97
93
127
80
96
94
129
121
126
84
126
78
98
127
85
119
103
90
123
88
b^l
75
80
98
107
101
82
81
91
81
91
91
72
89
77
99
97
90
85
113
74
88
86
117
109
113
79
118
74
92
118
82
108
95
85
109
85
ntllO
74
79
97
105
99
81
80
90
80
90
90
71
88
76
98
96
89
84
112
73
87
85
116
108
112
78
117
73
91
117
81
107
94
84
108
84
pb) and population
t!07
73
78
96
102
96
79
79
88
78
88
88
70
87
75
97
95
89
83
111
73
87
84
114
106
110
77
116
72
90
117
80
106
93
83
107
84
tno
71
76
94
99
94
77
77
86
76
86
86
68
85
74
95
93
88
82
109
72
86
83
112
104
108
76
114
71
89
116
79
104
92
82
105
83
t!20
69
74
90
93
89
74
75
83
72
82
83
66
81
72
93
91
86
80
106
70
84
81
110
102
106
75
112
70
87
114
78
102
90
80
103
81
u!07
73
78
96
102
96
79
79
88
78
88
88
70
87
75
97
95
89
83
111
73
87
84
114
106
110
77
116
72
90
117
80
106
93
83
107
84
ullO
71
76
94
99
94
77
77
86
76
86
86
68
85
74
95
93
88
82
109
72
86
83
112
104
108
76
114
71
89
116
79
104
92
82
105
83
u!20
69
74
90
93
89
74
75
83
72
82
83
66
81
72
93
91
86
80
106
70
84
81
110
102
106
75
112
70
87
114
78
102
90
80
103
81
Pop'n.
248,253
86,088
247,052
178,386
207,949
146,568
80,298
35,233
468,973
69,120
192,493
834,054
851,659
281,131
106,107
462,361
113,269
503,341
151,309
250,836
65,797
74,361
182,132
427,239
736,014
729,268
692,134
606,900
818,584
111,183
170,936
101,154
757,027
123,372
215,686
61,236
C-32
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-ll. One-hour county design values (p
State
WI
WI
WI
WI
WI
wv
wv
wv
wv
wv
wv
County
Vernon
Walworth
Washington
Waukesha
Winnebago
Cabell
Greenbrier
Hancock
Kanawha
Ohio
Wood
dvl
85
100
106
109
98
122
99
106
110
107
116
bgl
81
92
95
99
91
98
73
92
91
85
89
ntllO
80
91
94
98
90
97
72
91
90
84
88
pb) and population
t!07
80
90
93
97
89
95
71
90
88
83
87
tno
79
89
91
95
88
93
70
89
86
81
85
t!20
78
87
89
93
86
91
68
87
82
79
83
u!07
80
90
93
97
89
95
71
90
88
83
87
ullO
79
89
91
95
88
93
70
89
86
81
85
u!20
78
87
89
93
86
91
68
87
82
79
83
Pop'n.
51,372
48,741
115,400
863,518
50,871
403,662
118,710
120,284
339,574
264,806
227,813
Table C-12. Eight-hour county design values (ppb) and population
All 8-hour county design values are listed in this table. The design value headings in the
this table are defined as follows:
dv8—measured 8-hour design value 1995-1997.
bg8—projected design value in 2007 after ROTR.
nt810—interpolated design value in 2010 without Tier 2/Sulfur controls.
t807—projected design values in 2007 after ROTR and Tier 2/Sulfur (OMS4).
t810—interpolated design values in 2010 after ROTR and Tier 2/Sulfur.
t820—projected design values in 2020 after ROTR and Tier 2/Sulfur (OMS3).
u807—interpolated design values in 2007 after ROTR and Tier 2/Sulfur.
u810—interpolated design values in 2010 after ROTR and Tier 2/Sulfur.
u820—interpolated design values in 2020 after ROTR and Tier 2/Sulfur.
The "u" values correspond to today's proposal. The "t" values correspond to OMS3 and
OMS4. For discussing the effects of today's proposal on design values, we projected county
design values using the rollback method for three modeling runs: 2007 post ROTR, OMS3, and
OMS4. All other design values have been linearly interpolated based on NOX. See the main text
for further details.
Table C-12. Eight-hour county design values (
State
AL
AL
AL
AL
County
Clay
Elmore
Geneva
Jefferson
dv8
86
77
69
92
bg8
72
65
63
82
nt810
71
64
62
81
)pb) and population
t807
70
63
61
80
t810
68
61
60
78
t820
66
59
58
74
u807
70
63
61
80
u810
68
61
60
78
u820
66
59
58
74
cntypop
13,252
49,210
23,647
651,525
C-33
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-12. Eight-hour county design values (ppb) and population
State
AL
AL
AL
AL
AL
AL
AR
AR
AR
AR
CT
CT
CT
CT
CT
CT
CT
DC
DE
DE
DE
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
FL
GA
GA
County
Lawrence
Madison
Mobile
Montgomery
Shelby
Sumter
Crittenden
Montgomery
Newton
Pulaski
Fairfield
Hartford
Litchfield
Middlesex
New Haven
New London
Tolland
District of Columbia
Kent
New Castle
Sussex
Brevard
Broward
Duval
Escambia
Hillsborough
Lee
Leon
Manatee
Orange
Osceola
Palm Beach
Pasco
Pinellas
Polk
Sarasota
Seminole
St Johns
St Lucie
Volusia
Chatham
DeKalb
dv8
81
84
79
70
95
66
95
67
71
81
104
89
94
98
107
101
91
94
94
99
93
71
69
81
84
82
71
66
75
78
72
67
75
73
74
76
72
72
64
69
72
100
bg8
70
73
75
63
83
60
87
63
66
79
94
78
81
84
93
88
79
80
79
83
81
69
69
79
82
83
70
61
76
78
71
63
73
72
72
76
70
69
62
67
66
86
nt810
69
72
74
62
82
59
86
62
65
78
93
77
80
83
92
87
78
79
78
82
80
68
69
78
81
82
69
60
75
77
70
62
72
71
71
75
69
68
61
66
65
85
t807
68
71
73
61
80
58
85
61
64
77
93
76
79
82
91
86
77
79
77
81
79
68
69
77
81
81
69
59
74
76
69
61
71
70
70
74
67
67
61
64
64
82
t810
67
69
72
59
78
57
83
60
63
75
92
75
77
80
89
84
76
79
75
79
77
66
69
75
79
80
68
57
72
74
67
59
69
69
68
72
65
65
59
62
63
79
t820
65
66
70
57
74
55
81
58
61
72
91
73
75
78
87
82
74
79
72
76
74
64
69
72
77
78
66
53
70
70
64
57
66
67
66
70
62
62
57
59
61
74
u807
68
71
73
61
80
58
85
61
64
77
93
76
79
82
91
86
77
79
77
81
79
68
69
77
81
81
69
59
74
76
69
61
71
70
70
74
67
67
61
64
64
82
u810
67
69
72
59
78
57
83
60
63
75
92
75
77
80
89
84
76
79
75
79
77
66
69
75
79
80
68
57
72
74
67
59
69
69
68
72
65
65
59
62
63
79
u820
65
66
70
57
74
55
81
58
61
72
91
73
75
78
87
82
74
79
72
76
74
64
69
72
77
78
66
53
70
70
64
57
66
67
66
70
62
62
57
59
61
74
cntypop
31,513
238,912
378,643
209,085
99,358
16,174
49,939
7,841
7,666
349,660
827,645
851,783
174,092
143,196
804,219
254,957
128,699
606,900
110,993
441,946
113,229
398,978
1,255,488
672,971
262,798
834,054
335,113
192,493
211,707
677,491
107,728
863,518
281,131
851,659
405,382
277,776
287,529
83,829
150,171
370,712
216,935
545,837
C-34
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-12. Eight-hour county design values (ppb) and population
State
GA
GA
GA
GA
GA
GA
GA
GA
IA
IA
IA
IA
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IL
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
IN
County
Fannin
Fulton
Glynn
Gwinnett
Muscogee
Paulding
Richmond
Rockdale
Linn
Polk
Scott
Van Buren
Adams
Champaign
Cook
Du Page
Effmgham
Jersey
Kane
Lake
Macon
Macoupin
Madison
McHenry
Peoria
Randolph
Rock Island
Sangamon
St Clair
Will
Winnebago
Allen
Clark
Elkhart
Floyd
Hamilton
Hancock
La Porte
Lake
Madison
Marion
Porter
dv8
76
110
75
91
83
89
87
106
62
66
78
68
74
81
91
75
81
84
85
86
83
83
93
85
80
78
70
76
78
79
79
90
92
90
90
97
95
104
95
91
95
96
bg8
66
95
70
78
75
77
77
90
59
62
74
63
67
68
90
70
67
72
75
81
69
69
79
77
71
69
65
65
70
68
71
81
76
79
77
84
83
93
86
79
82
85
nt810
65
94
69
77
74
76
76
89
58
61
73
62
66
67
90
70
66
71
74
81
68
68
78
76
70
68
64
64
69
67
70
80
75
78
77
83
82
92
85
78
81
84
t807
63
91
68
74
71
73
74
86
58
61
73
61
65
67
91
70
66
70
74
81
67
67
76
76
69
67
64
63
69
67
70
79
74
78
77
83
82
92
85
78
80
84
t810
61
89
67
71
69
71
72
83
57
60
72
60
64
66
91
69
65
68
73
80
66
65
74
75
68
66
63
61
68
66
69
78
73
77
76
82
81
91
84
76
79
83
t820
58
84
65
66
64
67
68
78
56
59
71
59
62
64
93
69
63
64
72
79
64
63
71
73
66
64
62
59
66
65
68
76
72
76
76
82
80
89
83
74
78
81
u807
63
91
68
74
71
73
74
86
58
61
73
61
65
67
90
70
66
70
74
81
67
67
76
76
69
67
64
63
69
67
70
79
74
78
77
83
82
92
85
78
80
84
u810
61
89
67
71
69
71
72
83
57
60
72
60
64
66
91
69
65
68
73
80
66
65
74
75
68
66
63
61
68
66
69
78
73
77
76
82
81
91
84
76
79
83
u820
58
84
65
66
64
67
68
78
56
59
71
59
62
64
92
69
63
64
72
79
64
63
71
73
66
64
62
59
66
65
68
76
72
76
76
82
80
89
83
74
78
81
cntypop
15,992
648,951
62,496
352,910
179,278
41,611
189,719
54,091
168,767
327,140
150,979
7,676
66,090
173,025
5,105,067
781,666
31,704
20,539
317,471
516,418
117,206
47,679
249,238
183,241
182,827
34,583
148,723
178,386
262,852
357,313
252,913
300,836
87,777
156,198
64,404
108,936
45,527
107,066
475,594
130,669
797,159
128,932
C-35
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-12. Eight-hour county design values (ppb) and population
State
IN
IN
IN
IN
KS
KS
KS
KS
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
KY
LA
LA
LA
LA
LA
LA
LA
County
St. Joseph
Vanderburgh
Vigo
Warrick
Miami
Pawnee
Sedgwick
Wyandotte
Bell
Boone
Boyd
Bullitt
Campbell
Christian
Daviess
Edmonson
Fayette
Graves
Greenup
Hancock
Hardin
Henderson
Jefferson
Jessamine
Kenton
Lawrence
Livingston
McCracken
McLean
Oldham
Perry
Pike
Pulaski
Scott
Simpson
Ascension
Beauregard
Bossier
Caddo
Calcasieu
East Baton Rouge
Grant
dv8
91
93
88
93
74
71
74
85
76
82
85
85
90
80
87
82
85
74
80
89
81
86
91
80
90
70
86
80
83
87
69
74
76
78
83
87
75
79
82
85
95
75
bg8
80
80
72
77
72
71
72
83
62
68
69
70
79
66
73
68
75
60
64
67
62
69
75
66
78
53
72
65
69
74
58
62
61
68
72
81
69
74
77
77
88
69
nt810
79
79
71
76
71
71
71
82
61
67
68
69
78
65
72
67
74
59
63
66
61
68
75
65
77
52
71
64
68
73
57
61
60
67
71
80
68
73
76
76
87
68
t807
79
78
70
75
71
71
71
82
60
67
67
68
77
65
72
67
74
59
63
66
61
68
75
64
77
52
71
64
67
73
57
60
60
66
70
80
68
73
75
76
86
68
t810
78
77
69
74
70
71
70
81
59
66
66
67
76
64
71
66
73
58
62
65
60
67
74
63
76
51
70
63
66
72
56
58
59
65
69
79
67
72
74
75
84
67
t820
77
76
68
72
69
71
69
81
57
65
65
66
74
62
70
64
73
56
60
63
58
65
74
61
74
50
69
61
65
70
54
56
58
63
67
77
66
70
72
75
81
66
u807
79
78
70
75
71
71
71
82
60
67
67
68
77
65
72
67
74
59
63
66
61
68
75
64
77
52
71
64
67
73
57
60
60
66
70
80
68
73
75
76
86
68
u810
78
77
69
74
70
71
70
81
59
66
66
67
76
64
71
66
73
58
62
65
60
67
74
63
76
51
70
63
66
72
56
58
59
65
69
79
67
72
74
75
84
67
u820
77
76
68
72
69
71
69
81
57
65
65
66
74
62
70
64
73
56
60
63
58
65
74
61
74
50
69
61
65
70
54
56
58
63
67
77
66
70
72
75
81
66
cntypop
247,052
165,058
106,107
44,920
23,466
7,555
403,662
161,993
31,506
57,589
51,150
47,567
83,866
68,941
87,189
10,357
225,366
33,550
36,742
7,864
89,240
43,044
664,937
30,508
142,031
13,998
9,062
62,879
9,628
33,263
30,283
72,583
49,489
23,867
15,145
58,214
30,083
86,088
248,253
168,134
380,105
17,526
C-36
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-12. Eight-hour county design values (ppb) and population
State
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
LA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MA
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
MD
ME
ME
ME
ME
ME
ME
ME
County
Iberville
Jefferson
Lafayette
Lafourche
Livingston
Orleans
Ouachita
Pointe Coupee
St Bernard
St Charles
St James
St John The Baptist
StMary
West Baton Rouge
Barnstable
Berkshire
Bristol
Essex
Hampden
Hampshire
Middlesex
Plymouth
Suffolk
Worcester
Anne Arundel
Baltimore
Baltimore City
Calvert
Carroll
Cecil
Charles
Harford
Kent
Montgomery
Prince George's
Cumberland
Hancock
Kennebec
Knox
Oxford
Penobscot
Piscataquis
dv8
96
83
84
85
88
71
77
83
80
81
84
82
81
86
100
77
97
89
89
97
87
71
76
87
107
96
103
85
93
110
93
107
96
93
100
94
83
75
87
59
73
64
bg8
91
80
79
81
82
66
73
77
75
77
79
77
76
80
89
67
84
79
81
87
76
62
70
75
93
82
93
69
76
92
77
93
82
78
86
83
72
67
76
53
65
58
nt810
90
79
78
81
81
65
72
76
74
76
78
76
75
79
88
66
83
78
80
86
75
61
70
74
92
81
92
68
75
91
76
92
81
77
85
82
71
66
75
52
64
57
t807
90
79
77
81
80
65
72
76
74
76
77
76
75
78
87
65
82
77
80
85
75
61
70
73
90
79
91
67
74
90
75
91
80
76
84
80
70
66
74
52
64
56
t810
89
78
76
80
78
64
71
74
73
75
76
75
74
76
85
64
80
75
79
84
73
60
70
71
88
77
90
65
73
88
73
89
78
74
82
78
69
64
72
50
62
55
t820
88
78
74
79
76
63
70
72
72
75
75
74
73
73
83
62
78
73
78
82
71
58
70
69
86
75
89
63
71
85
71
86
76
72
80
76
67
62
70
48
60
53
u807
90
79
77
81
80
65
72
76
74
76
77
76
75
78
87
65
82
77
80
85
75
61
70
73
90
79
91
67
74
90
75
91
80
76
84
80
70
66
74
52
64
56
u810
89
78
76
80
78
64
71
74
73
75
76
75
74
76
85
64
80
75
79
84
73
60
70
71
88
77
90
65
73
88
73
89
78
74
82
78
69
64
72
50
62
55
u820
88
78
74
79
76
63
70
72
72
75
75
74
73
73
83
62
78
73
78
82
71
58
70
69
86
75
89
63
71
85
71
86
76
72
80
76
67
62
70
48
60
53
cntypop
31,049
448,306
164,762
85,860
70,526
496,938
142,191
22,540
66,631
42,437
20,879
39,996
58,086
19,419
186,605
139,352
506,325
670,080
456,310
146,568
1,398,468
435,276
663,906
709,705
427,239
692,134
736,014
51,372
123,372
71,347
101,154
182,132
17,842
757,027
729,268
243,135
46,948
115,904
36,310
52,602
146,601
18,653
C-37
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-12. Eight-hour county design values (ppb) and population
State
ME
ME
ME
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MI
MN
MN
MN
MN
MN
MO
MO
MO
MO
MO
MO
MO
MO
MO
MO
MS
MS
MS
MS
MS
County
Sagadahoc
Somerset
York
Allegan
Benzie
Berrien
Cass
Clinton
Genesee
Huron
Ingham
Kalamazoo
Kent
Lenawee
Macomb
Mason
Muskegon
Oakland
Ottawa
St. Clair
Washtenaw
Wayne
Anoka
Dakota
Lake
St Louis
Washington
Clay
Greene
Jackson
Jefferson
Monroe
Platte
St Louis City
St. Charles
St. Louis
Ste. Genevieve
Adams
Choctaw
DeSoto
Hancock
Hinds
dv8
95
69
96
98
88
98
94
74
84
82
83
87
88
83
91
98
99
78
84
92
83
88
81
68
63
66
74
94
78
73
90
84
86
83
100
88
87
77
55
88
80
76
bg8
82
62
85
86
79
86
83
64
75
73
74
75
79
75
82
88
89
71
75
84
74
80
79
66
63
66
72
90
70
70
83
78
81
76
90
78
77
74
49
84
76
70
nt810
81
61
84
85
78
85
82
63
75
72
73
74
78
74
82
87
88
70
74
83
73
79
79
66
63
66
71
89
69
69
82
77
81
75
89
77
76
73
48
83
75
69
t807
80
61
83
85
78
85
81
63
75
72
73
74
77
74
82
87
87
70
73
82
72
79
79
66
63
66
71
88
68
69
79
76
82
75
87
75
75
73
48
83
74
68
t810
78
59
81
84
77
84
80
62
74
71
72
73
76
73
81
86
86
69
72
81
71
79
78
65
63
66
70
86
66
68
78
74
81
74
85
74
73
72
47
82
73
67
t820
76
57
79
83
75
82
78
61
73
69
71
71
75
71
81
84
84
68
71
80
70
79
78
65
63
66
69
84
63
66
76
72
79
73
82
72
70
70
46
81
72
65
u807
80
61
83
85
78
85
81
63
75
72
73
74
77
74
82
87
87
70
73
82
72
79
79
66
63
66
71
88
68
69
79
76
82
75
87
75
75
73
48
83
74
68
u810
78
59
81
84
77
84
80
62
74
71
72
73
76
73
81
86
86
69
72
81
71
79
78
65
63
66
70
86
66
68
78
74
81
74
85
74
73
72
47
82
73
67
u820
76
57
79
83
75
82
78
61
73
69
71
71
75
71
81
84
84
68
71
80
70
79
78
65
63
66
69
84
63
66
76
72
79
73
82
72
70
70
46
81
72
65
cntypop
33,535
49,767
164,587
90,509
12,200
161,378
49,477
57,883
430,459
34,951
281,912
223,411
500,631
91,476
717,400
25,537
158,983
1,083,592
187,768
145,607
282,937
2,111,687
243,641
275,227
10,415
198,213
145,896
153,411
207,949
633,232
171,380
9,104
57,867
396,685
212,907
993,529
16,037
35,356
9,071
67,910
31,760
254,441
C-38
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-12. Eight-hour county design values (ppb) and population
State
MS
MS
MS
MS
MS
MS
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
NC
ND
NE
NE
NH
NH
NH
NH
County
Jackson
Lauderdale
Lee
Madison
Sharkey
Warren
a Alexander
Buncombe
Caldwell
Camden
Caswell
Chatham
Cumberland
Davie
Duplin
Durham
Edgecombe
Forsyth
Franklin
Granville
Guilford
Haywood
Johnston
Lincoln
Martin
Mecklenburg
New Hanover
Northampton
Person
Pitt
Rockingham
Rowan
Swain
Wake
Yancey
Cass
Douglas
Lancaster
Belknap
Carroll
Cheshire
Coos
dv8
86
72
75
77
80
76
79
75
79
83
89
85
87
88
72
83
84
89
86
94
85
85
87
86
75
97
79
86
84
88
84
93
66
89
84
67
67
56
64
62
74
77
bg8
88
65
68
71
74
71
71
67
71
74
74
71
79
78
65
74
76
80
73
83
76
75
76
77
65
94
72
71
68
77
73
82
57
82
71
67
65
54
57
55
66
70
nt810
87
64
67
70
73
70
70
66
70
73
73
70
78
77
64
73
75
79
72
82
75
74
75
76
64
93
71
70
67
76
72
81
56
81
70
67
64
54
56
54
65
69
t807
87
63
66
69
73
70
68
65
68
72
71
68
76
75
62
71
72
77
69
79
73
72
73
73
63
90
70
69
65
74
70
79
55
78
69
67
64
54
56
53
64
67
t810
86
62
65
68
72
69
66
63
66
70
69
66
74
73
60
69
70
75
67
76
71
70
71
71
61
87
68
67
63
72
68
77
53
75
67
67
63
53
55
52
63
65
t820
85
60
63
67
71
68
62
59
63
68
66
63
69
69
58
64
66
71
63
71
66
67
67
67
58
82
66
64
60
68
65
73
51
70
65
67
63
53
53
51
61
63
u807
87
63
66
69
73
70
68
65
68
72
71
68
76
75
62
71
72
77
69
79
73
72
73
73
63
90
70
69
65
74
70
79
55
78
69
67
64
54
56
53
64
67
u810
86
62
65
68
72
69
66
63
66
70
69
66
74
73
60
69
70
75
67
76
71
70
71
71
61
87
68
67
63
72
68
77
53
75
67
67
63
53
55
52
63
65
u820
85
60
63
67
71
68
62
59
63
68
66
63
69
69
58
64
66
71
63
71
66
67
67
67
58
82
66
64
60
68
65
73
51
70
65
67
63
53
53
51
61
63
cntypop
115,243
75,555
65,581
53,794
7,066
47,880
27,544
174,821
70,709
5,904
20,693
38,759
274,566
27,859
39,995
181,835
56,558
265,878
36,414
38,345
347,420
46,942
81,306
50,319
25,078
511,433
120,284
20,798
30,180
107,924
86,064
110,605
11,268
423,380
15,419
102,874
416,444
213,641
49,216
35,410
70,121
34,828
C-39
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-12. Eight-hour county design values (ppb) and population
State
NH
NH
NH
NH
NH
NH
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
County
Grafton
Hillsborough
Merrimack
Rockingham
Strafford
Sullivan
Atlantic
Bergen
Camden
Cumberland
Essex
Gloucester
Hudson
Hunterdon
Mercer
Middlesex
Monmouth
Morris
Ocean
Union
Albany
Bronx
Chautauqua
Chemung
Dutchess
Erie
Essex
Hamilton
Herkimer
Jefferson
Kings
Madison
Monroe
Niagara
Oneida
Onondaga
Orange
Putnam
Queens
JAichmond
Saratoga
Schenectady
dv8
57
87
71
95
78
71
100
94
106
93
92
105
98
98
101
103
100
100
108
83
81
95
85
73
90
76
86
75
70
88
88
77
81
85
72
79
91
90
91
103
76
74
bg8
51
77
63
86
71
64
87
84
92
78
79
92
94
83
88
89
87
84
92
71
72
97
74
65
78
66
78
67
62
79
79
69
74
77
64
70
79
78
90
92
68
66
nt810
50
76
62
85
70
63
86
83
91
77
78
91
94
82
87
88
86
83
91
70
71
97
73
64
77
65
77
66
61
78
78
68
73
76
63
69
78
77
90
91
67
65
t807
50
75
62
84
69
62
85
82
90
76
78
90
95
81
86
87
85
82
89
69
71
98
72
63
76
65
76
65
61
77
77
67
73
76
63
69
77
76
90
90
67
65
t810
49
74
61
82
68
61
83
81
88
74
77
88
94
79
84
86
83
80
87
68
70
98
71
62
74
64
74
64
60
75
76
66
72
75
62
68
76
74
90
88
66
64
t820
47
72
60
80
66
59
81
80
85
71
76
86
94
76
82
84
81
78
84
66
68
99
69
60
72
63
72
62
58
73
74
64
70
74
60
66
74
72
92
86
64
62
u807
50
75
62
84
69
62
85
82
90
76
78
90
95
81
86
87
85
82
89
69
71
97
72
63
76
65
76
65
61
77
77
67
73
76
63
69
77
76
89
90
67
65
u810
49
74
61
82
68
61
83
81
88
74
77
88
94
79
84
86
83
80
87
68
70
98
71
62
74
64
74
64
60
75
76
66
72
75
62
68
76
74
90
88
66
64
u820
47
72
60
80
66
59
81
80
85
71
76
86
94
76
82
84
81
78
84
66
68
98
69
60
72
63
72
62
58
73
74
64
70
74
60
66
74
72
91
86
64
62
cntypop
74,929
336,073
120,005
245,845
104,233
38,592
224,327
825,380
502,824
138,053
778,206
230,082
553,099
107,776
325,824
671,780
553,124
421,353
433,203
493,819
292,594
1,203,789
141,895
95,195
259,462
968,532
37,152
5,279
65,797
110,943
2,300,664
69,120
713,968
220,756
250,836
468,973
307,647
83,941
1,951,598
378,977
181,276
149,285
C-40
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-12. Eight-hour county design values (ppb) and population
State
NY
NY
NY
NY
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OK
OK
OK
OK
OK
PA
PA
PA
PA
County
Suffolk
Ulster
Wayne
Westchester
Allen
Ashtabula
Butler
Clark
Clermont
Clinton
Cuyahoga
Franklin
Hamilton
Jefferson
Knox
Lake
Lawrence
Licking
Logan
Lorain
Lucas
Madison
Mahoning
Medina
Miami
Montgomery
Portage
Preble
Stark
Summit
Trumbull
Warren
Washington
Cleveland
Comanche
Me Clain
Oklahoma
Tulsa
Allegheny
Beaver
Berks
Blair
dv8
102
82
83
94
89
85
93
93
89
97
88
87
95
84
91
99
82
92
86
87
89
91
90
89
88
91
88
86
88
91
93
99
89
79
60
77
83
88
105
87
92
90
bg8
89
72
75
82
75
74
77
79
76
80
77
73
87
71
77
86
65
78
73
78
80
76
78
78
75
76
77
70
76
80
81
83
69
78
57
74
83
87
95
76
78
75
nt810
88
71
74
81
74
73
76
78
75
79
76
72
87
70
76
85
64
77
72
77
79
75
77
77
74
75
76
69
75
79
80
82
68
77
56
73
82
86
94
75
77
74
t807
86
70
73
80
74
73
76
77
75
78
75
71
88
69
76
85
64
77
71
76
79
74
77
77
73
75
76
68
75
79
79
81
67
76
56
72
81
86
94
75
76
73
t810
84
68
72
79
73
72
75
76
74
77
74
70
88
68
75
84
63
76
70
75
78
73
75
76
72
74
74
67
74
78
77
79
66
74
55
71
80
85
93
74
74
72
t820
82
66
71
77
72
70
74
74
72
75
73
69
88
66
73
82
61
74
68
74
76
71
73
74
71
73
72
65
73
76
75
77
64
72
54
69
78
84
91
72
71
70
u807
86
70
73
80
74
73
76
77
75
78
75
71
88
69
76
85
64
77
71
76
79
74
77
77
73
75
76
68
75
79
79
81
67
76
56
72
81
86
94
75
76
73
u810
84
68
72
79
73
72
75
76
74
77
74
70
88
68
75
84
63
76
70
75
78
73
75
76
72
74
74
67
74
78
77
79
66
74
55
71
80
85
93
74
74
72
u820
82
66
71
77
72
70
74
74
72
75
73
69
88
66
73
82
61
74
68
74
76
71
73
74
71
73
72
65
73
76
75
77
64
72
54
69
78
84
91
72
71
70
cntypop
1,321,864
165,304
89,123
874,866
109,755
99,821
291,479
147,548
150,187
35,415
1,412,140
961,437
866,228
80,298
47,473
215,499
61,834
128,300
42,310
271,126
462,361
37,068
264,806
122,354
93,182
573,809
142,585
40,113
367,585
514,990
227,813
113,909
62,254
174,253
111,486
22,795
599,611
503,341
1,336,449
186,093
336,523
130,542
C-41
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-12. Eight-hour county design values (ppb) and population
State
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
PA
RI
RI
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
SC
TN
TN
TN
County
Bucks
Cambria
Dauphin
Delaware
Erie
Franklin
Lackawanna
Lancaster
Lawrence
Lehigh
Luzerne
Lycoming
Mercer
Montgomery
Northampton
Perry
Philadelphia
Washington
Westmoreland
York
Kent
Providence
a Abbeville
Aiken
Anderson
Barnwell
Berkeley
Charleston
Cherokee
Chester
Darlington
Edgefield
Oconee
Pickens
Richland
Spartanburg
Union
Williamsburg
York
Anderson
Blount
Davidson
dv8
102
88
88
100
87
86
86
96
84
95
90
73
92
97
90
84
102
95
88
87
96
87
77
81
88
81
71
76
84
87
78
76
76
83
83
86
81
70
82
87
95
88
bg8
86
74
75
87
75
71
75
83
74
80
74
61
81
83
76
70
94
74
76
74
82
74
70
73
79
71
63
70
74
78
68
66
68
76
74
79
73
61
75
75
83
82
nt810
85
73
74
86
74
70
74
82
73
79
73
60
80
82
75
69
94
73
75
73
81
73
69
72
78
70
62
69
73
77
67
65
67
75
73
78
72
60
74
74
82
81
t807
83
72
72
85
73
69
73
81
72
78
72
60
78
81
74
68
94
73
74
72
79
72
67
70
76
68
61
68
71
75
66
63
65
72
71
75
70
59
72
72
81
81
t810
81
71
70
83
72
68
72
79
70
76
71
59
76
79
72
66
93
71
73
70
77
71
65
68
74
66
60
67
69
73
64
61
63
70
69
73
68
58
70
70
79
80
t820
79
69
67
81
70
66
70
76
68
73
69
57
74
76
70
64
92
69
71
68
75
69
62
64
70
64
58
65
65
70
61
58
59
66
64
68
65
56
66
68
76
79
u807
83
72
72
85
73
69
73
81
72
78
72
60
78
81
74
68
94
73
74
72
79
72
67
70
76
68
61
68
71
75
66
63
65
72
71
75
70
59
72
72
81
81
u810
81
71
70
83
72
68
72
79
70
76
71
59
76
79
72
66
93
71
73
70
77
71
65
68
74
66
60
67
69
73
64
61
63
70
69
73
68
58
70
70
79
80
u820
79
69
67
81
70
66
70
76
68
73
69
57
74
76
70
64
92
69
71
68
75
69
62
64
70
64
58
65
65
70
61
58
59
66
64
68
65
56
66
68
76
79
cntypop
541,174
163,029
237,813
547,651
275,572
121,082
219,039
422,822
96,246
291,130
328,149
118,710
121,003
678,111
247,105
41,172
1,585,577
204,584
370,321
339,574
161,135
596,270
23,862
120,940
145,196
20,293
128,776
295,039
44,506
32,170
61,851
18,375
57,494
93,894
285,720
226,800
30,337
36,815
131,497
68,250
85,969
510,784
C-42
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-12. Eight-hour county design values (ppb) and population
State
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TN
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
TX
VA
VA
VA
VA
VA
VA
VA
VA
VA
VA
County
Hamilton
Haywood
Humphreys
Jefferson
Knox
Putnam
Rutherford
Sevier
Shelby
Sullivan
Sumner
Williamson
Wilson
Bexar
Brazoria
Cameron
Collin
Dallas
Denton
Ellis
Galveston
Gregg
Harris
Hidalgo
Jefferson
Nueces
Orange
Smith
Tarrant
Travis
Victoria
Webb
Alexandria City
Arlington
Caroline
Charles City
Chesterfield
Fairfax
Fauquier
Frederick
Hampton City
Hanover
dv8
90
82
77
96
95
80
79
93
95
88
99
88
90
87
92
66
101
95
104
82
105
91
117
56
93
83
86
89
97
81
78
60
87
91
84
90
87
92
81
84
87
90
bg8
82
71
65
80
85
67
67
81
90
77
88
80
80
85
87
66
93
89
94
78
99
84
109
55
86
81
80
82
89
77
75
60
76
79
70
77
74
79
64
67
78
77
nt810
81
70
64
79
84
66
66
80
89
76
87
79
79
84
86
66
92
88
93
77
98
83
109
55
85
81
79
81
88
76
74
60
75
79
69
76
73
78
63
66
77
76
t807
79
69
64
78
83
65
65
78
88
76
86
78
78
84
86
66
91
88
92
76
98
83
110
55
85
81
79
80
87
75
74
60
75
79
67
75
71
78
62
65
76
74
t810
77
68
63
76
81
64
63
76
86
75
84
76
76
83
85
66
89
87
90
75
97
81
110
55
84
81
78
79
86
74
73
60
75
78
65
73
69
76
61
63
74
72
t820
73
66
61
73
78
62
61
73
84
73
81
74
73
82
84
66
87
85
87
73
97
79
111
55
82
81
77
77
84
72
73
60
75
78
62
70
67
74
59
61
72
68
u807
79
69
64
78
83
65
65
78
88
76
86
78
78
84
86
66
91
88
92
76
98
83
109
55
85
81
79
80
87
75
74
60
75
79
67
75
71
78
62
65
76
74
u810
77
68
63
76
81
64
63
76
86
75
84
76
76
83
85
66
89
87
90
75
97
81
110
55
84
81
78
79
86
74
73
60
75
78
65
73
69
76
61
63
74
72
u820
73
66
61
73
78
62
61
73
84
73
81
74
73
82
84
66
87
85
87
73
97
79
110
55
82
81
77
77
84
72
73
60
75
78
62
70
67
74
59
61
72
68
cntypop
285,536
19,437
15,795
33,016
335,749
51,373
118,570
51,043
826,330
143,596
103,281
81,021
67,675
1,185,394
191,707
260,120
264,036
1,852,810
273,525
85,167
217,399
104,948
2,818,199
383,545
239,397
291,145
80,509
151,309
1,170,103
576,407
74,361
133,239
111,183
170,936
19,217
6,282
209,274
818,584
48,741
45,723
133,793
63,306
C-43
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Table C-12. Eight-hour county design values (ppb) and population
State
VA
VA
VA
VA
VA
VA
VA
VA
VT
VT
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
WI
wv
wv
wv
wv
wv
County
Henri co
Henry
Madison
Prince William
Roanoke
Stafford
Suffolk City
Wythe
Bennington
Chittenden
Brown
Columbia
Dane
Dodge
Door
Florence
Fond Du Lac
Jefferson
Kenosha
Kewaunee
Manitowoc
Marathon
Milwaukee
Oneida
Outagamie
Ozaukee
Polk
Racine
Rock
Sauk
Sheboygan
St Croix
Vernon
Walworth
Washington
Waukesha
Winnebago
Cabell
Greenbrier
Hancock
Kanawha
Ohio
dv8
90
82
86
88
78
86
82
78
79
69
82
81
79
78
92
64
79
80
96
93
95
71
94
64
80
98
69
92
85
74
92
67
69
83
80
81
81
88
83
85
81
86
bg8
77
67
71
73
66
69
70
61
71
63
75
76
74
72
83
60
73
73
88
84
85
67
89
61
75
94
66
85
78
70
83
65
66
77
74
75
76
71
61
73
68
68
nt810
76
66
70
72
65
68
69
60
70
62
74
75
73
71
82
59
72
72
87
83
84
66
88
60
74
93
65
84
77
69
82
64
65
76
73
74
75
70
60
72
67
67
t807
74
64
69
71
64
67
68
59
69
61
74
74
73
71
81
59
72
72
86
82
83
66
87
59
74
93
65
84
77
68
81
64
65
76
73
73
74
69
59
72
65
67
t810
72
62
67
69
62
65
66
58
67
60
73
73
72
70
80
58
71
71
85
81
82
65
86
58
73
92
64
82
76
67
80
63
64
75
72
72
73
68
58
71
63
66
t820
69
60
65
67
59
63
63
56
65
59
71
71
70
68
78
57
70
69
83
79
80
63
84
57
71
90
63
80
75
66
79
62
63
73
70
70
71
66
57
69
61
64
u807
74
64
69
71
64
67
68
59
69
61
74
74
73
71
81
59
72
72
86
82
83
66
87
59
74
93
65
84
77
68
81
64
65
76
73
73
74
69
59
72
65
67
u810
72
62
67
69
62
65
66
58
67
60
73
73
72
70
80
58
71
71
85
81
82
65
86
58
73
92
64
82
76
67
80
63
64
75
72
72
73
68
58
71
63
66
u820
69
60
65
67
59
63
63
56
65
59
71
71
70
68
78
57
70
69
83
79
80
63
84
57
71
90
63
80
75
66
79
62
63
73
70
70
71
66
57
69
61
64
cntypop
217,881
56,942
11,949
215,686
79,332
61,236
52,141
25,466
35,845
131,761
194,594
45,088
367,085
76,559
25,690
4,590
90,083
67,783
128,181
18,878
80,421
115,400
959,275
31,679
140,510
72,831
34,773
175,034
139,510
46,975
103,877
50,251
25,617
75,000
95,328
304,715
140,320
96,827
34,693
35,233
207,619
50,871
C-44
-------
Appendix C: One-hour and Eight-hour County Design Values
Table C-12. Eight-hour county design values (ppb) and population
State
WV
County
Wood
dv8
89
bg8
70
nt810
69
t807
68
t810
67
t820
65
u807
68
u810
67
u820
65
cntypop
86,915
Table C-13. Metropolitan areas and rural counties with
design values that exceeded
standard prior to ROTR controls, and were projected to meet but remain within 15%
hour standard after ROTR controls.
Name
Manitowoc WI
Door WI
Mason MI
KentMD
Sagadahoc ME
Atlanta, GA MSA
Barnstable- Yarmouth, MA MSA
Baton Rouge, LA MSA
Birmingham, AL MSA
Boston-Worcester-Lawrence, MA-NH-ME-CT CMSA
Chicago-Gary-Kenosha, IL-IN-WI CMSA
Dallas-Fort Worth, TX CMSA
Grand Rapids-Muskegon-Holland, MI MSA
Houma,LAMSA
Kansas City, MO-KS MSA
Lancaster, PA MSA
Louisville, KY-IN MSA
Milwaukee-Racine, WI CMSA
Pittsburgh, PA MSA
Providence-Fall River-Warwick, RI-MA MSA
St. Louis, MO-IL MSA
Springfield, MA MSA
Design Value
(ppb)
113
113
113
110
109
124
118
122
117
119
124
124
121
122
122
109
111
118
121
114
117
118
sum
MA count
MA sum
Cnty count
Cnty sum
the 1-hour
of the 1-
Pop'n.
r
80,421
25,690
25,537
17,842
33,535
2,959,500
134,954
528,261
839,942
5,455,403
8,239,820
4,037,282
937,891
182,842
1,582,874
422,822
949,012
1,607,183
2,394,811
1,134,350
2,492,348
587,884
34,670,204
17
34,487,179
5
183,025
C-45
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Appendix C. References
1. U.S. EPA. Green Book Web Site: http://www.epa.gov/oar/oaqps/greenbook/oytc.html.
C-46
-------
Appendix D: Legal Authority for Gasoline Sulfur Control
Appendix D: EPA's Legal Authority for Proposing
Gasoline Sulfur Controls
We are proposing gasoline sulfur controls pursuant to our authority under Section
21 l(c)(l) of the Clean Air Act. This section gives us the authority to "control or prohibit the
manufacture, introduction into commerce, offering for sale, or sale" of any fuel or fuel additive
(A) whose emission products, in the judgment of the Administrator, cause or contribute to air
pollution "which may be reasonably anticipated to endanger the public health or welfare" or (B)
whose emission products "will impair to a significant degree the performance of any emission
control device or system which is in general use, or which the Administrator finds has been
developed to a point where in a reasonable time it would be in general use" were the fuel control
or prohibition adopted. The following sections describe current our regulatory requirements that
affect gasoline sulfur content, and explain our bases for proposing to control gasoline sulfur
under Section 21 l(c)(l).
A. EPA's Current Regulatory Requirements for Gasoline
We currently have regulatory requirements for conventional and reformulated gasoline
(RFG), adopted under Sections 21 l(c) and 21 l(k) of the Act. RFG is required to be sold in
certain ozone nonattainment areas. Gasoline sold in the rest of the country is subject to the
conventional gasoline requirements. See 40 CFR part 80.
Both the RFG and conventional gasoline (CG) requirements include a NOx performance
standard that requires refiners to achieve a certain level of NOx control compared to 1990
baseline levels. As a practical matter, compliance with this performance standard results in
limiting sulfur levels in RFG. The NOx reductions required by the Phase 2 RFG requirements,
effective on January 1, 2000, are expected to result in RFG sulfur levels of about 150 ppm. In
addition, EPA's regulations require compliance with the RFG and CG standards (including the
NOx performance standard) to be calculated using the Complex Model beginning in 1998. This
model contains range limits for RFG for a number of fuel parameters that affect NOx
performance, including a range of zero to 500 ppm for sulfur. Therefore, the requirement to use
the Complex Model effectively limits sulfur levels in RFG to no more than 500 ppm. The sulfur
Complex Model range limit for RFG is the only direct regulation of sulfur content under Section
21 l(c)(l). However, the NOx performance standards for RFG and CG have an indirect effect on
sulfur content."™"1
All gasoline is subject to Section 21 l(f) of the Act, which prohibits fuel or fuel additive
"""""Because sulfur is directly or indirectly controlled by EPA requirements, and will be controlled directly
under today's proposal, states are preempted from initiating sulfur control programs unless they are identical to the
federal requirements. See the discussion in Section V.B on this subject.
D-l
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
manufactures from introducing into commerce, or increasing the concentration in use of, any fuel
or fuel additive for general use in light duty motor vehicles which is not "substantially similar" to
the fuel used in the certification of model year 1975 or later vehicles or engines. We have
interpreted "substantially similar" for unleaded gasoline to include any gasoline meeting the 1988
ASTM specifications for unleaded gasoline (ASTM D 4814-88"™), which limits the sulfur
content of unleaded gasoline to 0.1 weight percent (1000 ppm) sulfur.
B. How the Proposed Gasoline Sulfur Control Program Meets the CAA Section
211(c) Criteria
Under Section 21 l(c)(l), EPA may adopt a fuel control if at least one of the following
two criteria is met: 1) the emission products of the fuel cause or contribute to air pollution which
may reasonably be anticipated to endanger public health or welfare, or 2) the emission products
of the fuel will significantly impair emissions control systems in general use or which would be
in general use were the fuel control to be adopted. We are proposing to control sulfur levels in
gasoline based on both of these criteria. Under the first criterion, we believe that emissions
products of sulfur in gasoline used in Tier 1 and LEV technology vehicles contribute to ozone
pollution, air toxics, and PM. Under the second criterion, we believe that gasoline sulfur in fuel
that will be used in Tier 2 technology vehicles will significantly impair the emissions control
systems expected to be used in such vehicles. The following sections summarize our analysis of
each criterion.
1. Health and Welfare Concerns of Air Pollution Caused by Sulfur in Gasoline
We believe that the emission products of gasoline sulfur contribute to air pollution that
can reasonably be anticipated to endanger public health and welfare. The combustion products
of the sulfur-containing compounds in gasoline (SO2 and other sulfur oxides) contribute to air
pollution that has adverse impacts on public health and welfare. The greatest impact of gasoline
sulfur on pollution is the increase in emissions of hydrocarbons (including hazardous air
pollutants such as benzene and 1,3-butadiene), NOx, particulate matter, and compounds such as
nitrates and sulfates that become particulates in the atmosphere. As explained below and in the
RIA, these increased emissions result primarily from the adverse impact of high sulfur levels on
the automotive catalysts used in the vehicles which have recently entered the fleet or will be used
to comply with the proposed Tier 2 standards. The health and welfare implications of the
emissions of these compounds are discussed in greater detail in Section III of the Preamble.
Section 21 l(c)(2)(A) requires that, prior to adopting a fuel control based on a finding that
the fuel's emission products contribute to air pollution that can reasonably be anticipated to
endanger public health or welfare, EPA consider "all relevant medical and scientific evidence
"""Standard Specification for Automotive Spark-Ignition Engine Fuel
D-2
-------
Appendix D: Legal Authority for Gasoline Sulfur Control
available, including consideration of other technologically or economically feasible means of
achieving emission standards under [section 202 of the Act]." EPA's analysis of the medical and
scientific evidence relating to the emissions impact of sulfur in gasoline is described in more
detail in the RIA.
2. Impact of Gasoline Sulfur Emission Products on Emission Control Systems
EPA believes that sulfur in gasoline can significantly impair the emissions control
technology of vehicles designed to meet the proposed Tier 2 emissions standards. We know that
gasoline sulfur has a negative impact on vehicle emission controls. This is not a new
development. Vehicles depend on the catalytic converter to oxidize or reduce emissions of HC,
CO, and NOx. Sulfur and sulfur compounds attach or "adsorb" to the precious metals which are
required to convert these emissions. Sulfur also blocks sites on the catalyst designed to store
oxygen which are necessary to optimize NOx emissions conversions. While the amount of sulfur
contamination can vary depending on the metals used in the catalyst and other aspects of the
design and operation of the vehicle, some level of sulfur contamination will occur in any catalyst.
For older vehicles designed to meet Tier 0 and Tier 1 emission standards, this sulfur
contamination increases emissions of NMHC and NOx by almost 17 percent when one of these
vehicles is operated on gasoline containing 330 ppm sulfur (approximately the current national
average sulfur level) compared to operation on gasoline with 30 ppm sulfur (which is close to
California's current average sulfur level, and is the average sulfur level proposed in this notice).
Thus, Tier 0 and Tier 1 vehicles have higher emissions when they are exposed to sulfur levels
substantially higher than the proposed sulfur standard. This increase is generally not enough to
cause a vehicle to exceed the full useful life emission standards in practice, because car
manufacturers design the vehicles with a margin of safety to compensate for deterioration in
emissions performance over the life of the vehicle. However, it does lead to greater in-use
emissions than achieved with the proposed control on gasoline sulfur content.
The sulfur impact on the catalysts used in later model vehicles is clearly significant. High
sulfur levels have been shown to significantly impair the emissions control systems of cleaner,
later technology vehicles. The California LEV standards and Federal NLEV standards, as well as
California's new LEV-II standards and our proposed Tier 2 standards, require catalysts to be
extremely efficient to adequately reduce emissions over the full useful life of the vehicle. Recent
test programs conducted by the automotive and oil industries show that LEV and ULEV vehicles
can experience, on average, a 40 percent increase in NMHC and 134 percent increase in NOx
emissions when operated on 330 ppm sulfur fuel compared to 30 ppm sulfur fuel. This level of
emissions increase is significant enough that it would undermine the technical and economic
feasibility of the Tier 2 standards proposed today.
This level of impact on emission control system efficiency would mean actual in-use
emissions reductions from the proposed tier 2 standards would be undercut to such a degree that
the resulting limited in-use emissions reductions would not warrant the expense imposed by the
D-3
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Tier 2 standards, and would not achieve the in-use emissions reductions from these motor
vehicles needed to address the air quality problems described elsewhere in this notice. In
addition, the concerns about irreversibility of the damage to the catalyst mean it would not be
feasible to design an emission control system that would offset this level of impact on the
efficiency of the control system in order to comply over the useful life of the vehicles. Average
sulfur levels in the U.S. are currently high enough to significantly impair the emissions control
systems in new technology vehicles, and to potentially cause these vehicles to fail emission
standards required for vehicles up through 100,000 miles (or more) of operation.
Based on this information, we have concluded that the sensitivity of automotive catalysts
to sulfur has increased to such a degree in vehicle technology currently available, and expected to
be used to meet the proposed Tier 2 standards, that sulfur levels in gasoline must be reduced to
enable these catalysts to operate properly. Not only will harmful emissions from vehicles on the
road today be reduced through lowering gasoline sulfur levels, but the emissions control systems
expected to be used to attain the proposed Tier 2 standards will be significantly impaired if sulfur
levels are not substantially reduced from current levels. A lesser reduction in gasoline sulfur
levels nationwide would likely require us to reduce the stringency of the proposed Tier 2
standards. The same kind of analysis for Tier 0 and Tier 1 vehicles could arguably support a
determination that gasoline sulfur levels significantly impair the emissions control systems of
these vehicles. This is because the effect of sulfur in reducing catalyst efficiency and thereby
increasing emissions exists for all vehicles at issue here (Tier 0 through tier 2), presenting more a
question of difference in degree than in the nature of the effects.
Sulfur in gasoline can also significantly impair the onboard diagnostic (OBD) systems in
current and future vehicles. EPA regulations require all vehicles to be equipped with OBD
systems that monitor catalyst performance and other emissions-related performance, and warns
the vehicle owner if the emissions control system is not functioning properly. In a 1997 staff
paper, EPA concluded that sulfur in gasoline can directly impact OBD systems by affecting the
OBD system's oxygen sensors.1 It is possible that high sulfur levels may impair the OBD system
in such a way that it does not recognize an improperly functioning catalyst, and fails to warn the
owner. In addition, it is not clear that the conditions which may reverse some of sulfur's effect
on the catalyst will also reverse this impact on the OBD system's oxygen sensors. The impact of
sulfur on OBD systems in cleaner technology vehicles may be even more significant, since the
OBD malfunction thresholds are expressed as multiples of the applicable hydrocarbon standard.
Therefore, the impact of sulfur on OBD systems in vehicles meeting more stringent hydrocarbon
standards would be more significant in relative terms.
3. Sulfur Levels that Tier 2 Vehicles Can Tolerate
We believe that Tier 2 vehicles that operate on gasoline will have to use fuel with sulfur
levels no greater than 80 ppm to avoid significant impairment of their emissions control systems.
Furthermore, on average, these vehicles will not be able to be exposed to sulfur levels
substantially greater than 30 ppm to achieve the desired emission performance and avoid
D-4
-------
Appendix D: Legal Authority for Gasoline Sulfur Control
significantly impairing the emissions control system. These conclusions are based on data
collected on vehicles currently sold in California or being developed for sale in California and
the Northeast (the latter under the NLEV program).
The test data from industry test programs and individual automotive and catalyst
manufacturers show that even very low levels of sulfur have some impact on catalyst
performance. The data also show that the greatest increase in emissions comes as the sulfur level
is increased from the lowest levels. At higher sulfur levels, the catalyst is approaching the point
of being saturated with sulfur, and its performance is already impaired, such that an additional
increase in sulfur content has a smaller impact on emissions. This trend applies generally for all
of the regulated pollutants (NMHC, CO, NOx). However, for most vehicles, the impact is
greatest for NOx.
While the overall trends demonstrate that high sulfur levels significantly impair the
emissions control system of newer technology vehicles, the data also shows that some vehicles
are much less sensitive to sulfur than others. The reasons for these vehicle-to-vehicle variations
are not fully understood. We have identified a number of factors involved in the vehicle design
and operation which appear to contribute to the variation. To summarize briefly, sulfur
sensitivity is impacted not only by the catalyst formulation (the types and amounts of precious
metals used in the catalyst) but also by the following factors:
the materials used to provide oxygen storage capacity in the catalyst, as well as the general design
of the catalyst,
the location of the catalyst relative to the engine, which impacts the temperatures inside the
catalyst,
the mix of air and fuel entering the engine over the course of operation, which is varied by the
engine's computer in response to the driving situation and affects the mix of gases entering the
catalyst from the engine, and
the speeds the car is driven at and the load the vehicle is carrying, which also impact the
temperatures experienced by the catalyst.
All of these factors contribute not only to the degree to which sulfur will poison a catalyst, but
also whether and how easily the sulfur will be removed during a vehicle's normal operation.
This cycle of sulfur collection (adsorption) and removal (desorption) in the catalyst is what
ultimately affects sulfur's net impact on emissions and the emissions control system, both short
and long term. Since these factors vary for every vehicle, the sulfur impact varies for every
vehicle to some degree. There is no single factor that guarantees that a vehicle will be very
sensitive or very insensitive to sulfur. None of the data that we have reviewed indicates a vehicle
design which is completely insensitive to sulfur, or even capable of tolerating average sulfur
levels above 30 ppm without a significant impairment of its emissions control system.
D-5
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
Based on the data and information obtained from catalyst manufacturers, we have also
concluded that there are no viable emission control alternatives that could achieve the same level
of emission standards without reducing commercial gasoline sulfur levels, as explained in the
next section.
4. Sulfur Sensitivity of Other Catalysts
Section 21 l(c)(2)(B) requires that, prior to adopting a fuel control based on a significant
impairment to vehicle emissions control systems, EPA consider available scientific and
economic data, including a cost benefit analysis comparing emissions control devices or systems
which are or will be in general use that require the proposed fuel control with such devices or
systems which are or will be in general use that do not require the proposed fuel control. As
described below, we conclude that the emissions control systems expected to be used to meet the
proposed Tier 2 standards would be significantly impaired by operation on high sulfur gasoline.
Our analysis of the available scientific and economic data can be found in the Preamble,
including an analysis of the environmental benefits of the proposed control, an analysis of the
costs and the technological feasibility of controlling sulfur to the proposed levels, and a cost-
benefit analysis of the proposed sulfur control and Tier 2 vehicle emissions standards. Under
Section 21 l(c)(2)(B), EPA is also required to compare the costs and benefits of achieving vehicle
emissions standards through emissions control systems that are not sulfur-sensitive, if any such
systems are or are will be in general use.
We have determined that there are not (and will not be in the foreseeable future) emission
control devices available for gasoline-powered vehicles that can meet the proposed Tier 2
emission standards and would not be significantly impaired by gasoline with high sulfur levels.
All catalysts are sensitive to sulfur to some degree. As explained in Section IV. A of the
Preamble, we cannot identify one or more factors that definitively determine sulfur sensitivity,
because sulfur sensitivity seems to be due to a combination of many factors that vary by vehicle.
Hence, it is not possible to identify alternative designs that can tolerate existing (or even
intermediate) sulfur levels and that can reasonably be expected to be applied to all cars and light
trucks meeting Tier 2 standards.
As described in Section IV. A. of the Preamble, EPA anticipates that all the gasoline
vehicle technologies expected to be used to meet the proposed Tier 2 standards will require the
use of low sulfur gasoline. If we do not control gasoline sulfur to the proposed levels, we will
not be able to set Tier 2 standards as stringent as those we are proposing today. Moreover,
vehicles already on the road would continue to emit at higher levels than they would if operated
on low sulfur fuel. Consequently, EPA concludes that the benefits that would be achieved
through implementation of the proposed vehicle and sulfur control programs cannot be achieved
through the use of emission control technology that is not sulfur-sensitive.
This also means that if EPA were to adopt vehicle emissions control standards without
controlling gasoline sulfur content, the standards would be significantly less stringent than those
D-6
-------
Appendix D: Legal Authority for Gasoline Sulfur Control
proposed today based on what would be technologically feasible with current sulfur levels. The
cost of the vehicle emissions control technology would likely be similar to the costs of meeting
the proposed Tier 2 standards, because the same technologies would be used. However, the
emissions benefits of those technologies would be significantly less than what would be achieved
by the program proposed today, because the emissions control technology for gasoline vehicles
currently in use, and expected to be used in the future, would be significantly impaired by high
sulfur fuel.
5. Effect of Gasoline Sulfur Control on the Use of Other Fuels or Fuel Additives
Section 21 l(c)(2)(C) requires that, prior to prohibiting a fuel or fuel additive, EPA
establish that such prohibition will not cause the use of another fuel or fuel additive "which will
produce emissions which endanger the public health or welfare to the same or greater degree"
than the prohibited fuel or additive. This finding is required by the Act only prior to prohibiting
a fuel or additive, not prior to controlling a fuel or additive. Since EPA is not proposing to
prohibit sulfur in gasoline, but rather to control the levels of sulfur in gasoline, this finding is not
required prior to regulation. However, EPA does not believe that the proposed sulfur control will
result in the use of any other fuel or additive that will produce emissions that will endanger
public health or welfare to the same or greater degree as the emissions produced by gasoline with
uncontrolled sulfur levels.
We believe that gasoline formulated to meet the proposed low sulfur standards will have
a net benefit to public health due to reduced emissions of harmful compounds. The composition
of the emissions from combustion of low sulfur gasolines will be different than the composition
of the emissions from the high sulfur gasolines they replace. Furthermore, other changes to the
composition of the gasoline are likely to accompany the reduction in sulfur content. While some
of these changes may involve increases in the content of certain compounds that tend to lead to
more harmful emission products, we believe that the improved catalyst performance enabled by
the low sulfur fuel will more than offset any slight increase in harmful emissions that would
otherwise result (if sulfur levels remained constant but the other properties were increased).
It is difficult to quantify this impact because it depends on the specific approaches that
each refiner takes to reduce their gasolines' sulfur levels, as well as the composition of the
gasoline overall. However, some general trends can be identified, and based on these trends we
have drawn the conclusion that low sulfur gasoline will pose no new, increased risk to human
health relative to the higher sulfur gasoline it replaces.
Some refiners already make gasolines that meet the proposed standards. Others will
make modest changes in the way in which they blend refinery streams to produce low sulfur
gasoline. But most refiners will have to install some desulfurization technology and/or otherwise
substantially change their operation. If a refiner chooses a traditional route to desulfurize
gasoline, he will likely select a desulfurization technology which has the undesirable side effect
of reducing the octane content of the gasoline streams. To make up that octane, the refiner has
D-7
-------
Tier 2/Sulfur Draft Regulatory Impact Analysis - April 1999
several options. All of these options, whether increasing the aromatics or olefins content of the
gasoline through other processing changes, or through the addition of oxygenates such as ethanol
or MTBE, could lead to increased emissions of air toxics (benzene, 1,3-butadiene, aldehydes) if
the emissions performance of the vehicle catalyst remained constant. However, since low sulfur
gasoline will enable very low emitting catalysts and will improve the performance of existing
catalysts, the catalyst will be able to convert these toxic emissions into less harmful compounds.
Because of the diversity among refineries, it is impossible to estimate with any certainty how
many refiners may choose this route.
If a refiner chooses one of the improved technologies for sulfur removal, the technologies
on which much of our economic analysis for this proposal is based (as discussed in Sections IV.C
and IV.D of the Preamble), there will be less of a need to increase high octane compounds in the
gasoline. These improved technologies are designed to reduce the octane loss that occurs with
the traditional technologies. Because the need to increase high octane components is reduced if
these technologies are used, the net benefit of low sulfur gasoline is even greater, because there
are even fewer toxic compounds for the catalyst to have to convert. (No catalyst yet developed is
able to convert 100 percent of the pollutants that come from the engine.)
D-8
-------
Appendix D: Legal Authority for Gasoline Sulfur Control
Appendix D. References
1. EPA Staff Paper on Gasoline Sulfur Issues (EP A420-R-98-004), May 1998.
D-9
------- |