United States Air and Radiation EPA420-R-99-023
Environmental Protection December 1999
Agency
vvEPA Regulatory Impact
Analysis - Control of Air
Pollution from New Motor
Vehicles: Tier 2 Motor
Vehicle Emissions
Standards and Gasoline
Sulfur Control
Requirements
> Printed on Recycled Paper
-------
EPA420-R-99-023
December 1999
-
Tier 2
Engine Programs and Compliance Division
Office of Mobile Sources
U.S. Environmental Protection Agency
-------
Executive Summary
Executive Summary
Vehicle Standards
Today's action sets new federal emission standards ("Tier 2 standards") for passenger
cars, light trucks, and larger passenger vehicles. The program is designed to focus on reducing
the emissions most responsible for the ozone and particulate matter (PM) impact from these
vehicles — nitrogen oxides (NOx) and non-methane organic gases (NMOG), consisting primarily
of hydrocarbons (HC) and contributing to ambient volatile organic compounds (VOC). The
program will also, for the first time, apply the same set of federal standards to all passenger cars,
light trucks, and medium-duty passenger vehicles. Light trucks include "light light-duty trucks"
(or LLDTs), rated at less than 6000 pounds gross vehicle weight and "heavy light-duty trucks"
(or HLDTs), rated at more than 6000 pounds gross vehicle weight).1 "Medium-duty passenger
vehicles" (or MDPVs) form a new class of vehicles introduced by this rule that includes SUVs
and passenger vans rated at between 8,500 and 10,000 GVWR. The program thus ensures that
essentially all vehicles designed for passenger use in the future will be very clean vehicles.
The Tier 2 standards finalized today will 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 will phase
in beginning in 2004, with the standards to be fully phased in by 2007.2 For heavy LDTs and
MDPVs, the Tier 2 standards will 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 the primary Tier 2 standards will have to meet an interim average standard of 0.30
g/mi NOx, equivalent to the current NLEV standards for LDVs.3 During the period 2004-2008,
heavy LDTs and MDPVs not certified to the final Tier 2 standards will phase in to an interim
program with an average standard of 0.20 g/mi NOx, with those not covered by the phase-in
meeting a per-vehicle standard (i.e., an emissions "cap") of 0.60 g/mi NOx (for FtLDTs) and 0.09
1 A vehicle's "Gross Vehicle Weight Rating," or GVWR, is the curb weight of the vehicle plus its
maximum recommended load of passengers and cargo.
2 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
(larger light trucks are not covered). For further comparison, the standards met by today's Tier 1 vehicles range
from 0.60 g/mi to 1.53 g/mi.
3 There 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. of this
preamble.
iii
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999 Executive Summary
g/mi NOx (for MDPVs). The average standards for NOx will allow manufacturers to comply
with the very stringent new standards in a flexible way, assuring that the average emissions of a
company's production meet the target emission levels while allowing the manufacturer to choose
from several more- and less-stringent emission categories for certification.
We are also setting stringent particulate matter standards that will be especially important
if there is substantial future growth in the sales of diesel vehicles. With higher sales of diesel
cars and light trucks, these vehicles could easily contribute between one-half and two percent of
the PM10 concentration allowed by the NAAQS, with some possibility that the contribution
could be as high as five to 40 percent in some roadside situations with heavy traffic. These
increases would make attainment even more difficult for eight counties which we already predict
to need further emission reductions even without an increase in diesel sales, and would put at risk
another 18 counties which are now within 10 percent of a NAAQS violation. Thus, by including
a more stringent PM standard in the program finalized today, we help address environmental
concerns about the potential growth in the numbers of light-duty diesels on the road - even if
that growth is substantial. The new requirements also include more stringent hydrocarbon
controls (exhaust NMOG and evaporative emissions standards).
Gasoline Sulfur Standards
The other major part of today's action will significantly reduce average gasoline sulfur
levels nationwide. We expect these reductions could begin to phase in as early as 2000, with full
compliance for most refiners occurring by 2006. Refiners will generally install advanced refining
equipment to remove sulfur during the production of gasoline. Importers of gasoline will be
required to import and market only gasoline meeting the sulfur limits. Temporary, less stringent
standards will apply to refineries who produce fuel for use in the Geographic Phase-in Area4
through 2006 and a few small refiners through 2007.
This significant new control of gasoline sulfur content will have two important effects.
The lower sulfur levels will enable the much-improved emission control technology necessary to
meet the stringent vehicle standards of today's rule to operate effectively over the useful life of
the new vehicles. In addition, as soon as the lower sulfur gasoline is available, all gasoline
vehicles already on the road will have reduced emissions-from less degradation of their catalytic
converters and from fewer sulfur compounds in the exhaust.
Today's action will require that most refiners and importers meet a corporate average
gasoline sulfur standard of 120 ppm and a cap of 300 ppm beginning in 2004. By 2006, the cap
will be reduced to 80 ppm and most individual refineries must produce gasoline averaging no
"Alaska, Colorado, Idaho, Montana, New Mexico, North Dakota, Utah, and Wyoming
iv
-------
Executive Summary
more than 30 ppm sulfur. The program builds upon the existing regulations covering gasoline
composition 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.
We intend for the credit program to ease compliance uncertainties by providing refiners the
flexibility to phase in early controls in 2000-2003 and use credits gained in these years to delay
some control until as late as 2006. As finalized today, the program will achieve the needed
environmental benefits while providing substantial flexibility to refiners.
Cost-Effectiveness of the Tier2/Sulfur Program
A comparison of the costs of our program with the emission reductions it is estimated to
achieve leads us to conclude that it is a cost-effective means of reducing pollution. As shown in
Chapter VI, the cost-effectiveness of Tier 2/gasoline sulfur falls within the range of cost-
effectiveness of 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 finalizing 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 are expected to result from our standards near full implementation in 2030. We
estimate that our Tier 2/gasoline sulfur standards would, in the long term, result in substantial
benefits, such as: the yearly avoidance of approximately 4300 premature deaths, approximately
2300 cases of bronchitis, and significant numbers of hospital visits, lost work days, and multiple
respiratory ailments (especially those that affect children). Our standards will also produce
welfare benefits relating to agricultural crop damage, visibility, and nitrogen deposition in rivers
and lakes. Total monetized benefits, however, are driven primarily by the value placed on the
reductions in premature deaths. The adoption of a value for the projected reduction in the risk of
premature mortality is the subject of continuing discussion within the economic and public policy
analysis community within and outside the Administration. In response to the sensitivity on this
issue, we provide estimates reflecting two alternative approaches. The first approach-supported
by some in the above community and preferred by EPA-uses a Value of a Statistical Life (VSL)
approach developed for the Clean Air Act Section 812 benefit-cost studies. This VSL estimate
of $5.9 million (1997$) was derived from a set of 26 studies identified by EPA using criteria
established in Viscusi (1992), as those most appropriate for environmental policy analysis
applications.
An alternative, age-adjusted approach is preferred by some others in the above
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999 Executive Summary
community both within and outside the Administration. This approach was also developed for
the Section 812 studies and addresses concerns with applying the VSL estimate -reflecting a
valuation derived mostly from labor market studies involving healthy working-age manual
laborers- to PM-related mortality risks that are primarily associated with older populations and
those with impaired health status. This alternative approach leads to an estimate of the value of a
statistical life year (VSLY), which is derived directly from the VSL estimate. It differs only in
incorporating an explicit assumption about the number of life years saved and an implicit
assumption that the valuation of each life year is not affected by age.5 The mean VSLY is
$360,000 (1997$); combining this number with a mean life expectancy of 14 years yields an age-
adjusted VSL of $3.6 million (1997$).
Both approaches are imperfect, and raise difficult methodological issues which are
discussed in depth in the recently published Section 812 Prospective Study, the draft EPA
Economic Guidelines, and the peer-review commentaries prepared in support of each of these
documents. For example, both methodologies embed assumptions (explicit or implicit) about
which there is little or no definitive scientific guidance. In particular, both methods adopt the
assumption that the risk versus dollars trade-offs revealed by available labor market studies are
applicable to the risk versus dollar trade-offs in an air pollution context.
EPA currently prefers the VSL approach because, essentially, the method reflects the
direct, application of what EPA considers to be the most reliable estimates for valuation of
premature mortality available in the current economic literature. While there are several
differences between the labor market studies EPA uses to derive a VSL estimate and the
particulate matter air pollution context addressed here, those differences in the affected
populations and the nature of the risks imply both upward and downward adjustments. For
example, adjusting for age differences may imply the need to adjust the $5.9 million VSL
downward as would adjusting for health differences, but the involuntary nature of air pollution-
related risks and the lower level of risk-aversion of the manual laborers in the labor market
studies may imply the need for upward adjustments. In the absence of a comprehensive and
balanced set of adjustment factors, EPA believes it is reasonable to continue to use the $5.9
million value while acknowledging the significant limitations and uncertainties in the available
literature. Furthermore, EPA prefers not to draw distinctions in the monetary value assigned to
the lives saved even if they differ in age, health status, socioeconomic status, gender or other
characteristic of the adult population.
Those who favor the alternative, age-adjusted approach (i.e. the VSLY approach)
Specifically, the VSLY estimate is calculated by amortizing the $5.9 million mean VSL estimate over the 35 years of
life expectancy associated with subjects in the labor market studies. The resulting estimate, using a 5 percent discount rate, is
$360,000 per life-year saved in 1997 dollars. This annual average value of a life-year is then multiplied times the number of
years of remaining life expectancy for the affected population (in the case of PM-related premature mortality, the average number
of $ life-years saved is 14.
vi
-------
Executive Summary
emphasize that the value of a statistical life is not a single number relevant for all situations.
Indeed, the VSL estimate of $5.9 million (1997 dollars) is itself the central tendency of a number
of estimates of the VSL for some rather narrowly defined populations. When there are
significant differences between the population affected by a particular health risk and the
populations used in the labor market studies - as is the case here - they prefer to adjust the VSL
estimate to reflect those differences. While acknowledging that the VSLY approach provides an
admittedly crude adjustment (for age though not for other possible differences between the
populations), they point out that it has the advantage of yielding an estimate that is not
presumptively biased. Proponents of adjusting for age differences using the VSLY approach
fully concur that enormous uncertainty remains on both sides of this estimate - upwards as well
as downwards - and that the populations differ in ways other than age (and therefore life
expectancy). But rather than waiting for all relevant questions to be answered, they prefer a
process of refining estimates by incorporating new information and evidence as it becomes
available.
The results indicate that using EPAs preferred approach to valuing reductions in
premature mortality, total monetary benefits realized after nearly a full turnover of the fleet to
Tier 2 vehicles would be approximately $25.2 billion in 2030. Using the alternative, age-
adjusted approach to value reductions in premature mortality yields total monetized benefits of
$13.9 billion in 2030. Comparing this estimate of the economic benefits with the adjusted cost
estimate indicates that the net economic benefit of the tier 2/gasoline sulfur standards to society
are approximately $20 billion in 2030. Using the alternative, age-adjusted approach to valuing
premature mortality, net benefits are approximately $8.5 billion. Due to the uncertainties
associated with this estimate of net benefits, it should be considered along with other components
of this RIA, such as: total cost, cost-effectiveness, and other considerations of benefits and costs
that could not be monetized.
vn
-------
Table of Contents
Table of Contents
Executive Summary iii
List of Acronyms xvi
List of Tables xx
List of Figures xxvii
Chapter I: Introduction 1-1
A. Background Information for Today's Final Rule 1-1
B. Overview of the Final Rule 1-2
1. Vehicle Emission Standards 1-3
2. Gasoline Sulfur Standards 1-4
Chapter II: Health and Welfare Concerns D-l
A. Health and Welfare Effects of Ozone D-l
B. Health and Welfare Effects of Particulate Matter H-4
C. Carbon Monoxide D-6
D. Visibility and Regional Haze D-7
Chapter n. References D-9
Chapter HI: Environmental Impact ffi-1
A. Inventory Impacts of Tier 2/Sulfur III-l
1. NOx m-4
a. Light-Duty NOx Trends Without Tier 2/Sulfur ffl-4
b. NOx Reductions Due To Tier 2/Sulfur IH-10
2. VOC IH-16
a. Light-Duty VOC Trends Without Tier 2/Sulfur IE-16
b. VOC Reductions Due To Tier 2/Sulfur IH-21
3. SOx IH-27
a. Light-Duty SOx Trends Without Sulfur Control IH-27
b. SOx Reductions Due To Sulfur Control IH-29
4. Particulate Matter ffl-30
a. "No Growth" Diesel Sales Scenario IH-31
b. "Increased Growth" Sales Scenario ffi-35
B. Ozone: Baseline Nonattainment and Program Impacts ID-41
1. General Description of the EPA Ozone Modeling Used In This
IX
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Rulemaking ffi-41
a. Modeling Methodology in-41
b. Determining Need for Additional Emissions Reductions ... 111-45
2. Ozone Reductions Expected from this Rule 111-46
3. Ozone Modeling and Analysis in 1-Hour State Implementation Plan
Submittals and Other Local Ozone Modeling in-47
a. Overview in-47
b. CAA Requirements and EPA Policy ffi-51
c. Local Ozone Modeling in SIP Submissions 111-53
d. Conclusions from the Local Modeling in SIP Submittals . . . 111-56
e. Other Local Ozone Modeling in-58
f. Need for Further Reductions in Emissions in Order to Attain and
Maintain 111-59
C. Particulate Matter and Visibility/Regional Haze ID-61
1. Particulate Matter ffl-61
a. Background on Particulate Matter ID-61
b. PM10 Role of Cars and Light Trucks ffl-63
c. Current PM10 Nonattainment 111-67
d. Future Nonattainment 111-68
e. Diesel PM IH-73
f. Reductions In Ambient PM ffl-80
2. Visibility/Regional Haze IE-SI
D. Air Toxics ffl-82
1. Health Effects IH-82
a. Benzene in-83
b. 1,3-Butadiene IH-84
c. Formaldehyde 111-85
d. Acetaldehyde IH-87
e. Diesel Particulate Matter ffl-87
2. Assessment of Emissions and Exposure 111-89
a. Emissions Modeling 111-90
b. Nationwide Toxic Emissions Estimates - Baseline Scenario 111-96
c. Exposure - Baseline Scenario in-97
d. Impact of Potential Vehicle and Fuel Controls in-99
E. Carbon Monoxide ffl-103
Chapter HI References ffl-106
Chapter IV: Technological Feasibility IV-1
A. Feasibility of Tier 2 Exhaust Emission Standards for Vehicles IV-1
1. NMOG and NOx Emissions from Gasoline-Fueled Vehicles IV-1
a. Technology Description IV-2
-------
Table of Contents
b. Data Supporting Tier 2 Technical Feasibility IV-14
c. Lean-Burn Technology IV-32
2. CO Emissions from Gasoline Fueled Vehicles IV-33
3. Formaldehyde Emissions from Gasoline Fueled Vehicles IV-33
4. Evaporative Emissions IV-34
5. Diesel Vehicles IV-35
B. Feasibility of Removing Sulfur from Gasoline IV-39
1. Source of Gasoline Sulfur IV-39
2. Current Levels of Sulfur in Gasoline IV-40
3. Feasibility of Meeting the Final Gasoline Sulfur Standards IV-42
4. Meeting a Low Sulfur Gasoline Standard IV-43
a. Background IV-43
b. FCC Feed Hydrotreating IV-46
c. FCC Gasoline Hydrotreating IV-47
5. Expected Desulfurization Technology to be Used by Refiners IV-53
6. Feasibility for a Low Gasoline Sulfur Standard in 2004 IV-54
7. Phase In of Compliance with the Proposed Sulfur Standards and Early
Credit Generation IV-58
Chapter IV References IV-66
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-4
c. Hardware Costs for Evaporative Emissions Control V-17
d. Assembly Costs V-19
e. Development and Capital Costs V-20
f. Total Near-term and Long-term Manufacturer Costs V-22
2. Tier 2 Vehicle Consumer Costs V-26
3. Annual Total Nationwide Costs for Tier 2 Vehicles V-27
a. Overview of Nationwide Vehicle Costs V-27
b. Methodology V-28
c. Estimates of Total Nationwide Vehicle Costs by Vehicle Class
V-29
B. Gasoline Desulfurization Costs V-38
1. Overview of Changes Since the NPRM V-39
2. Cost Estimation Methodology V-41
a. Technology and Cost Inputs V-41
b. Capital Costs V-46
c. Fixed Operating Cost V-48
xi
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
d. Variable Operating Cost V-48
e. Determination of Blendstock Sulfur Levels V-50
f. Phase-In Desulfurization V-58
g. Decreasing Costs in Future Years V-59
3 The Cost of Desulfurizing Gasoline V-60
a. EPA Costs V-60
b. Other Low Sulfur Cost Studies V-64
c. Cost of Meeting a 5 ppm Averaging Standard V-70
4. Other Effects of This Program V-74
a. Effect of the Cap Standard V-74
b. Other Effects on the Refining Industry V-76
5. Per Vehicle Life-Cycle Fuel Costs V-80
6. Aggregate Annual Fuel Costs V-83
a. Methodology V-84
b. Explanation of Results V-85
C. Combined Vehicle and Fuel Costs V-89
1. Combined Costs Per Vehicle V-89
2. Combined Total Annual Nationwide Costs V-91
Chapter V References V-92
Chapter VI: Cost-Effectiveness VI-1
A. Overview of the Analysis VI-1
1. Temporal and Geographic Applicability VI-2
2. Baselines VI-3
B. Costs VI-5
1. Near and Long-Term Cost Accounting VI-5
2. Vehicle and Fuel Costs VI-6
3. Cost Crediting for PM and SO2 VI-6
C. Emission Reductions VI-8
1. NOx and NMHC VI-9
2. Irreversibility VI-12
3. Primary Particulate Matter VI-15
4. Sulfur Dioxide VI-16
D. Aggregate Cost-Effectiveness VI-17
E. Results VI-17
APPENDIX VI-A : Discounted Lifetime Tonnage Values for Exhaust Emissions VI-22
APPENDIX VI-B : Discounted Lifetime Tonnage Values for Evaporative Emissions . . VI-26
APPENDIX VI-C : Aggregate Annual Tons and Costs VI-27
xn
-------
Table of Contents
Chapter VI. References VI-28
Chapter VH: Benefit-Cost Analysis VH-1
A. Emissions VII-7
B. Air Quality Impacts VH-8
1. Ozone Air Quality Estimates VD-10
a. Modeling Domain VD-10
b. Simulation Periods VII-11
c. Converting UAM-V Outputs to Full-Season Profiles for Benefits
Analysis VD-13
d. Ozone Air Quality Results VH-13
2. PM Air Quality Estimates VH-16
a. Development of the S-R Matrix VII-16
b. Fugitive Dust Adjustment Factor VII-17
c. Normalizing S-R Matrix Results to Observed Data VII-18
d. PM Air Quality Results VH-19
3. Visibility Degradation Estimates VD-21
4. Nitrogen Deposition Estimates VD-24
C. Benefit Analysis VH-25
1. Methods for Estimating Benefits from Air Quality Improvements . VII-25
2. Methods for Describing Uncertainty VD-28
a. Unquantifiable Environmental Benefits and Costs VD-32
b. Projected Population and Income Growth VD-33
D. Assessment of Human Health Benefits VD-33
1. Estimating Baseline Incidences for Health Effects VD-34
2. Accounting for Potential Health Effect Thresholds VD-34
3. Quantifying and Valuing Individual Health Endpoints VD-35
a. Premature Mortality: Quantification VD-40
b. Premature Mortality: Valuation VII-42
c. Chronic Bronchitis: Quantification VD-46
d. Chronic Bronchitis: Valuation VD-47
e. Chronic Asthma: Quantification VD-47
f. Chronic Asthma: Valuation VII-48
g. Hospital Admissions: Quantification VII-49
h. Hospital Admissions: Valuation VII-49
i. Other Health Effects: Quantification VH-50
j. Other Health Effects: Valuation VH-53
k. Lost Worker Productivity: Quantification and Valuation . . VII-53
1. Estimated Reductions in Incidences of Health Endpoints and
Associated Monetary Values VII-53
E. Assessment of Human Welfare Benefits VD-56
1. Visibility Benefits VH-56
xiii
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
2. Agricultural and Forestry Benefits VII-60
3. Benefits from Reductions in Materials Damage VII-62
4. Benefits from Reduced Ecosystem Damage VII-62
5. Estimated Values for Welfare Endpoints VII-65
F. Total Benefits VH-66
G. Summary of Cost Results VII-76
I. Comparison of Costs and Benefits VII-78
H. References VH-80
Appendix VII-A
Supplementary Benefit Estimates and Sensitivity Analyses of Key Parameters in the
Benefits Analysis VII-89
A. Introduction and Overview VII-89
B. Supplementary Benefit Estimates VII-89
C. Sensitivity Analyses VII-92
1. Alternative Lag Structures VII-92
2. PM Health Effect Threshold VH-94
D. References VH-97
Chapter VHI: Regulatory Flexibility VIH-1
A. Requirements of the Regulatory Flexibility Act Vffi-1
B. Description of Affected Entities Vffi-2
1. Small Refiners VIH-3
2. Small Petroleum Marketers Vffl-4
3. Small Certifiers of Covered Vehicles VIH-4
C. Projected Costs of the Proposed Gasoline Sulfur Standards Vffi-5
D. The Types and Number of Small Entities to Which the Proposed Rule Would
Apply VIH-5
E. Projected Reporting, Recordkeeping, and Other Compliance Requirements of the
Proposed Rule VIH-6
F. Other Relevant Federal Rules Which May Duplicate, Overlap, or Conflict with the
Proposed Rule VIH-7
G. Regulatory Alternatives Vffi-7
1. Small Refiner Interim Sulfur Standards VIH-7
a. Extensions Beyond 2007 for Qualifying Small Refiners .... Vffi-9
b. Compliance Plans for Demonstrating a Commitment to Produce
Low Sulfur Gasoline VIH-10
2. Small Certifiers of Covered Vehicles VIH-11
Chapter VHI. References VIH-13
xiv
-------
Table of Contents
Appendix A: 47-State and Four-Cities Emission Inventories A-l
Appendix B: Evidence Supporting the Irreversibility of Sulfur's Emission Impact B-l
A. Exhaust Emission Sensitivity to Sulfur Content B-2
B. Theory Supporting the Reversibility and Irreversibility of Sulfur's Emission
Impact B-9
C. Results of Sulfur Irreversibility Test Programs B-12
1. Pre-SFTP LEVs B-12
a. Coordinating Research Council (CRC) Sulfur Irreversibility
Program B-12
b. American Petroleum Institute Sulfur Irreversibility Program . B-l4
c. Ford Sulfur Irreversibility Program B-20
d. EPA Sulfur Irreversibility Test Program B-21
e. ATL Sulfur Irreversibility Program B-22
f. Irreversibility for Long-Term Sulfur Exposure B-23
2. SFTP-compliant LEV and Tier 2 vehicles B-24
a. DaimlerChrysler Sulfur Irreversibility Program B-24
b. EPA Sulfur Reversibility Program B-25
c. ATL Sulfur Reversibility Program B-26
D. Criteria for Evaluating Sulfur Reversibility Data B-26
E. Projected Levels of Sulfur Irreversibility In-Use B-29
1. Pre-SFTP Vehicles B-29
2. SFTP-Compliant LEV and Tier 2 vehicles B-30
Appendix C: Refinery Energy and Global Warming Impacts and Emissions 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 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-4
3. Sulfur Levels that Tier 2 Vehicles Can Tolerate D-5
4. Sulfur Sensitivity of Other Catalysts D-7
5. Effect of Gasoline Sulfur Control on the Use of Other Fuels or Fuel
Additives D-9
xv
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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
air/fuel ratio
acute my el old 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
xvi
-------
Table of Contents
GPA
GVWR
HAPEM
HC
HDV
HEGO
I/M
ICI
IRFA
LOT
LDV
LEV
LPG
MDPV
MECA
MLE
MRAD
MSCF
MTBE
NAAQS
NAPAP
NFRAQS
NLEV
NMHC
NMOG
NO2
NOx
NPC
NPRA
NPRM
OAQPS
OBD
OMB
Geographic Phase-in Area
gross vehicle weight rating
Hazardous Air Pollutant Exposure Model
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 passenger 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
xvn
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
QMS
QMS
ORNL
OSTP
OTAG
PADD
PCM
Pd
PM
PNGV
ppm
Pt
R&D
RFA
RfC
RFG
Rh
ROI
ROTR
RPE
RVP
S-R Matrix
S&PDRI
SAB
SBA
SB ARP or the Panel
SBREFA
SCR
SER
SFTP
SIC
SIGMA
SIP
Office of Mobile Sources
Office of Mobile Sources
Oak Ridge National Laboratory
(White House) Office of Science and Technology Policy
Ozone Transport Assessment Group
Petroleum Administrative Districts for Defense
powertrain control module
palladium
paniculate 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
xvin
-------
Table of Contents
SO2
SOx
SULEV
SVM
SVM
SwRI
TOG
TW
UAM
UCL
UEGO
ULEV
UV
VMT
VNA
voc
WLD
WTP
sulfur dioxide
oxides of sulfur
super ultra low emission vehicle
small volume manufacturer
small volume manufacturer (of vehicles)
Southwest Research Institute
total organic gases
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
xix
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
List of Tables
Table IE. A-1. 47-State Light Duty NOx Emissions Without Tier 2/Sulfur IH-6
Table IH.A-2. Light-Duty Contribution to Total NOx Inventory Without Tier 2/Sulfur ... ffl-10
Table IH.A-3. 47-State Light-Duty NOx Reductions Due To Tier 2/Sulfur IH-11
Table ffl.A-4. Light-Duty Contribution to Total NOx Inventory With Tier 2/Sulfur ffl-16
Table IH.A-5. 47-State Light-Duty VOC Emissions Without Tier 2/Sulfur IH-17
Table ffl.A-6. Light-Duty Contribution to Total VOC Inventory Without Tier 2/Sulfur . . . ffl-21
Table IH.A-7. 47-State Light-Duty VOC Reductions Due to Tier 2/Sulfur IH-22
Table ffl.A-8. Light-Duty Contribution to Total VOC Inventory With Tier 2/Sulfur ffl-27
Table IH.A-9. 47-State SOx Emissions Without Sulfur Control IH-28
Table IE. A-10. 47-State Light-Duty SOx Reductions Due To Sulfur Control ffl-29
Table ffl.A-11. 47 State Light-Duty Direct Exhaust PM10 Emissions Without Tier 2/Sulfur
No Growth in Diesel Sales ffl-32
Table IH.A-12. 47-State Light-Duty Direct Exhaust PM10 Reductions Due To Tier 2/Sulfur
No Growth in Diesel Sales in-34
Table IHA-13. Diesel LDT Sales Penetration Under Increased Growth Scenario in-36
Table ffl.A-14. 47 State Light-Duty Direct Exhaust PM10 Emissions Without Tier 2/Sulfur
Increased Diesel Growth Scenario in-37
Table ffl.A-15. 47-State Light-Duty Direct Exhaust PM10 Reductions Due To Tier 2/Sulfur
Increased Diesel Growth Scenario in-39
Table ffl.B-1. Comparison of eastern U.S. regional model performance statistics between the
Ozone Transport Assessment Group (OTAG) modeling used to support the NOX SIP call
and the Tier 2/Sulfur modeling. The units are percentages ffl-44
Table ffl.C-1. NFRAQS Compositional Analysis of PM25 Samples ffl-65
Table in.C-2. Percentage of PM25 Coming from Gasoline Vehicles in-65
Table ffl.C-3. Percentage of Total PM25 From Gasoline Vehicles in-66
Table ffl.C-4. Gasoline Vehicle/Engine Contribution to Ambient PM - from source
apportionment reports 111-67
Table ffl.C-5. Fifteen PM10 Nonattainment Areas Violating the PM10 NAAQS in 1996-1998
IH-69
Table ffl.C-6. Eight areas with a high risk of failing to attain and maintain
the PM10 NAAQS without further reductions in emissions ffl-71
Table ffl.C-7. Five areas with a significant risk of failing to attain and maintain
the PM10 NAAQS without further reductions in emissions 111-72
Table ffl.C-8. Thirteen metropolitan statistical area counties with 1997 and/or 1998 ambient
PM10 concentrations within 10 percent of the annual or 24-hour the PM10 NAAQS. 111-73
Table ffl.C-9. Ambient Diesel PM Concentrations Reported from Chemical Mass Balance
Modeling ffl-77
xx
-------
Table of Contents
Table in.D-1. Areas Included in Toxic Emissions Modeling 111-93
Table in.D-2. Example of Data File Format for Toxic Adjustment Factors 111-94
Table in.D-3. 47 State Highway Vehicle Toxic Emissions (tons)
In 1990, 1996, 2007, and 2020, for Baseline Scenarios ffl-99
Table IHD-4. Average 47 State Highway Vehicle Toxic Exposure (|ig/m3)
In 1990, 1996, 2007, and 2020, for Baseline Scenarios ffl-100
Table IHD-5. Percentage of sales fleet expected to be diesel in each respective year under "most
likely" scenario, as estimated in A. D. Little, Inc. report ffi-102
Table in.D- 6. 47 State Highway Vehicle Toxic Emissions (tons)
in 2007, for Various Scenarios III-103
Table IHD-7. 47 State Highway Vehicle Toxic Emissions (tons)
in 2020, for Various Scenarios ffl-103
Table IHD-8. Average 47 State Highway Vehicle Toxic Exposures for the Entire Population
(|ig/m3) in 2007, for Various Scenarios III-104
Table IHD-9. Average 47 State Highway Vehicle Toxic Exposures for the Entire Population in
2007, for Various Exposures ffl-104
Table IV-1. Emission Control Hardware and Techniques
Projected to Meet Tier 2 Vehicle Standards IV-2
Table IV-2. Number of 1999 Model Year Engine Families with One or More Engine/Vehicle
Configurations with Low Full-life NOx Levels IV-15
Table IV-3. 1999 MY Engine Families with One or More Vehicle Configurations
with Full-life NOx Certification Levels at or below 0.07 g/mile NOx IV-16
Table IV-4. 2000 MY Engine Families with One or More Vehicle Configurations
with Full-life NOx Certification Levels at or below 0.07 g/mile NOx IV-19
Table IV-5. MECA Test Program: Emissions with Catalysts Aged to 100,000 Miles IV-23
Table IV-6. CARS Production LEV LDV Passenger Car Emission Data IV-23
Table IV-7. ARE Modified Passenger Car Emission Data IV-24
Table IV-8. CARS Ford Expedition Baseline Emission Test Results IV-24
Table IV-9. CARB Expedition Emission Results with Advanced Catalyst Systems IV-25
Table IV-10: EPA Test Vehicle Specifications IV-26
Table IV-11: Catalyst Specifications IV-27
Table IV-12. CO Emissions from California LEVs IV-33
Table IV-13. Formaldehyde Emissions from California LEVs IV-34
Table IV-14. Estimated Average Sulfur Levels by PADD and for the Nation IV-42
Table IV-15. Projected Use of Desulfurization Technology Types by
Refiners During the Phase-in Period IV-54
Table IV-16. Leadtime Required Between Promulgation of the Final Rule and Implementation
of the Gasoline Sulfur Standard (years) IV-55
Table IV-17. Number of U.S. Refineries with 1998 Sulfur Averages Falling Into theSpecified
Range of Sulfur Content (ppm) IV-60
Table IV-18. Number of New Desulfurization Units Operating by January 1 of Indicated Year
and National Pool Average Sulfur Levels Under the Final Sulfur Standards IV-63
xxi
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-l. Main or Underfloor Catalyst Cost Breakdown V-7
Table V-2. Total Estimated Per Vehicle Manufacturer
Incremental Hardware Costs for the Tier 2 Standards V-l 1
Table V-3. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDV Compared to
NLEV LDV V-12
Table V-4. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT1 Compared to
NLEV LDT1 V-13
Table V-5. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT2 Compared to
NLEV LDT2 V-14
Table V-6. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT3 Compared to
Current LDT3s V-15
Table V-7. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT4s and MDPVs
Compared to Current Vehicles V-16
Table V-8. Potential Evaporative Improvements and Their Costs to Manufacturers V-l9
Table V-9. Per Vehicle Fixed Costs V-23
Table V-10. Total Per Vehicle Manufacturer Costs - Near Term V-23
Table V-l 1. Long-term Total Incremental Per Vehicle Manufacturer Costs V-26
Table V-12. Incremental Per Vehicle Costs to Consumers for Tier 2 Vehicles V-27
Table V-13. Estimated Annual Nationwide Costs V-29
Table V-14. Estimated Annual 49-State Vehicle Sales V-30
Table V-15. Projected Overall Industry Phase-in of Tier 2 Vehicles and Improved Evaporative
Emissions Controls For Purposes of the Aggregate Cost Analysis V-31
Table V-16. Annual Nationwide Costs For Tier 2 LDVs V-33
Table V-17. Annual Nationwide Costs For Tier 2 LDTls V-34
Table V-18. Annual Nationwide Costs For Tier 2 LDT2s V-35
Table V-19. Annual Nationwide Costs For Tier 2 LDT3s V-36
Table V-20. Annual Nationwide Costs For Tier 2 LDT4s V-37
Table V-21 (A). Annual Nationwide Costs For Tier 2 LDVs, LDTs and MDPVs V-38
Table V-21 (B). Non-Annualized Nationwide Vehicle Costs For Tier 2 LDVs, LDTs and
MDPVs V-39
Table V-23. Projected Use of Desulfurization Technology Types by Refiners During the Phase-
in Period V-44
Table V-24. Process Operations Information for FCC Naphtha Desulfurization Processes V-45
Table V-25. Process Operations Information for Additional Units
used for Desulfurization Cost Analysis V-47
Table V-26. Offsite and Location Factors Used for Estimating Capital Costs V-48
Table V-27. Economic Cost Factors Used in Calculating the Capital Amortization Factor . V-49
Table V-28. Summary of Costs Taken From EIA and NPC Data Tables V-51
Table V-29. PADD 1 Gasoline Blendstock and Pool Sulfur Levels and Pool Fractions . . . V-54
Table V-30. PADD 2 Gasoline Blendstock and Pool Sulfur Levels and Pool Fractions . . . V-55
Table V-31. PADD 3 Gasoline Blendstock and Pool Sulfur Levels and Pool Fractions . . . V-56
Table V-32. PADD 4 Gasoline Blendstock and Pool Sulfur Levels and Pool Fractions . . . V-57
xxii
-------
Table of Contents
Table V-33. PADD 5 Outside of California Gasoline Blendstock
and Pool Sulfur Levels and Pool Fractions V-58
Table V-34. Projected Volume of Gasoline Produced by an Average Refinery in each PADD and
Projected Gasoline Consumption for the U.S.* in 2004 V-59
Table V-35. Cumulative Fraction of the Gasoline Pool Desulfurized by PADD and by Year V-60
Table V-36. Estimated U.S. Aggregate Operating and Capital Cost, and Per-Gallon Cost of
Desulfurizing Gasoline to 30 ppm V-62
Table V-37. Post Phase-in Cost (year 2008) of Desulfurizing Gasoline to 30 ppm
Based on Different Capital Amortization Rates V-64
Table V-38. Estimated Average Per-Refinery and Aggregate Capital and Operating Cost of
Desulfurizing Gasoline to 30 ppm V-65
Table V-39. API Gasoline Desulfurization Estimate, Adjusted and Compared to EPA's . V-66
Table V-40. NPRA PADD 4 Gasoline Desulfurization Estimate, Adjusted and Compared to
EPA's V-68
Table V-41. AIAM Gasoline Desulfurization Estimate for PADD 4, Adjusted and Compared to
EPA's V-69
Table V-42. DOE Gasoline Desulfurization Estimate, Adjusted and Compared to EPA's . V-70
Table V-43. Process Operation Information for Deep Desulfurization of FCC Naphtha .. V-73
Table V-44. Other Refinery Process Changes Potentially Needed to Meet a 5 ppm Sulfur
Standard V-74
Table V-45. Estimated Cost of Meeting a 5 ppm Sulfur Standard V-75
Table V-46. Undiscounted Per-vehicle Costs of Low Sulfur Gasoline V-83
Table V-47. Discounted Per-vehicle Costs of Low Sulfur Gasoline V-84
Table V-48. Fleet Average Per-vehicle Costs V-84
Table V-49. Summary of the Increased Annualized Social Cost of Gasoline
as a Result of the Tier 2 Gasoline Sulfur Controls V-85
Table V-50. Calculation of Gasoline Consumption by Highway Sources V-88
Table V-51. Aggregate Annualized Fuel Costs per Year from 2004 to 2030 V-89
Table V-52. Total Incremental Per Vehicle Costs to Consumers
Over the Life of a Tier 2 Vehicle V-90
Table V-53. Total Annualized Costs to the Nation for
Tier 2 Vehicles and Low Sulfur Gasoline V-92
Table VI-1. Fleet-average, Per-vehicle Costs Used in Cost-effectiveness VI-6
Table VI-2. Fleet Average Per-vehicle Costs Used in Cost-effectiveness VI-8
Table VI-3. Weighting Factors for NOx and NMHC Lifetime Tonnage Values VI-11
Table VI-4. Vehicle Class Sales Weighting Factors VI-11
Table VI-5. Fleet-average, Per-vehicle Discounted
Lifetime Tons for the NLEV Baseline VI-12
Table VI-6. Fleet-average, Per-vehicle Discounted
Lifetime Tons for Tier 2 Standards VI-12
Table VI-7. Fleet-average, Per-vehicle Discounted
Lifetime Tons Used in Cost-effectiveness Analysis VI-15
xxiii
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table VI-8. Per-vehicle cost-effectiveness of the Tier 2/gasoline sulfur standards VI-18
Table VI-9. Aggregate cost-effectiveness of the standards VI-18
Table VI-10. Cost-effectiveness of Previously Implemented
Mobile Source Programs (Costs Adjusted to 1997 Dollars) VI-19
Table VII-1. Human Health and Welfare Effects of Pollutants Affected by the Tier 2/Gasoline
Sulfur Rule VH-4
Table VII-2. Summary of UAM-V Derived Hourly Ozone Air Quality for 2030 Base Case and
Change Due to Tier 2 Standards VH-14
Table VII-3. Summary of 2030 Base Case PM Air Quality and Changes Due to Tier 2 Standards
VH-20
Table VII-4. Summary of Absolute and Relative Changes in PM Air Quality Due to Tier 2
Standards VH-21
Table VII-5. Summary of 2030 Visibility Degradation Estimates by Region: Residential
VH-23
Table VII-6. Summary of 2030 Visibility Degradation Estimates by Region: Recreational
(Annual Average Deciviews) VD-23
Table VII-7. Summary of 2030 Nitrogen Deposition in Selected Estuaries and Changes Due to
Tier 2 Standards VH-25
Table VII-8. Primary Sources of Uncertainty in the Benefit Analysis VII-31
Table VII-9. Endpoints and Studies Included in the Primary Analysis VD-36
Table VII-10. Unit Values Used for Economic Valuation of Health Endpoints VII-38
Table VH-11. Summary of Mortality Valuation Estimates3 VH-44
Table VII-12. Estimated Annual Health Benefits Associated With Air Quality Changes
Resulting from the Tier 2/Gasoline Sulfur Rule in 2030 VII-55
Table VII-13. Reduction Goals and Nitrogen Loads to Selected Eastern Estuaries
(tons per year) VD-63
Table VII-14. Estimated Annual Reductions in Nitrogen Loadings in Selected Eastern Estuaries
for the Final Tier 2/Gasoline Sulfur Rule in 2030 VH-64
Table VII-15. Estimated Annual Monetary Values for Welfare Effects Associated With
Improved Air Quality Resulting from the Tier 2/Gasoline Sulfur Rule in 2030 . . . VII-66
Table VII-16. EPA Preferred Estimate of Annual Quantified and Monetized Benefits Associated
With Improved Air Quality Resulting from the Tier 2/Gasoline Sulfur Rule in 2030
VH-70
Table VD-17. Final Tier 2/Gasoline Sulfur Rule: 2030 Monetized Benefits Estimates for
Alternative Premature Mortality Valuation Approaches VII-71
Table VII-18. Alternative Benefits Calculations for the Tier 2/Gasoline Sulfur Rule in 2030
VH-73
Table VH - 19. Adjusted Cost for Comparison to Benefits VH-78
Table VII-20. 2030 Annual Monetized Costs, Benefits, and Net Benefits for the Final Tier
2/Gasoline Sulfur Rule: EPA Preferred Estimates Using the Value of Statistical Lives
Saved Approach to Value Reductions in Premature Mortality3 VII-80
Table VII-21. 2030 Annual Monetized Costs, Benefits, and Net Benefits for the Final Tier
xxiv
-------
Table of Contents
2/Gasoline Sulfur Rule: Alternative Estimates Using the Value of Statistical Life Years
Saved Approach to Value Reductions in Premature Mortalitya VII-80
Table VII-A-1
Supplemental Benefit Estimates for the Final Tier 2 Rule for the 2030 Analysis Year
VH-92
Table VII-A-2
Sensitivity Analysis of Alternative Lag Structures for PM-related Premature Mortality
VH-94
Table VIII-1. Industries Containing Small Businesses
Potentially Affected by Today's Proposed Rule Vffi-3
Table Vffi-2. Costs for a 19,000 bbls gasoline/day
Refinery to Produce 30 ppm Gasoline Vffi-5
Table Vffi-3. Types and Number of Small Entities to
Which the Proposed Tier 2/Gasoline Sulfur Rule Would Apply Vffi-6
Table B-l. Sulfur Sensitivity: New Data B-3
Table B-2. Vehicles Tested After Short-Term vs. Long-Term
Exposure to Higher Sulfur Fuel B-5
Table B-3. Vehicle-by-Vehicle Short-Term vs. Long-Term Sulfur Sensitivity B-6
Table B-4. Percent Difference Between Short-Term vs. Long-Term Sulfur Sensitivity B-8
Table B-5. CRC Test Vehicles B-13
Table B-6. Sulfur Irreversibility: CRC Test Program (%) B-14
Table B-7. API Test Vehicles B-16
Table B-8. Sulfur Irreversibility: API Test Program
Low Mileage Catalysts, Short-Term Exposure to High Sulfur Fuel (%) B-l7
Table B-9. Sulfur Irreversibility: API Test Program
Low Mileage Catalysts, 1,000 Mile Exposure to High Sulfur Fuel (%) B-18
Table B-10. Sulfur Irreversibility: API with 100K Aged Catalysts Test Program (%) B-19
Table B-l 1. Sulfur Sensitivity: API Test Program
Low Mileage Catalysts, Short-Term Exposure to High Sulfur Fuel (g/mi) B-20
Table B-12. Sulfur Irreversibility: Ford Test Program (%) B-21
Table B-13. Sulfur Irreversibility: EPA Test Program,
Short-Term and Long-Term Exposure (%) B-22
Table B-14. Sulfur Irreversibility: ATL Test Program B-23
Table B-l5. Sulfur Irreversibility: DaimlerChrysler Test Program (%) B-24
Table B-16. Sulfur Irreversibility: EPA Test Program (%) B-25
Table B-l7. Sulfur Irreversibility: ATL Test Program (%) B-26
Table B-18. Pre-SFTP Sulfur Irreversibility: Summary of Relevant Test Programs (%) B-29
Table B-19. SFTP-Compliant Sulfur Irreversibility:
Summary of Relevant Test Programs (%) B-31
xxv
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table C-l. Energy Consumed by PADD 3 Refineries in 1994, Projected to 1997 C-2
Table C-2. Estimated Annual Energy and Hydrogen Demand of CDTECH
Desulfurization Units for Half of PADD 3 C-4
Table C-3. Estimated Annual Energy and Hydrogen Demand of OCTGAIN
Desulfurization Units for Half of PADD 3 C-5
Table C-4. Estimated Annual Energy and Hydrogen Demand of Black and Veatch
Desulfurization Units for Half of PADD 3 C-6
Table C-5. Estimated Annual Energy and Hydrogen Demand of Phillips
Desulfurization Units for Half of PADD 3 C-7
Table C- 6. Increase in Energy Consumed and Carbon Dioxide Emissions Due to Desulfurizing
Gasoline (1997 energy use and emissions) C-9
xxvi
-------
Table of Contents
List of Figures
Figure ffl.A-1. 47-State Light-Duty NOx Emissions Without Tier 2/Sulfur
(Annualized Summer Tons) III-6
Figure ffl.A-2. Breakdown of Total 2030 47 State NOx Inventory Without Tier 2 ffl-9
Figure IH.A-3. Breakdown of Total 2030 Atlanta NOx Inventory Without Tier 2 IH-9
Figure ffl.A-4. 47-State Light-Duty NOx Emissions With Tier 2/Sulfur
(Annualized Summer Tons) ffi-11
Figure ffl.A-5. Breakdown of Total 2030 47-State NOx Inventory With Tier 2/Sulfur .... ffl-14
Figure IH.A-6. Breakdown of Total 2030 Atlanta NOx Inventory With Tier 2/Sulfur IH-14
Figure ffl.A-7. 47-State Light-Duty VOC Emissions Without Tier 2/Sulfur
(Annualized Summer Tons) 111-17
Figure ffl.A-8. Breakdown of Total 2030 47-State VOC Inventory Without Tier 2/Sulfur
m-20
Figure ffl.A-9. Breakdown of Total 2030 Atlanta VOC Inventory Without Tier 2/Sulfur . . ffl-20
Figure IE. A-10. 47-State Light-Duty VOC Emissions With Tier 2/Sulfur
(Annualized Summer Tons) 111-22
Figure m.A.-ll. Breakdown of Total 2030 47-State VOC Inventory With Tier 2/Sulfur .. IH-25
Figure IH.A-12. Breakdown of Total 2030 Atlanta VOC Inventory With Tier 2/Sulfur ... IH-25
Figure ffl.A-13. 47-State Light-Duty SOx Emissions Without Sulfur Control
(Annual Tons) ffl-28
Figure ffl.A-14. 47-State Light-Duty SOx Emissions With Sulfur Control
(Annual Tons) ffl-30
Figure ILL A-15. 47-State Light-Duty Direct Exhaust PM10 Emissions
Without Tier 2/Sulfur - No Diesel Growth (Annual Tons) ffl-32
Figure IE. A-16. 47-State Light-Duty Direct Exhaust PM10 Emissions With Tier 2/Sulfur - No
Diesel Growth (Annual Tons) 111-35
Figure ffl.A-17: 47-State Light-Duty Direct Exhaust PM10 Without Tier 2/Sulfur - Increased
Diesel Sales (Annual Tons) in-37
Figure IHD-1. Example Plot of Target Fuel Benzene Versus
Baseline Fuel TOG under FTP Conditions ffl-96
Figure IV-1. Impact of Coating Architecture on HC and NOx Emissions IV-11
Figure IV-2: 50,000 mile equivalent NOx vs. NMHC levels for a number of hardware and engine
calibration configurations tested with a 1999 GM Chevrolet Silverado Pickup (5.3L V8)
originally certified to the LEV MDV-2 standard (0.4 g/mi NOx, 0.16 g/mi NMOG). FV-30
Figure IV-3: 50,000 mile equivalent NOx vs. NMHC emissions levels for a number of hardware
and engine calibration configurations tested with a 1999 Ford Expedition (5.4L V8)
originally certified to the LEV MDV-3 standard (0.6 g/mi NOx, 0.195 g/mi NMOG)
IV-31
Figure IV-4. Map of U.S. Petroleum Administrative Districts for Defense IV-41
Figure IV-5. Diagram of a Typical Complex Refinery IV-45
xxvn
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Figure V-l. Distribution of Progress Ratios V-25
Figure V-2. Total Annualized Costs of Tier 2 Vehicles and Low Sulfur Gasoline V-91
xxvin
-------
Chapter 1: Introduction
Chapter I: Introduction
We prepared this Regulatory Impact Analysis (RIA) for our final rule on Tier 2 Motor
Vehicle Emissions Standards and Gasoline Sulfur Control Requirements. The purpose of this
RIA is to present our 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.
This chapter summarizes the events that lead to today's action as well as the provisions
incorporated within it. 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 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 rule, including the impact of the 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 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 vehicle and fuel
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 standards to the expected costs.
• Chapter VIII concludes this RIA with a presentation of the Final Regulatory Flexibility
Analysis for the rule. This analysis evaluates the impacts of the Tier 2 and gasoline sulfur
standards on small businesses.
A. Background Information for Today's Final Rule
Through the Clean Air Act, Congress directed EPA to assess the air quality need,
1-1
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999 Chapter 1: Introduction
technical feasibility, and cost-effectiveness of more stringent motor vehicle emission
standards-emission standards more stringent than federal "Tier 1" standards. On July 31, 1998,
we submitted our Tier 2 Report to Congress, a formal report which contained the results of our
draft Tier 2 Study.1 In our study, we examined the appropriateness of requiring more stringent
emission standards for new passenger cars and light-duty trucks.
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. We also found
ample evidence that technologies would be available to meet more stringent Tier 2 standards.
The Tier 2 Study also 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.
In addition, on May 1, 1998, we released our Staff Paper on Gasoline Sulfur Issues which
presented our 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. 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, we stated in a Notice of Proposed Rulemaking (May 13, 1999, 64 FR 26004) that
new, more stringent emission standards are indeed needed, technologically feasible, and cost
effective. In June, 1999, we held four public hearings to obtain feedback on our proposal.
B. Overview of the Final Rule
Today's final rule, described below, incorporates changes to the proposed program based
upon updated analyses as well as comments heard at the public hearings and those submitted in
1 On April 28, 1998, We 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 we received.
1-2
-------
Chapter 1: Introduction
writing.
1. Vehicle Emission Standards
Today's action sets new federal emission standards ("Tier 2 standards") for passenger
cars, light trucks, and larger passenger vehicles. The program is designed to focus on reducing
the emissions most responsible for the ozone and particulate matter (PM) impact from these
vehicles — nitrogen oxides (NOx) and non-methane organic gases (NMOG), consisting primarily
of hydrocarbons (HC) and contributing to ambient volatile organic compounds (VOC). The
program will also, for the first time, apply the same set of federal standards to all passenger cars,
light trucks, and medium-duty passenger vehicles. Light trucks include "light light-duty trucks"
(or LLDTs), rated at less than 6000 pounds gross vehicle weight and "heavy light-duty trucks"
(or HLDTs), rated at more than 6000 pounds gross vehicle weight).2 "Medium-duty passenger
vehicles" (or MDPVs) form a new class of vehicles introduced by this rule that includes SUVs
and passenger vans rated at between 8,500 and 10,000 GVWR. The program thus ensures that
essentially all vehicles designed for passenger use in the future will be very clean vehicles.
The Tier 2 standards finalized today will 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 will phase
in beginning in 2004, with the standards to be fully phased in by 2007.3 For heavy LDTs and
MDPVs, the Tier 2 standards will 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 the primary Tier 2 standards will have to meet an interim average standard of 0.30
g/mi NOx, equivalent to the current NLEV standards for LDVs.4 During the period 2004-2008,
heavy LDTs and MDPVs not certified to the final Tier 2 standards will phase in to an interim
program with an average standard of 0.20 g/mi NOx, with those not covered by the phase-in
meeting a per-vehicle standard (i.e., an emissions "cap") of 0.60 g/mi NOx (for FtLDTs) and 0.09
g/mi NOx (for MDPVs). The average standards for NOx will allow manufacturers to comply
with the very stringent new standards in a flexible way, assuring that the average emissions of a
2 A vehicle's "Gross Vehicle Weight Rating," or GVWR, is the curb weight of the vehicle plus its
maximum recommended load of passengers and cargo.
3 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
(larger light trucks are not covered). For further comparison, the standards met by today's Tier 1 vehicles range
from 0.60 g/mi to 1.53 g/mi.
4 There 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. of this
preamble.
1-3
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999 Chapter 1: Introduction
company's production meet the target emission levels while allowing the manufacturer to choose
from several more- and less-stringent emission categories for certification.
We are also setting stringent particulate matter standards that will be especially important
if there is substantial future growth in the sales of diesel vehicles. With higher sales of diesel
cars and light trucks, these vehicles could easily contribute between one-half and two percent of
the PM10 concentration allowed by the NAAQS, with some possibility that the contribution
could be as high as five to 40 percent in some roadside situations with heavy traffic. These
increases would make attainment even more difficult for eight counties which we already predict
to need further emission reductions even without an increase in diesel sales, and would put at risk
another 18 counties which are now within 10 percent of a NAAQS violation. Thus, by including
a more stringent PM standard in the program finalized today, we help address environmental
concerns about the potential growth in the numbers of light-duty diesels on the road - even if
that growth is substantial. The new requirements also include more stringent hydrocarbon
controls (exhaust NMOG and evaporative emissions standards).
2. Gasoline Sulfur Standards
The other major part of today's action will significantly reduce average gasoline sulfur
levels nationwide. We expect these reductions could begin to phase in as early as 2000, with full
compliance for most refiners occurring by 2006. Refiners will generally install advanced refining
equipment to remove sulfur during the production of gasoline. Importers of gasoline will be
required to import and market only gasoline meeting the sulfur limits. Temporary, less stringent
standards will apply to refineries who produce fuel for use in the Geographic Phase-in Area5
through 2006 and a few small refiners through 2007.
This significant new control of gasoline sulfur content will have two important effects.
The lower sulfur levels will enable the much-improved emission control technology necessary to
meet the stringent vehicle standards of today's rule to operate effectively over the useful life of
the new vehicles. In addition, as soon as the lower sulfur gasoline is available, all gasoline
vehicles already on the road will have reduced emissions — from less degradation of their
catalytic converters and from fewer sulfur compounds in the exhaust.
Today's action will require that most refiners and importers meet a corporate average
gasoline sulfur standard of 120 ppm and a cap of 300 ppm beginning in 2004. By 2006, the cap
will be reduced to 80 ppm and most refineries must produce gasoline averaging no more than 30
ppm sulfur. The program builds upon the existing regulations covering gasoline composition 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. We intend for the
5Alaska, Colorado, Idaho, Montana, New Mexico, North Dakota, Utah, and Wyoming
1-4
-------
Chapter 1: Introduction
credit program to ease compliance uncertainties by providing refiners the flexibility to phase in
early controls in 2000-2003 and use credits gained in these years to delay some control until as
late as 2006. As finalized today, the program will achieve the needed environmental benefits
while providing substantial flexibility to refiners.
1-5
-------
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. Health and Welfare Effects of 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.1 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 precursors to 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.
1 Carbon monoxide also participates in the production of ozone, albeit at a much slower rate than most
VOC and NOx compounds.
n-i
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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
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.
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, chronic bronchitis and chronic
asthma.
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
from 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.
Evidence also exists of a possible relationship between daily increases in ozone levels
and increases in daily mortality levels. While the magnitude of this relationship is still too
uncertain to allow for direct quantification, the full body of evidence indicates a likely positive
n-2
-------
Chapter II: Health and Welfare Concerns
relationship between ozone exposure and premature mortality.
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 HI contains more information about air toxics.
Besides their role as an ozone precursor, NOx emissions produce a wide variety of health
and welfare effects.34 These problems are caused in part by emissions of nitrogen oxides 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.
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.5 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 the value placed on scenic views.
n-3
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Section n.D. further describes information about visibility impairment and regional haze.
B. Health and Welfare Effects of 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.2 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.
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
2Sulfuric acid is a paniculate, while nitric acid is a gas at ambient conditions.
n-4
-------
Chapter II: Health and Welfare Concerns
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.
Scientific studies have linked paniculate matter (alone or in combination with other air
pollutants) with a series of health effects.6 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 HI contains a discussion of the toxic health effects from paniculate matter in
diesel fuel exhaust.
Paniculate matter also causes a number of adverse effects on the environment. Fine
paniculate matter is the major cause of reduced visibility in parts of the U.S., including many of
our national parks and wilderness areas. (Section n.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
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
n-5
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 reduction 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.3 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 9 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.
3This value of the CO concentration decrease is measured by the composite average of the annual second
highest 8-hour concentration.
H-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.7
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
H-7
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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
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 1999, EPA promulgated a program to
address regional haze in the nation's most treasured national parks and wilderness areas (see 64
FR 35714, July 1, 1999).
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.8 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.9 The Commission identifies mobile source
pollutants of concern as VOC, NOX, and elemental and organic carbon.
H-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, 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
4. U.S.EPA, 1993, Air Quality Criteria for Oxides of Nitrogen, EP A/600/8-9 l/049aF.
5. 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.
6. U.S. EPA, 1996, Air Quality Criteria for Particulate Matter, EPA/600/P-95/001aF.
7. "National Air Quality and Emissions Trends Report, 1996", EPA Document Number
454/R-97-013.
8. Letter from Governor Michael Leavitt of Utah, on behalf of the Western Governors'
Association, to EPA Administrator Carol Browner, dated June 29, 1998.
9. "Report of the Grand Canyon Visibility Transport Commission to the United States
Environmental Protection Agency", June 1996.
H-9
-------
Chapter III: Environmental Impact
Chapter III: Environmental Impact
A. Inventory Impacts of Tier 2/Sulfur
Today's action will reduce NOx, VOC, paniculate, 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 action will 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 for LDVs and LDTs under 6000 pounds and Tier 1 for LDTs
above 6000 pounds on in-use fuel as currently specified; sulfur levels for Conventional Gasoline
are estimated at 330 ppm1, summertime Phase 2 RFG levels are estimated at 150 ppm (i.e.,
baseline case). "With Tier 2/Sulfur" refers to implementation of 120 ppm sulfur nationwide in
2004, 90 ppm in 2005 and 30 ppm in 2006 in conjunction with the phase-in of NOx, VOC, and
PM vehicles standards finalized under today's action (i.e., control case).2
For the proposal, separate emission inventories were used as the basis for the inventory
projections, air quality analysis and cost-benefit analysis. The emission inventory estimates were
based on updated estimates of on-highway mobile source emissions as planned for use in EPA's
MOBILE6 model, using a spreadsheet program known as the Tier 2 Model. However, the air
quality and cost-benefit analyses presented in the proposal were based on previous work which
relied on both MOBILES and the Tier 2 model. This discrepancy has been reconciled for the
final rule. Subsequent to the proposal, new emission inventories were generated for every county
in the nation in 1996, 2007 and 2030. These inventories reflect the on-highway mobile source
emission updates contained in the NPRM version of the Tier 2 Model, and were used as the basis
'For NOx and VOC, the modeling performed in support of our air quality and cost-benefit analyses was
based on a sulfur baseline of 330 ppm for conventional gasoline. This estimate has since been updated to 300 ppm,
reflecting recent in-use fuel data and our expectation that refiners will shift sulfur to summertime conventional
gasoline in order to meet the Phase 2 RFG NOx performance specification. This change has been incorporated into
our updated Tier 2 Model, as detailed in a memorandum from John Koupal to Docket A-97-10 entitled
"Development of Light-Duty Emission Inventory Estimates in the Final Rulemaking for Tier 2 and Sulfur
Standards". Our PM and SOx estimates are based on PARTS, which uses a sulfur level of 339 ppm for
conventional gasoline.
2These sulfur levels would occur under our Sulfur Averaging, Banking and Trading (ABT) program, for
which reduced sulfur levels could occur as early as 2000. We did not include sulfur reductions prior to 2004 in
our air quality analysis inventory estimates, but early sulfur reductions were accounted for in our updated Tier 2
Model.
m-i
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
for the inventory projections, air quality analysis and cost-benefit analysis (an updated version of
the Tier 2 Model, which incorporates data and comments received subsequent to the proposal, is
discussed below). Specifically, the modeling 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 modeling 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.
The final rule modeling also incorporates planned changes in emission rates for heavy-duty
gasoline and diesel vehicles, as well as the effects of heavy-duty NOx defeat device.
The inventories generated from this round of modeling are presented as our "official"
inventory estimates in this chapter; since they formed the basis of our final rule air quality
analysis this work is referred to as the "air quality analysis" modeling, to distinguish from
subsequent inventory modeling performed using updated inputs. These estimates of on-highway
emissions were generated in several steps: First, MOBILES and PARTS where run for every
county in the country. Next, multiplicative adjustment factors were applied to the output of
MOBILES to account for changes planned for MOBILE6 for light-duty and heavy-duty vehicles;
for 2007 and 2030, multiplicative adjustment factors were also developed which incorporated
the benefits of the Tier 2/Sulfur program, for all pollutants. Finally, excess NOx emissions due
to air conditioning usage were applied as a function of ambient temperature.1 This approach
enabled a significant improvement over what was done in the proposal, in that the final rule
inventories are estimated on a county-by-county basis using locally-specific modeling inputs and
assumptions where available, "corrected" (for NOx and VOC) to our best estimate of MOBILE6
(at the time of the proposal) with and without Tier 2/Sulfur control. Thus the 47-state numbers
presented here are not based just on national default inputs as was done for the proposal, but
instead are based on the sum of the inventories for every county in the 47 states, including local
average temperatures, fleet characteristics, I/M programs, fuel properties, and roadway
type/speed distributions. Likewise, the city-specific inventories presented here, which are based
on the same county-specific inventories, are substantially more accurate than the ones presented
in the proposal, which are based on national default modeling input.
In the proposal, we used a combination of methods to project future highway vehicle
miles traveled (VMT). The proposal approach used the 1997 National Emissions Trends (NET)
Report VMT up to 2010, a compounded growth rate of 2.1 percent from 2010 to 2015 and a
simple linear growth rate of 2.1 percent after 2015. For the final rule, we have chosen to project
future growth using a linear extrapolation of the NET projections. This is a simpler, more
consistent approach, which results in lower estimates of future VMT (a simple linear growth rate
of 1.7 percent from 2007 to 2030). Total light-duty VMT estimates were split into light-duty
vehicle (LDV), light-duty trucks below 6,000 pounds (LDT1/2) and light-duty trucks above
6,000 pounds (LDT3/4) using the methodology developed for the proposal,2 accounting for the
recent growth trends in LDT sales.
m-2
-------
Chapter III: Environmental Impact
For the final rule, we have also updated the emission inventories for stationary, area, and
nonroad sources. The development of inventories for all sources are provided in a separate
document available in the Tier 2 docket.3
This section focuses on projections of the emissions inventory with and without today's
action for the United States excluding California, Alaska, and Hawaii,3 derived from the final
rule air quality analysis. Estimates from the final rule air quality analysis were also used to
generate the relative contribution of light-duty vehicles and trucks to the total NOx and VOC
inventories nationwide, as well as four urban areas: New York, Chicago, Atlanta and Charlotte.
Comprehensive inventories (47-state and four city) are presented in Appendix A with and
without Tier 2/sulfur control for 1996, 2007 and 2030, the three years for which the inventories
were generated. For VOC and NOx, the nationwide inventories are presented as "annualized
summer tons," meaning that inventory results for a typical July day were multiplied by 365 days.
The purpose of this is solely to present a consistent comparison of emission trends and reductions
based on summer conditions; the actual air quality and cost-benefit analyses relied on the
seasonal inventories generated in the air quality analysis.4 Nationwide SOx and PM results
presented in this section are true annual estimates.
For purposes of the analysis presented in this section, we also needed to estimate
inventories in other years between 2007 and 2030. Estimates for light-duty vehicles and trucks in
these intermediate years were derived from the NPRM version of the Tier 2 Model, adjusted to
reflect differences in this model and the air quality analysis results.4 To estimate stationary and
area source inventories in other years, we did a simple linear interpolation of the 2007 and 2030
estimates. For nonroad and on-highway inventories, we know that the emissions inventory is
non-linear between those years due to fleet turnover. To estimate intervening years, we first
calculated the ratio of the air quality analysis nonroad and on-highway inventories in 2007 and
2030 to the inventories generated for the proposal for these sectors for those same years. Then
we interpolated between the 2007 and 2030 ratios to estimate the ratio that we would expect in
intervening years. We then multiplied those interpolated ratios by the NPRM inventories for
those years. The result is an estimate of what the inventories would have been if we had done
3The 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.
4 In 2007, the 47-state annual VOC emissions reductions from the Tier 2/Sulfur program are
approximately 13% larger than the annualized summer VOC emissions reductions (the smaller evaporative VOC
emissions reductions in non-summer conditions are more than offset by higher exhaust VOC emissions reductions
during colder weather), while the annual 47-state NOx emissions reductions are 6-7% smaller than the annualized
NOx emissions reductions.
m-3
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
complete county-by-county inventories for them.
As mentioned, for the final rule air quality analysis the modeling assumptions and inputs
for light-duty vehicles and trucks were identical to those included in the NPRM version of the
Tier 2 Model. Subsequent to the air quality analysis modeling, our Tier 2 Model was updated to
reflect several new inputs stemming from a) our response to Tier 2 comments, b) new sulfur
sensitivity data and c) alignment with methodologies planned for use in MOBILE6, as well as
changes to the sulfur control program.5 Results from the updated Tier 2 Model are presented
side-by-side with the air quality analysis inventory modeling results for VOC and NOx. The
major updates to the model are summarized below:
1) Tier 1 and later NOx emission rates were updated to reflect a significantly larger
sample of vehicles certified to the 0.4 gram/mile NOx standard, in response to
comments.
2) Sulfur effects for LDV and LDT LEVs were increased significantly, in response to
new data showing that the effect of sulfur on emissions is much larger when a
vehicle operates on high sulfur fuel for a few thousand miles.
3) The model incorporates the effects of sulfur irreversibility, which results when
vehicles sustain permanent catalyst degradation from exposure to sulfur levels
higher than what they typically operate on. Inventory estimates are presented for
irreversibility levels consistent with those discussed in Appendix B.
4) The model incorporated the effects of representative in-use activity data planned
for use in MOBILE6, including vehicle speed, roadway type, and trip activity.
5) The Tier 2/Sulfur control case reflects the sulfur program contained in today's
action, as well as the effects of Averaging, Banking and Trading (ABT), interim
provisions for small refiners, and the geographic phase-in of low sulfur fuel.
With these updates, the Tier 2 Model now includes many of the key exhaust elements
planned for MOBILE6, and is the most up-to-date tool available for assessing trends in
nationwide light-duty exhaust emissions and the emission reductions gained from the Tier
2/Sulfur program. Overall, the updated model indicates that NOx and exhaust VOC emissions
without Tier 2/Sulfur control will be substantially higher than originally projected either in the
proposal or by the air quality analysis modeling, particularly for NOx. Although the inventory
estimates, air quality results and economic benefits assessment presented in this document show
conclusively the need for and benefit of today's action, we believe based on the updated Tier 2
Model that the estimates of emissions reductions underlying these analyses are in fact very
conservative.
1. NOx
a. Light-Duty NOx Trends Without Tier 2/Sulfur
m-4
-------
Chapter III: Environmental Impact
Total NOx emissions produced annually in the 47 states by cars and trucks without Tier
2/Sulfur controls are shown in Table ni.A-1 and Figure in.A-1, 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 mentioned, the air quality analysis inventory results are based on annualized
summer day results in 2007 and 2030, with intermediate years developed based on the NPRM
version of the Tier 2 Model. As shown, the estimates derived from the air quality analysis
modeling show total light-duty emissions declining from approximately 3.6 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 3.7 million tons by 2030,
such that the gains from the Tier 1, NLEV and SFTP control programs are more than offset by
VMT growth.
The estimates derived from the updated Tier 2 Model suggest a much more dire situation.
While the updated modeling estimates are similar to the air quality analysis estimates in 2000,
emissions are projected to rise steadily from this point; by 2030, emission are projected to
increase 50 percent from the 2000 levels.
m-5
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table III.A-1. 47-State Light Duty NOx Emissions Without Tier 2/Sulfur
(Annualized Summer Tons)
Year
2000
2004
2007
2010
2015
2020
2030
Light-Duty
Emissions: Air
Quality Analysis
Modeling
3,591,547
3,362,528
3,095,698
2,962,093
2,968,707
3,160,155
3,704,747
Contribution by Vehicle Class
LDV
52.1%
43.5%
37.3%
33.0%
28.6%
26.9%
27.1%
LDT1/2
30.0%
37.1%
41.0%
42.7%
44.3%
45.2%
45.5%
LDT3/4
17.9%
19.4%
21.7%
24.3%
27.1%
27.8%
27.4%
Light-Duty
Emissions: Updated
Tier 2 Model
3,548,883
3,612,395
3,681,990
3,817,070
4,116,074
4,502,761
5,323,860
6,000,000 -f
5,000,000
4,000,000 -
3,000,000
2,000,000
1,000,000
UPDATED T ER 2 MODEL
2000 2005
2010
2015 2020 2025 2030
LDV DLDT1/2
LDT3/4
Figure III.A-1. 47-State Light-Duty NOx Emissions Without Tier 2/Sulfur
(Annualized Summer Tons)
m-6
-------
Chapter III: Environmental Impact
The impact of steady truck growth on overall light-duty NOx emissions is clearly
demonstrated in the preceding figure. In 2000, we project that trucks will produce nearly 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
in.A-1, the decrease in overall light-duty emission levels estimated in the air quality analysis
modeling 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 updated Tier 2
Model shows that emission gains from these control programs are completely offset by high
sulfur sensitivity for LEVs, VMT growth and the increased penetration of light-duty trucks.
Figures ni.A-2 and in.A-3 show our projections of the contribution of light-duty vehicles
and trucks to the total NOx inventory (i.e., NOx emissions from all sources, including stationary,
area, nonroad) in the 47 states and in Atlanta based on the air quality analysis modeling for a
typical ozone season day. Table HI. A-2 shows this same contribution across all four cities from
2007 through 2030 based on the air quality analysis modeling, and for the 47 states based on the
air quality analysis modeling and the updated Tier 2 Model. Across the 47 states, the air quality
analysis modeling estimates that cars and trucks produce 16 percent of total emissions in 2007,
growing to nearly one-fifth of total NOx emissions by 2030. The updated modeling projects a
light-duty contribution of 19 percent in 2007, growing to one-quarter by 2030. In all cases, the
relative contribution of cars and trucks to total NOx emissions is projected to grow steadily.
Light-duty NOx contribution in urban areas is generally higher than in 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. Atlanta provides the most striking example of this; the air quality analysis
modeling projects that 34 percent of all NOx emissions will be produced by cars and trucks in
2007, growing to 41 percent by 2030.5 The light-duty contribution in New York and Charlotte
5The air quality analysis modeling reflects a baseline sulfur level of 330 ppm for Atlanta, our estimate of
national average conventional gasoline at the time of the proposal. Atlanta has recently implemented a summertime
lower sulfur fuel program requiring 150 ppm fuel in 1999 and 30 ppm fuel in 2003; Georgia has submitted a SIP
revision including this program and requested EPA's approval of a waiver of federal preemption and the SIP
revision. Approval of the waiver and SIP revision is still pending.
m-7
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
are higher than the national estimates, while the contribution in Chicago is slightly less than the
47-state estimate. Based on the national results, our updates to the modeling would result in even
larger contributions from cars and trucks in these cities, and in urban areas nationwide.
m-8
-------
Chapter III: Environmental Impact
Other On-Highway
7%
Nonroad
29%
Light-Duty Vehicles
and Trucks
19%
Stationary and Area
45%
Figure III.A-2. Breakdown of Total 2030 47 State NOx Inventory Without Tier 2
Other On-Highway
12%
Nonroad
31%
Light-Duty Vehicles
and Trucks
41%
Stationary and Area
16%
Figure III.A-3. Breakdown of Total 2GJ09^tlanta NOx Inventory Without Tier 2
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table III.A-2. Light-Duty Contribution to Total NOx Inventory Without Tier 2/Sulfur
(Typical Ozone Season Day)
Year
2007
2010
2015
2020
2030
Air Quality Analysis Modeling
47 -State
16%
16%
17%
17%
19%
New York
18%
-
-
-
22%
Chicago
14%
-
-
-
16%
Atlanta
34%
-
-
-
41%
Charlotte
24%
-
-
-
27%
Updated Model
47 -State
19%
20%
22%
23%
25%
b.
NOx Reductions Due To Tier 2/Sulfur
Today's action will provide substantial reductions in NOx emissions from cars and
trucks. The implementation of low sulfur fuel will 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 2028 the projected upward trend in light-duty NOx emissions that
would occur with no control. Table ni.A-3 contains annual tons of NOx we project will be
reduced by today's action, encompassing benefits of low sulfur fuel and the introduction of Tier
2 light-duty vehicle and light-duty truck standards. Figure HI. A-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.
m-io
-------
Chapter III: Environmental Impact
Table III.A-3. 47-State Light-Duty NOx Reductions Due To Tier 2/Sulfur
(Annualized Summer Tons)
Year
2004
2007
2010
2015
2020
2030
Air Quality Analysis Modeling
Emissions
Reduced
338,231
856,471
1,235,882
1,816,767
2,220,210
2,795,551
Percent Reduction in
Baseline Inventory
Light-Duty
10%
28%
42%
61%
70%
75%
All
Sources*
-
5%
7%
10%
12%
15%
Updated Tier 2 Model
Emissions
Reduced
326,556
956,512
1,554,442
2,527,309
3,205,571
4,049,687
Percent Reduction in
Baseline Inventory
Light-Duty
9%
26%
41%
61%
71%
76%
All
Sources*
-
5%
8%
13%
16%
20%
1 Includes emission reductions from Heavy-Duty Gasoline Vehicles due to sulfur control
6,000,000
5,000,000
4,000,000
3,000,000
2,000,000
1,000,000
WITHOUT TIER 2/SULFUR CONTROLS
WITH TIER 2/SULFUR CONTROLS
2000 2005 2010 2015 2020 2025
2030
•Air Quality Analysis Modeling
- - - - Updated Tier 2 Model
Figure III.A-4. 47-State Light-Duty NOx Emissions With Tier 2/Sulfur
(AnnualizeeFSnmmer Tons)
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
The projected reductions in 2004 are based on 120 ppm sulfur fuel, reflecting our final
Tier 2 sulfur program. Our modeling shows an immediate benefit of over 300,000 tons in 2004,
a 10 percent drop in uncontrolled light-duty emissions; this is the equivalent of emissions
produced by 19 million pre-Tier 2 cars and trucks.8'6 In the early years of sulfur control, nearly
all of the benefits would be due to reduced emissions from Tier 0, Tier 1 and NLEV vehicles.
Although not shown, emission reductions due to sulfur control could be realized as early as 2000
under the sulfur ABT program.
After 2004, emissions 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 70 percent reduction in 2020 light-
duty emissions without Tier 2/Sulfur, equivalent to the emissions from nearly 164 million pre-
Tier 2 cars and trucks. Total U.S. NOx emissions from all human sources would be reduced by
12 to 16 percent.
We project that light-duty emissions will 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 will consist of Tier 2 vehicles. The benefit of 2.8 million tons projected by the air
quality analysis modeling represents a three-quarters reduction in 2030 light-duty emissions
without Tier 2/Sulfur, equivalent to the emissions from over 200 million pre-Tier 2 cars and
trucks. These emission reductions are projected to be 15 percent of total NOx emissions in that
year in the absence of today's action. The benefits projected by the updated modeling are even
more substantial: emission reductions of over four million tons are projected, representing a 76
percent reduction of baseline light-duty emissions and a one-fifth reduction in NOx from all
sources.
The estimated percentage reductions in total inventory presented in Table HI. A-3 include
benefits that will be realized from 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 when 30 ppm fuel is in place. In addition, reductions from "Medium Duty Passenger
Vehicles" (e.g. passenger vehicles above 8500 pounds included as part of Tier 2 vehicle program)
are estimated to be approximately 37,000 tons in 2030.
Concurrently, we project that the light-duty contribution to total NOx emissions will drop
significantly. Figures IHA-5 and ni.A-6 show our 2030 projections of this contribution in the
47 states and in Atlanta with Tier 2/Sulfur control. Table IHA-4 shows this same contribution
across the 47 states from 2007 through 2030, and in the four cities in 2007 and 2030. In 2030,
the air quality analysis modeling projects that the light-duty contribution will drop to five percent
8i.e., vehicles that would be on the road in the absence of Tier 2/Sulfur control.
m-12
-------
Chapter III: Environmental Impact
nationally, from 19 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. Based on the air
quality analysis modeling we project that with Tier 2/Sulfur control, car and truck emissions
would contribute six percent of total emissions in New York (down from 22 percent), four
percent in Chicago (down from 16 percent), six percent in Charlotte (down from 27 percent), and
11 percent in Atlanta (down from 41 percent) in 2030.
m-13
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Other On-Highway
Light-Duty Vehicles
and Trucks
5%
Nonroad
34%
Stationary and Area
53%
Figure III.A-5. Breakdown of Total 2030 47-State NOx Inventory With Tier 2/Sulfur
m-14
-------
Chapter III: Environmental Impact
Other On-Highway
18%
Light-Duty Vehicles
and Trucks
11%
Stationary and Area
24%
Nonroad
47%
Figure III.A-6. Breakdown of Total 2030 Atlanta NOx Inventory With Tier 2/Sulfur
m-15
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table III.A-4. Light-Duty Contribution to Total NOx Inventory With Tier 2/Sulfur
(Typical Ozone Season Day)
Year
2007
2010
2015
2020
2030
Air Quality Analysis Modeling
47-State
12%
10%
7%
6%
5%
New York
15%
-
-
-
6%
Chicago
11%
-
-
-
4%
Atlanta
26%
-
-
-
11%
Charlotte
18%
-
-
-
6%
Updated Model
47-State
14%
13%
10%
8%
8%
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 HI.A-5 and Figure IHA-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 2.8 million tons to 1.7 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 1.8 million tons by
2020 and 2.1 million tons by 2030. As shown in Figure in.A-7, our updated modeling projects
much higher emissions, and an upturn in total light-duty VOC beginning in 2015.
HI-16
-------
Chapter III: Environmental Impact
Table III.A-5. 47-State Light-Duty VOC Emissions Without Tier 2/Sulfur
(Annualized Summer Tons)
Year
2000
2004
2007
2010
2015
2020
2030
Light-Duty
Emissions:
Air Quality
Analysis
Modeling
2,751,002
2,323,874
2,026,945
1,828,506
1,733,981
1,788,057
2,108,765
Contribution by Emission Source / Vehicle Class
Evaporative
(All
LDV/LDT)
48%
50%
53%
55%
57%
57%
57%
Exhaust
LDV
23%
19%
16%
14%
11%
10%
9%
LDT1/2
15%
18%
17%
17%
17%
18%
18%
LDT3/4
14%
13%
13%
14%
16%
16%
15%
Light-Duty
Emissions:
Updated Tier
2 Model
3,202,293
2,794,249
2,544,842
2,356,512
2,291,030
2,389,757
2,845,573
m-17
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
4,000,000
3,000,000
2,000,000 -;
1,000,000
UPDATED TIER 2 MODEL
2005
2010
2015
2020
2025
2030
Evap (All LD) DLDV Exhaust D LDT1/2 Exhaust D LDT3/4 Exhaust
Figure III.A-7. 47-State Light-Duty VOC Emissions Without Tier 2/Sulfur
(Annualized Summer Tons)
m-18
-------
Chapter III: Environmental Impact
Evaporative emissions are projected to be about 50 percent of the light-duty inventory in
2000, with this percent contribution rising steadily through 2030. Exhaust emissions from trucks
also play an increasingly significant role in shaping the overall VOC trend. In 2000, we project
that trucks will produce approximately 56 percent of exhaust VOC emissions; by 2015, trucks
account for 75 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.
Figures HI.A-8 and ILL A-9 show our projections of the contribution of light-duty vehicles
and trucks to the total anthropogenic (i.e., human-caused) 2030 VOC inventory in the 47 states
and in Atlanta (on an annualized summer basis) that were used in the air quality and economic
benefits analysis. Table HI. A-6 shows this same contribution across the 47 states and all four
cities from 2007 through 2030. Nationally, cars and trucks produce 13 percent of total VOC
emissions in 2007; this percentage declines in subsequent years before stabilizing at 11 percent
by 2015. Relative to the national average, the light-duty contributions are lower in New York
and Chicago and higher in Atlanta and Charlotte. For the latter two cities, we project that cars
and trucks will contribute 17 and 15 percent of all VOC emissions in 2007.
HI-19
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Nonroad
12%
Other On-Highway
3%
Light-Duty Vehicles
and Trucks
11%
Stationary and Area
74%
Figure III.A-8. Breakdown of Total 2030 47-State VOC Inventory Without Tier 2/Sulfur
Nonroad
12%
Other On-Highway
4%
Light-Duty Vehicles
and Trucks
15%
Stationary and Area
69%
Figure III.A-9. Breakdown of Total 2030^jtk
-------
Chapter III: Environmental Impact
Table III.A-6. Light-Duty Contribution to Total VOC Inventory Without Tier 2/Sulfur
(Typical Ozone Season Day)
Year
2007
2010
2015
2020
2030
Air Quality Analysis Modeling
47 State
13%
12%
11%
11%
11%
New York
6%
-
-
-
6%
Chicago
6%
-
-
-
5%
Atlanta
17%
-
-
-
15%
Charlotte
15%
-
-
-
12%
Updated Modeling
47 State
15%
14%
14%
14%
15%
b.
VOC Reductions Due To Tier 2/Sulfur
Table IHA-7 contains annual nationwide tons of VOC we project will be reduced due to
today's action, 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
IE.A-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.
HI-21
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table III.A-7. 47-State Light-Duty VOC Reductions Due to Tier 2/Sulfur
(Annualized Summer Tons)
Year
2004
2007
2010
2015
2020
2030
Air Quality Analysis Modeling
Emissions
Reduced
85,688
143,507
178,886
244,080
305,470
400,968
Percent Reduction in
Baseline Inventory
Light-
Duty
4%
7%
10%
14%
17%
19%
All
Sources*
-
0.9%
1.1%
1.5%
1.8%
2.2%
Updated Tier 2 Model
Emissions
Reduced
127,957
262,174
349,126
491,336
615,239
806,343
Percent Reduction in
Baseline Inventory
Light-
Duty
5%
10%
15%
21%
26%
28%
All
Sources*
-
1.6%
2.1%
2.9%
3.5%
4.2%
* Includes emission reductions from Heavy-Duty Gasoline Vehicles due to sulfur control
m-22
-------
Chapter III: Environmental Impact
4,000,000
3,000,000
2,000,000
1,000,000 -
WITHOUT TIER 2/SULFUR CONTROLS
WITH TIER 2/SULFUR CONTROLS
2000 2005 2010 2015 2020 2025 2030
• Air Quality Analysis Modeling Updated Tier2 Model
Figure IILA-10. 47-State Light-Duty VOC Emissions With Tier 2/Sulfur
(Annualized Summer Tons)
Although not shown, emission reductions due to sulfur control could be realized as early
as 2000 under the sulfur ABT program. In 2004, we project that the implementation of 120 ppm
fuel will 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 seven million
pre-Tier 2 cars and trucks. After 2004, further sulfur reductions and the introduction of LDT2s,
LDT3s, and LDT4s complying with the Tier 2 NMOG standard and operating on low sulfur fuel
reduce emission further. By 2030, the air quality analysis modeling projects that baseline light-
duty VOC emissions are reduced 19 percent due to Tier 2/Sulfur control, the equivalent of
emissions from 51 million pre-Tier 2 cars and trucks. This represents a 2.2 percent reduction of
the total anthropogenic VOC inventory. Our more recent modeling suggests that reductions will
in fact be significantly larger than the air quality analysis results; by 2030, light-duty VOC
emissions are reduced nearly 30 percent, and total anthropogenic VOC emissions reduced 4.2
percent. Tier 2/Sulfur control is projected to delay the upturn in light-duty VOC emissions by
five years.
It should be noted that both the air quality analysis modeling and updated Tier 2 Model
assume a Tier 2 fleet average NMOG level of 0.09 grams/mile, which is the highest fleet average
possible under the certification bin system. A more likely fleet average is in the range of 0.07 to
HI-23
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
0.08 grams/mile; the actual VOC emission reductions realized from today's action will thus be
even larger than presented here.
We project that heavy-duty gasoline vehicles would decrease emissions by approximately
4,500 tons per year in 2007 due to sulfur control; these reductions are included in the estimates of
mobile source and all source percent reductions contained in Table IHA-7. In addition,
reductions from "Medium Duty Passenger Vehicles" (e.g. passenger vehicles above 8500 pounds
included as part of Tier 2 vehicle program) are estimated to be at least 9,500 tons in 2030.
Figures ni.A-11 and IHA-12 show the contribution of light-duty cars and trucks to total
2030 VOC inventory in the 47 states and in Atlanta with Tier 2/Sulfur control. Table m.A-8
shows this same contribution across the 47 states in 2007, 2010, 2015 and 2030, and in the four
cities in 2007 and 2030. By 2030, we project that the light-duty contribution will drop to 9
percent nationally, from 11 percent without Tier 2/Sulfur control. This trend will 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 will contribute four
percent of total emissions in New York and Chicago in 2030 (down from six and five percent,
respectively), eight percent in Charlotte (down from 12 percent), and 11 percent in Atlanta (down
from 15 percent).
HI-24
-------
Chapter III: Environmental Impact
Nonroad
12%
Other On-Highway
3%
Light-Duty Vehicles
and Trucks
9%
Stationary and Area
76%
Figure IILA.-ll. Breakdown of Total 2030 47-State VOC Inventory With Tier 2/Sulfur
m-25
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Nonroad
12%
Other On-Highway
5%
Light-Duty Vehicles
and Trucks
11%
Stationary and Area
72%
Figure IILA-12. Breakdown of Total 2030 Atlanta VOC Inventory With Tier 2/Sulfur
m-26
-------
Chapter III: Environmental Impact
Table III.A-8. Light-Duty Contribution to Total VOC Inventory With Tier 2/Sulfur
(Typical Ozone Season Day)
Year
2007
2010
2015
2020
2030
Air Quality Analysis Modeling
47 State
12%
11%
9%
9%
9%
New York
6%
-
-
-
4%
Chicago
5%
-
-
-
4%
Atlanta
15%
-
-
-
11%
Charlotte
14%
-
-
-
8%
Updated Modeling
47 State
14%
13%
11%
11%
11%
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 ffl.A-9 and Figure HI.A-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 196,000 tons of SOx; by 2030, this level is projected to be 314,000
tons, an increase of 60 percent. The absolute emission estimates presented in Table IH.A.-9 are
based on a 330 ppm average sulfur level for conventional gasoline; using our updated estimate
for conventional gasoline of 300 ppm conventional gasoline, the values in this table would be
reduced by approximately eight percent.
HI-27
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table III.A-9. 47-State SOx Emissions Without Sulfur Control (Annual Tons)
Year
2000
2004
2007
2010
2015
2020
2030
Emissions
From All
Sources
-
-
18,052,276
17,949,631
17,792,528
17,607,934
17,242,341
Light-Duty
Emissions
196,334
206,258
216,626
230,781
253,109
274,016
313,998
Light-Duty
Contribution
to All
Sources
-
-
1.2%
1.3%
1.4%
1.6%
1.8%
Contribution by Vehicle Class
LDV
49%
42%
37%
33%
29%
27%
27%
LDT1/2
34%
42%
47%
50%
53%
54%
55%
LDT3/4
16%
16%
16%
17%
18%
18%
18%
400,000
300,000
200,000
100,000
2005
2010
2015
2020
2025
LDV
LDT1/2 DLDT3/4
2030
Figure IILA-13. 47-State Light-Duty SOx Emissions Without Sulfur Control
(Annual Tons)
m-28
-------
Chapter III: Environmental Impact
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
steadily. In 2000, trucks account for roughly half of light-duty SOx emissions, growing to over
70 percent by 2015.
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 ni.A-10 contains
annual nationwide tons of gaseous SOx we project will be reduced from light-duty vehicles and
trucks due to sulfur control. Figure ni.A-14 shows SOx emissions after sulfur control, broken
down by LDV, LOT 1/2 and LDT3/4.
Table III.A-10. 47-State Light-Duty SOx Reductions Due To Sulfur Control (Annual Tons)
Year
2004
2007
2010
2015
2020
2030
Light-Duty
Emissions
Without Sulfur
Control
206,258
216,626
230,781
253,109
274,016
313,998
Light-Duty
Emissions
With Sulfur
Control
82,408
22,847
24,302
26,652
28,837
32,982
Emissions
Reduced
123,850
193,779
206,479
226,457
245,179
281,016
Percent Reduction in
Baseline Inventory
Light-Duty
60%
89%
89%
89%
89%
89%
All Sources*
-
1.2%
1.4%
1.5%
1.6%
1.9%
* Includes reductions from Heavy-Duty Gasoline Vehicles, Motorcycles and Nonroad Sources
HI-29
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
400,000
300,000
200,000
100,000
WITHOUT SULFUR CONTROL
2005
2010
2015
2020
2025
2030
LDV DLDT1/2
LDT3/4
Figure III.A-14. 47-State Light-Duty SOx Emissions With Sulfur Control
(Annual Tons)
As shown, a significant reduction in light-duty SOx emissions would be realized
immediately with sulfur control. The reductions presented above presume 120 ppm in 2004 and
90 ppm in 2005 under the sulfur ABT program; reductions could be realized as early as 2000
under this program. We project that nearly 90 percent of light-duty SOx emissions will be
reduced when 30 ppm fuel is introduced in 2006. This relative reduction will remain constant
beyond 2006 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 16,000 tons by 2030. In addition, emissions from all gasoline-powered
nonroad equipment would be reduced due to sulfur control. 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 IHA-IO.
4.
Particulate Matter
m-30
-------
Chapter III: Environmental Impact
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 the
Partnership for a New Generation of Vehicles (PNGV). Thus, we assessed PM emissions with
and without Tier 2/Sulfur control under two sales scenarios: a "no growth" scenario, for which
current diesel sales trends were assumed to continue, and an "increased growth" scenario, based
on projections developed by A.D. Little, Inc. known as the "Most Likely" scenario. These
projections estimate that diesels will grow to nine percent of light-duty vehicle sales and 24
percent of light-duty truck sales by 2015.7 The inventory results presented in this section are for
direct exhaust PM10 emissions, 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 emissions are presented separately in Appendix A.
a. "No Growth" Diesel Sales Scenario
/'. Light-Duty Direct Exhaust PM10 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 from the light-duty vehicle fleet 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 in.A-11 and Figure ni.A-15. In 2000, we project
that approximately 36,000 tons will be emitted annually by light-duty cars and trucks. This level
is projected to exceed 48,000 tons in 2020 and reach nearly 56,000 tons in 2030. The absolute
emission estimates presented in Table IH.A.-l 1 are based on a 330 ppm average sulfur level for
conventional gasoline; using our updated estimate for conventional gasoline of 300 ppm
conventional gasoline, the values in this table would be reduced by approximately six percent.
ni-3i
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table IILA-11. 47 State Light-Duty Direct Exhaust PM10 Emissions Without Tier 2/Sulfur
No Growth in Diesel Sales
(Annual Tons)
Year
2000
2004
2007
2010
2015
2020
2030
Emissions
From All
Sources*
-
-
2,907,819
2,945,927
3,058,635
3,168,482
3,431,450
Light-
Duty
Exhaust
Emissions
38,729
38,551
40,365
42,385
45,887
49,577
56,861
Light-Duty
Contribution
to All
Sources
-
-
1.4%
1.4%
1.5%
1.6%
1.7%
Contribution by Fuel Type / Vehicle Class
Diesel
LDV/LDT
7%
4%
2%
2%
2%
2%
2%
Gas
LDV
47%
45%
40%
36%
32%
31%
30%
Gas
LDT1/2
28%
37%
43%
47%
50%
51%
51%
Gas
LDT3/4
19%
14%
14%
15%
16%
16%
17%
* Excludes natural and miscellaneous sources (e.g., fugitive dust), but includes indirect sources such as tire and
brake wear.
m-32
-------
Chapter III: Environmental Impact
2000
2005
2010
2015
2020
2025
2030
Diesel
LDVGas pLDT1/2Gas
DLDT3/4 Gas
Figure IILA-15. 47-State Light-Duty Direct Exhaust PM10 Emissions
Without Tier 2/Sulfur - No Diesel Growth (Annual Tons)
m-33
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 PM10 inventory to grow to 70 percent by 2030.
/'/'. Direct Exhaust PM10 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 IHA-12 contains annual nationwide tons of direct exhaust PM10 we project will be reduced
from light-duty vehicles and trucks due to Tier 2/Sulfur control. Figure ni.A-16 shows PM10
emissions after Tier 2/Sulfur control broken down by diesel (all light-duty cars and trucks) and
gasoline LDV, LDT1/2 and LDT3/4.
Table IILA-12. 47-State Light-Duty Direct Exhaust PM10 Reductions Due To Tier 2/Sulfur
No Growth in Diesel Sales
(Annual Tons)
Year
2004
2007
2010
2015
2020
2030
Light-Duty
Emissions Without
Tier 2/Sulfur
38,551
40,365
42,385
45,887
49,577
56,861
Light-Duty
Emissions With
Tier 2/Sulfur
24,424
16,938
17,254
17,937
18,891
20,857
Emissions
Reduced
14,127
23,427
25,131
27,950
30,686
36,004
Percent Reduction in
Baseline Inventory
Light-Duty
37%
58%
59%
61%
62%
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.
m-34
-------
Chapter III: Environmental Impact
60,000
50,000
40,000
30,000
20,000
10,000
WITHOUT TIER 2/SULFUR CONTR
2000
2005
2010
2015
1
2020
2025
2030
Diesel
LDV Gas
DLDT1/2 Gas
DLDT3/4 Gas
Figure IILA-16. 47-State Light-Duty Direct Exhaust PM10 Emissions With Tier 2/Sulfur
No Diesel Growth (Annual Tons)
Reductions from gasoline vehicles would result almost entirely from sulfur control, rather
than the proposed PM10 exhaust standards. PM10 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 PM10 standards contained in today's proposal. As shown, a
significant reduction in light-duty PM10 emissions would be realized immediately with sulfur
control. The reductions presented above presume 120 ppm in 2004 and 90 ppm in 2005 under
the sulfur ABT program; reductions could be realized as early as 2000 under this program. We
project that nearly 60 percent of light-duty PM10 emissions will be reduced when 30 ppm fuel is
introduced in 2006.
In addition to light-duty PM benefits, sulfur control would reduce PM10 emissions from
heavy-duty gasoline vehicles. We estimate these benefits would be approximately 1,100 tons
per year beginning with 30 ppm fuel, increasing to 1,500 tons by 2030. Across all sources, we
project Tier 2/Sulfur control would reduce direct PM10 from all non-natural sources by about one
percent.
b.
'Increased Growth" Sales Scenario
m-35
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Our "increased growth" scenario has been revised since the proposal. For the proposal,
the increased growth scenario was developed with the intent of analyzing an upper bound for
diesel growth, and assumed very aggressive levels of diesel penetration in the light truck market.
Since the proposal, we have derived more realistic growth assumptions based on work by A.D
Little, Inc.; the resulting growth scenario is referred to as the A.D. Little "Most Likely" diesel
growth scenario. The original A.D. Little methodology presented sales penetrations for LDVs
and LDTs in five-year increments, through 2015. We filled in the missing years using linear
interpolation, and assumed no growth beyond 2015. For this analysis, we assumed that diesel
LOT sales penetration would be distributed equally between the four truck classes. The resulting
diesel sales penetrations are shown in Table ni.A-13.
Table IILA-13. Diesel LDT Sales Penetration Under Increased Growth Scenario
Model Year
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015 and later
Diesel Sales Penetration
LDV
0.1%
0.1%
0.1%
0.1%
0.3%
0.7%
1.0%
1.3%
1.7%
2.0%
3.4%
4.8%
6.2%
7.6%
9.0%
LDT
0.1%
1.5%
3.0%
4.5%
6.0%
8.2%
10.4%
12.6%
14.8%
17.0%
18.4%
19.8%
21.2%
22.6%
24.0%
m-36
-------
Chapter III: Environmental Impact
/'. Light-Duty Direct Exhaust PM10 Trends Without Tier 2/Sulfur
Our projections for light-duty direct exhaust PM10 under the increased diesel sales
scenario are shown in Table IH.A-14 and Figure HI.A-17. As expected, this scenario is projected
to result in dramatic increases in light-duty PM10 emissions. 2010 baseline emissions are
approximately 54,000 tons, 28 percent higher than the 42,000 tons projected in the no growth
diesel case from Table ni.A-11. However, by 2030, we project this scenario would result in
direct PM emissions of 123,000 tons, over two times higher than the emissions projected for the
no growth scenario in the same year.
Table IILA-14. 47 State Light-Duty Direct Exhaust PM10 Emissions Without Tier 2/Sulfur
Increased Diesel Growth Scenario
(Annual Tons)
Year
2000
2004
2007
2010
2015
2020
2030
Light-Duty
Emissions Without
Tier 2/Sulfur
38,729
39,418
44,749
54,094
76,309
97,708
122,608
Contribution by Fuel Type
Diesel LDV/LDT
7%
6%
13%
26%
47%
58%
63%
Gasoline LDV/LDT
93%
94%
87%
74%
52%
42%
37%
m-37
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
125,000
100,000
75,000
50,000
25,000
2000
2005
2010
2015
2020
2025
2030
D Diesel
Gasoline
Figure IILA-17: 47-State Light-Duty Direct Exhaust PM10 Without Tier 2/Sulfur
Increased Diesel Sales (Annual Tons)
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 2007, we project diesels
would already account for 13 percent of all light-duty emissions. Diesel contribution grows to
over 25 percent by 2010 and over 60 percent by 2030.
/'/'. Direct Exhaust PM10 Reductions Due To Tier 2/Sulfur
Tier 2/Sulfur control would effectively neutralize excess PM emissions generated under
our increased diesel penetration scenario. Table in.A-15 contains reductions in direct exhaust
PM10 emissions due to Tier 2/Sulfur standards for the increased diesel sales penetration case.
Figure in. A-18 shows these emissions with Tier 2/Sulfur control, broken down by diesel and
gasoline. It should be noted that these emission reductions assume an average PM certification
standard of 0.01 grams per mile for all vehicles, and hence reflect a "best case" scenario for
diesel growth. Under the certification bin system, it is likely that many diesels would certify in
the 0.02 g/mi PM bin; under the "worst-case" scenario in which all diesels certify in this bin, we
project that emissions in 2030 would be approximately 8,000 tons higher than the "With Tier
HI-38
-------
Chapter III: Environmental Impact
2/Sulfur" scenario shown in Table IHA-15.9
Table IILA-15. 47-State Light-Duty Direct Exhaust PM10 Reductions Due To Tier 2/Sulfur
Increased Diesel Growth Scenario
(Annual Tons)
Year
2004
2007
2010
2015
2020
2030
Light-Duty
Emissions Without
Tier 2/Sulfur
39,418
44,749
54,094
76,309
97,708
122,608
Light-Duty
Emissions With
Tier 2/Sulfur
25,184
19,017
19,502
20,511
21,928
24,838
Emissions
Reduced
14,234
25,732
34,592
55,798
75,780
97,770
Percent Reduction in
Baseline Inventory
Light-Duty
36%
58%
64%
73%
78%
80%
All Sources*
-
0.9%
1.2%
1.8%
2.4%
2.8%
* 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.
In 2004, 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. By
2015, over 70 percent of baseline light-duty exhaust PM10 inventory is reduced. By 2030, 80
percent of the baseline emissions would be reduced, nearly 3 percent of total inventory. As
shown, today's action would largely serve to mitigate the PM10 emission increases which would
result from rapid diesel market penetration.
9 Under this "worst-case" PM scenario, all diesels under the A.D. Little "Most Likely" diesel sales case
(Table III.A.-13) would certify in a bin with a 0.02 g/mi PM standard (0.2 or 0.15 g/mi NOx). The remaining
vehicles (all gasoline) would certify in bins with a PM standard of 0.01 g/mi and a NOx standard or 0.07 g/mi or
less.
m-39
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
125,000 -r
100,000 WITHOUT TIER 2/SULFUR CONTROLS
75,000
50,000
25,000
2000
2005
2010
2015
2020
2025
Diesel
Gasoline
2030
Figure IILA-16. 47-State Light-Duty Direct Exhaust PM10 Emissions With Tier 2/Sulfur
Increased Diesel Sales (Annual Tons)
m-40
-------
Chapter III: Environmental Impact
B. Ozone: Baseline Nonattainment and Program Impacts
1. General Description of the EPA Ozone Modeling Used In This Rulemaking
A series of air quality modeling simulations were completed to a) support determination
of the need for additional emissions reductions in order to meet the ozone NAAQS, b) assess the
impact of the Tier 2/Sulfur rule on future ozone levels, and c) provide projected air quality
information to support the benefits/cost analysis. The model simulations were performed for five
emissions scenarios: a 1996 base year, a 2007 baseline projection, a 2007 projection with Tier
2/Sulfur controls, a 2030 baseline projection, and a 2030 projection with Tier 2/Sulfur controls.
These scenarios and the underlying emissions inventories associated with each are described in
Section HI. A.
In conjunction with current air quality data, as explained below, the model output from
the 2007 and 2030 baselines was used to identify areas expected to exceed the ozone NAAQS in
2007 and 2030. (These areas became candidates for being determined to be residual exceedance
areas which will require additional emission reductions to attain and maintain the ozone
NAAQS. As described in Section B.2, the impacts of the Tier 2/Sulfur controls were determined
by comparing the model results in the future year control runs against the NLEV/high sulfur
baseline simulations of the same year. The procedures for using the air quality modeling results
in the benefits/costs analysis are described in Section IV. The remainder of this section provides
a summary of the ozone modeling methodologies used to support this rulemaking. Additional
details are provided in the Tier 2 Final Rule Air Quality Modeling Technical Support Document.
The structure of this section is as follows:
Subsection 1 .a. describes in general terms the ozone modeling that was used to
determine the need for, and estimate the effects of, Tier 2/Sulfur controls.
Subsection l.b. describes the method used to determine residual non-attainment
areas.
a. Modeling Methodology
A variable-grid version of the Urban Airshed Model (UAM-V) was utilized to estimate
base and future-year ozone concentrations over the continental U.S. for the various emissions
scenarios. UAM-V simulates the numerous physical and chemical processes involved in the
formation, transport, and destruction of ozone. This model is commonly used for purposes of
determining attainment/non-attainment as well as estimating the ozone reductions expected to
occur from a reduction in emitted pollutants. The following sections provide an overview of the
ozone modeling completed as part of this rulemaking. More detailed information is included in
the Tier 2 Air Quality Modeling Technical Support Document (TSD).
HI-41
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Modeling domains
Two separate modeling domains were utilized in the Tier 2/Sulfur analyses. The first
covered that portion of the U.S. east of west longitude 99 degrees. The second covered the
remainder of the U.S. west of west longitude 99 degrees. The model resolution was 36 km over
the outer portions of each domain and 12 km in the inner portion of the grids. The model grids
for the eastern U.S. modeling were portioned into nine vertical layers with a surface layer depth
of 50 meters and a top level well above the typical mixed layer height (4000 meters). The model
grids for the western U.S. modeling were portioned into eleven vertical layers with a surface
layer depth of 50 meters and a top level of 4800 meters.
A recent modeling study (LADCO, 1999) considered the sensitivity of regional modeling
strategies to grid resolution. This study showed that the spatial pattern and magnitude of the
ozone changes at 4 km in response to emissions reductions were slightly more pronounced, but
generally similar to the modeled changes at 12 km in the Lake Michigan area. The OTAG10
modeling application also investigated the effects of grid resolution on national/regional control
strategies (e.g., Tier 2/Sulfur). The OTAG Final Report concluded that: a) peak simulated ozone
is generally higher with more highly resolved grids, b) spatial concentration patterns are
comparable between the fine and the coarse grid, and c) NOx reductions produce widespread
ozone decreases and occasional limited ozone increases with either the fine or the coarse grid
(although the increases tend to be larger in magnitude when finer-scale grids are used). More
detail on the effect of grid size upon model results is provided in the response to comments and
the TSD for this rule.
Modeling episodes
Three multi-day meteorological scenarios during the summer of 1995 were used in the
model simulations over the eastern U.S.: 12-24 June, 5-15 July, and 7-21 August. These periods
featured ozone exceedances at various times over most areas of the eastern U.S.11. In general,
these episodes do not represent extreme ozone events but, instead, are generally representative of
ozone levels near local design values12. Five simulations were completed for the June and July
10 The OTAG modeling project is used as a benchmark for the Tier 2/Sulfur modeling because it is the
most extensive regional ozone modeling application completed to date in terms of days modeled, areas covered, and
efforts of the air pollution modeling community to obtain sound model performance.
11 Each modeling episode contains three days for which the modeling results are not considered. These
days are simulated to minimize the dependence of the modeling results on uncertain initial conditions.
12 The fourth highest daily maximum 1-hour average ozone concentration measured over a three-year
period at a given monitor is the design value. Design values are used to determine the attainment status of a given
region.
m-42
-------
Chapter III: Environmental Impact
episodes (1996 base, 2007 baseline, 2007 control, 2030 baseline, 2030 control). Three
simulations were completed for the August episode (1996 base, 2007 baseline, 2007 control).
Two episodes were modeled for the western U.S. domain: 5-15 July 1996 and 18-31 July
1996. Again, these 19 days contained design value level ozone exceedances over most of the
western U.S. allowing for an assessment of emission controls in polluted, but not infrequent,
conditions. The primary purpose of simulating the western episodes was to provide data for the
benefits/cost analysis for 2030. Thus, no 2007 simulations were made for the West.
Non-emissions modeling inputs
The meteorological data required for input into UAM-V (wind, temperature, vertical
mixing, etc.) were developed by a separate meteorological model, the Regional Atmospheric
Modeling System (RAMS) for the eastern U.S. 1995 episodes, and the Fifth-Generation National
Center for Atmospheric Research (NCAR) / Penn State University (PSU) Mesoscale Model
(MM5) for the western U.S. 1996 episodes. These models provided needed data at every grid
cell on an hourly basis. These meteorological modeling results were evaluated against observed
weather conditions before being input into UAM-V and it was concluded that the model fields
were adequate representations of the historical meteorology.
The modeling assumed background pollutant levels at the top and along the periphery of
the domain. Additionally, initial conditions were assumed to be relatively clean as well. Given
the ramp-up days and the expansive domains, it is expected that these assumptions will not affect
the modeling results, except in areas near the boundary (e.g., Dallas-Fort Worth TX). The other
non-emission UAM-V inputs (land use, photolysis rates, etc.) were developed using procedures
employed in the OTAG regional modeling. The development of model inputs is discussed in
greater detail in the Tier 2 Air Quality Technical Support Document.
Model performance evaluation
The purpose of the Tier 2/Sulfur base year modeling was to reproduce the atmospheric
processes resulting in the observed ozone concentrations over these domains and episodes. One
of the fundamental assumptions in ozone modeling is that a model which closely replicates
observed ozone in the base year can be used to support future-year policymaking.
As with previous regional photochemical modeling studies, the accuracy of the Tier
2/Sulfur model base year simulations of historical ozone patterns varies by day and by location
over this large modeling domain. From a qualitative standpoint, there appears to be considerable
similarity on most days between the observed and simulated ozone patterns. Additionally, where
possible to discern, the model appears to follow the regional-scale ozone trends fairly closely.
The values of two primary measures of model performance, mean normalized bias and
HI-43
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
mean normalized gross error, indicate that the Tier 2/Sulfur modeling over the eastern U.S. is
generally as good or better than the grid modeling done for OTAG13, as shown in Table ni.B-1.
As OTAG did not perform any modeling for the West, no comparison back to OTAG is possible
for the Tier 2 western U.S. model performance. Mean normalized bias is defined as the average
difference between model predictions and observations (paired in space and time) normalized by
the observations. Mean gross error is defined as the average absolute difference between model
predictions and observations, paired in space and time, normalized by the observations. EPA
guidance on local ozone attainment demonstration modeling (not the purpose of the Tier 2
modeling) suggests biases be less than 5-15 percent and error be less than 30-35 percent.
Table III.B-1. Comparison of eastern U.S. regional model performance statistics
between the Ozone Transport Assessment Group (OTAG) modeling used to support the
NOX SIP call and the Tier 2/Sulfur modeling. The units are percentages.
Mean
Normalized Bias
Domain
Midwest
Northeast
Southeast
Southwest
OTAG
1988
Episode
-8
-15
o
-6
+2
-6
OTAG
1991
Episode
-4
-8
-6
+15
+6
OTAG
1993
Episode
+1
-8
-8
+21
+2
OTAG
1995
Episode
+4
-5
+8
+9
+12
Tier 2
June 95
Episode
-10
-11
-17
-4
+2
Tier 2
July 95
Episode
-6 (-4)14
-13 (-8)
-9 (-9)
+4 (+5)
+8 (+8)
Tier 2
August 95
Episode
+2
+7
-9
+7
+6
Mean
Normalized
Gross Error
Domain
Midwest
Northeast
Southeast
Southwest
OTAG
1988
Episode
28
27
29
28
22
OTAG
1991
Episode
25
26
23
25
24
OTAG
1993
Episode
27
25
23
32
23
OTAG
1995
Episode
25
24
26
27
29
Tier 2
June 95
Episode
24
24
27
20
24
Tier 2
July 95
Episode
24 (24)
26 (25)
22 (21)
24 (24)
27 (26)
Tier 2
August 95
Episode
23
22
24
22
24
In general, the model underestimates ozone for the June and July eastern episodes in 1995
13 Again, the OTAG application is used as a relative benchmark for model performance because it is the
most detailed modeling to date over this region.
14 Values in parentheses are for the 10-15th only. These dates correspond with OTAG episode days.
HI-44
-------
Chapter III: Environmental Impact
and, especially, both western episodes in 1996. The under prediction bias in the western U.S.
modeling averages about 40 percent. The model is slightly biased toward overestimation in the
August 1995 eastern episode. Although the overall tendency is to underestimate the observed
ozone, there are several instances in which overestimations occurred. The net effect is expected
to be an underestimate of the total extent of future-year exceedances, although some individual
areas may be overstated.
Application of the modeling results
As discussed in the preamble and in other sections of this document, the grid modeling is
being utilized to support the need for the Tier 2/Sulfur rulemaking and to determine the effects of
the emissions reductions on ozone air quality, with results reported at the level of CMSAs and
MSAs, as described in Section ni.B.3, below. Section Vn of this document discusses how these
modeling results are used in the cost benefit analysis.
b. Determining Need for Additional Emissions Reductions
Table in.B-1 of the Tier 2/Sulfur preamble lists those metropolitan areas which were
determined to require additional emission reductions in order to attain and maintain the 1-hour
ozone NAAQS. This determination was made for all areas with current design values greater
than or equal to 125 ppb and with modeling evidence that exceedances will persist into the
future15. The following sections provide the details inherent in both parts of this determination.
Air quality design values
An ozone design value is the concentration that determines whether a monitoring site
meets the NAAQS for ozone. Because of the way they are defined, design values are determined
based on three consecutive-year monitoring periods. 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. 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 part per billion (ppb). Due to the precision with
which the standards are expressed (0.12 parts per million (ppm) for the 1-hour, a violation of the
1-hour standard is defined as a design value greater than or equal to 125 ppb.
For a county, the design value is the highest design value from among all the monitors
with valid design values within that county. If a county does not contain an ozone monitor, it
does not have a design value. For most of our analyses, county design values are consolidated
where possible into design values for consolidated metropolitan statistical areas (CMSA) or
15 Modeling evidence from non-EPA analyses were also considered, as described in Section III.B.3, below.
m-45
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
metropolitan statistical areas (MSA). The design value for a metropolitan area is the highest
design value among the included counties. Counties that are not in metropolitan areas are treated
separately. For the purposes of defining the current design value of a given area, the higher of
the 1995-1997 and 1996-1998 design values were chosen to provide greater confidence in
identifying areas likely to have an ozone problem in the future. The 1995-1997 and 1996-1998
design values are listed in the Tier 2 Air Quality Modeling Technical Support Document.
Method for projecting future exceedances
The exceedance method was used for interpreting the future-year modeling results to
determine where nonattainment is expected to occur in the 2007 and 2030 Base Cases16. As part
of this method the modeling grid cells are first assigned to individual areas. The daily maximum
1-hour ozone values predicted in grid cells assigned to an area are then checked to identify
whether there are any predictions >=125 ppb. Areas with current measured violations of the one-
hour ozone standard, one or more model-predicted exceedances, and no conflicting modeling
evidence to the contrary are projected to have a nonattainment problem in the future. This
procedure is further described in the Tier 2 Air Quality Modeling Technical Support Document.
2. Ozone Reductions Expected from this Rule
The large reductions in emissions of ozone precursors from today's standards will be very
beneficial to federal and State efforts to lower ozone levels and bring about attainment with the
current one-hour ozone standards. The air quality modeling for the final rule shows that
improvements in ozone levels are expected to occur throughout the country because of the Tier
2/Gasoline Sulfur program.17 EPA found that the program significantly lowers model-predicted
exceedances of the ozone standard. In 2007 the number of exceedances in CMS A/MS As is
forecasted to decline by nearly one-tenth and in 2030, when full turnover of the vehicle fleet has
occurred, the program lowers such exceedances by almost one-third. In these same areas, the
total amount of ozone above the NAAQS is forecasted to decline by about 15 percent in 2007
and by more than one-third in 2030. In the vast majority of areas, the air quality modeling
predicts that the program will lower peak summer ozone concentrations for both 2007 and 2030.
The reduction in daily maximum ozone is nearly 2 ppb, on average in 2007 and over 5 ppb, on
average in 2030. These reductions contribute to EPA's assessment that the program will provide
the large set of public health and environmental benefits summarized in Section Vn. The
forecasted impacts of the program on ozone in 2007 and 2030 are further described in the Tier 2
Air Quality Modeling Technical Support Document.
16 2030 is the relevant baseline scenario for the western U.S. domain
17EPA assessment of air quality changes for 2007 and 2030 focused on 37 States in the East because these
States cover most of the areas with 1-hour nonattainment problems.
HI-46
-------
Chapter III: Environmental Impact
During the public comment period on the proposed rule, EPA received several comments
that expressed concern about potential increases in ozone that might occur as a result of this rule.
As indicated above, the air quality modeling results indicate an overall reduction in ozone levels
in 2007 and 2030 during the various episodes modeled. In addition to ozone reductions, a few
areas had predicted ozone increases in portions of the area during parts of the episodes modeled.
In most of these cases, the overall decrease led EPA to conclude that there will be a net reduction
in ozone levels in these areas due to the Tier 2/Sulfur program. In the very small number of
exceptions to this, the Agency did find benefit from reduction of peak ozone levels. Based upon
a careful examination of this issue, including EPA's modeling results as well consideration of
the modeling and analyses submitted by commenters, it is clear that the significant ozone
reductions from this rule outweigh the limited ozone increases that may occur. Additional details
on this issue are provided in the Response to Comments Document and in the Tier 2 Air Quality
Modeling Technical Support Document.
Collectively, EPA believes these results indicate that it will be much easier for States to
provide EPA assurances that their State Implementation Plans will attain and maintain
compliance with the one-hour ozone standards. In the limited number of cases mentioned above,
EPA will work with States who will be conducting more detailed local modeling of their specific
local programs that they have designed to provide attainment. Notably, there are also other
upcoming federal measures to lower ozone precursors will aid these efforts. If the State
modeling of local programs shows a need, the Agency will work with states to plan further
actions to produce attainment with the NAAQS in order to protect the public's health and the
environment. Further details on EPA's modeling results can be found in the response to
comments and technical support documents.
3. Ozone Modeling and Analysis in 1-Hour State Implementation Plan
Submittals and Other Local Ozone Modeling
a. Overview
We have reviewed and recently proposed action on SIP submissions from 14 states
covering 10 serious and severe 1-hour ozone nonattainment areas. We received these
submissions as part of the three-phase SIP process allowed by EPA guidance memos or as part of
a request for an attainment date extension. These submissions also provided ozone modeling
results for two attainment areas in a downwind state. These submissions contain local ozone
modeling which we considered along with the results of the EPA ozone modeling described
above. We have also considered ozone modeling submitted as part of an attainment date
extension request for Beaumont-Port Arthur, TX, but have not yet taken action on that request.
We have also reviewed a status report on the results of modeling being conducted in anticipation
of submittal to EPA as part of an extension request for Dallas, TX. Finally, we have considered
information in the most recent SIP submittal from California for the South Coast Air Basin.
Table in.B-2 lists the areas involved, our overall conclusion as to whether the modeling
HI-47
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
demonstrates attainment without reductions that would be considered "further reductions" under
CAA section 202(i), and whether the area is included in our recent proposals for actions on SIP
submittals. The Federal Register notices for these recent proposals appeared together on
December 16, 1999, beginning at 64 FR 70318. This section discusses the background for the
submissions and our conclusions from them.
The local modeling analyses generally cover a modeling domain encompassing one or a
few closely spaced nonattainment areas and a limited upwind area. Because of this limited
domain, states have been able to use grid cells of 4 or 5 kilometers on a side, in keeping with
EPA guidance for such modeling. This fine grid size is an important factor in how much weight
we have given to this set of evidence. The future attainment date examined differs from state to
state depending on its current (or proposed extended) attainment deadline. In the state modeling,
ozone episode days were selected by the respective states based on days with high ozone in the
local domain being modeled. In all cases, the selection of episode days met our guidance. The
local modeling also makes use of more information on the local emission inventory and control
program than is practicable to include in broad scale modeling by EPA as described above.
The SIP submissions for these 10 nonattainment areas contain many legally required
elements in addition to the attainment demonstrations. After considering the attainment
demonstrations and these other elements, we have proposed appropriate action on each of these
submissions. In many cases, we have proposed alternative actions on our part, based on whether
the state submits additional SIP elements which we have described as necessary. We also
explained what each state m
ust provide us in order to allow us to take final approval or conditional approval action.
More specific descriptions of the ozone modeling contained in the SIPs, for areas where
we have recently proposed action on a submittal, and more explanation of our evaluation of it can
be obtained in the individual Federal Register notices and in the technical support document
prepared for each action.
HI-48
-------
Chapter III: Environmental Impact
Table III.B-2. Nonattainment Areas For Which EPA Has Recently Proposed Action On
SIP Submissions Containing 1-hour Ozone Attainment Demonstrations or Otherwise Has
Considered Results of Local Ozone Modeling
Nonattainment Area
(Major Metro Area)
Western
Massachusetts
(Springfield)
Greater Connecticut
(Hartford and other
MSAs)
New York City
Philadelphia
Baltimore
Washington, D.C.
Atlanta
Houston
Chicago
Milwaukee
Benton Harbor
Grand Rapids
Dallas
Affected
States
MA
CT
NY, CT,
NJ
PA, NJ,
DE, MD
MD
MD, VA,
D.C.
GA
TX
IL, IN
WI
MI
MI
TX
Attainment
Date
2003
(Requested
Extension)
2007
(Requested
Extension)
2007
2005
2005
2005
(Requested
Extension)
2003
(Requested
Extension)
2007
2007
2007
Not
Applicable
Not
Applicable
2007
Demonstrates
Attainment
Without
"Further
Reductions "
Yes
Yes
No
No
No
No
No
No
Yes
Yes
Yes
Yes
No
Proposed for Action in
December 16, 1999
Federal Register
(64 FR 703 18)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
m-49
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Nonattainment Area
(Major Metro Area)
Beaumont-Port
Arthur
South Coast Air
Basin
Affected
States
TX
CA
Attainment
Date
2007
(Requested
Extension)
2010
Demonstrates
Attainment
Without
"Further
Reductions "
No
No
Proposed for Action in
December 16, 1999
Federal Register
(64 FR 703 18)
Yes
No
m-50
-------
Chapter III: Environmental Impact
b. CAA Requirements and EPA Policy
The CAA, as amended in 1990, required EPA to designate as nonattainment any area that
was violating the 1-hour ozone standard, generally based on air quality monitoring data from the
three-year period from 1987-1989. CAA § 107(d)(4); 56 FR 56694 (Nov. 6, 1991). The CAA
further classified these areas, based on the area's design value, as marginal, moderate, serious,
severe or extreme. Marginal areas were suffering the least significant air pollution problems
while the areas classified as severe and extreme had the most significant air pollution problems.
The control requirements and dates by which attainment needs to be achieved vary with
the area's classification. Marginal areas are subject to the fewest mandated control requirements
and have the earliest attainment date. Severe and extreme areas are subject to more stringent
planning requirements but are provided more time to attain the standard. Serious areas were
required to attain the 1-hour standard by November 15, 1999 and severe areas are required to
attain by November 15, 2005 or November 15, 2007.
Under section 182(c)(2) and (d) of the CAA, serious and severe areas were required to
submit demonstrations of how they would attain the 1-hour standard and how they would achieve
reductions in VOCs and NOx emissions of 9 percent for each three-year period until the
attainment year (rate-of-progress or ROP) by November 15, 1994.
In general, an attainment demonstration SIP includes a modeling analysis component
showing how the area will achieve the standard by its attainment date and the control measures
necessary to achieve those reductions. Another component of the attainment demonstration SIP
is a motor vehicle emissions budget for transportation conformity purposes. Transportation
conformity is a process for ensuring that States consider the effects of emissions associated with
new or improved federally-funded roadways on attainment of the standard. As described in
section 176(c)(2)(A), attainment demonstrations necessarily include the estimates of motor
vehicle emissions that are consistent with attainment, which then act as a budget or ceiling for
the purposes of determining whether transportation plans and projects conform to the attainment
SIP.
Notwithstanding significant efforts by the States, in 1995 EPA recognized that many
States in the eastern half of the United States could not meet the November 1994 time frame for
submitting an attainment demonstration SIP because emissions of NOx and VOCs in upwind
States (and the ozone formed by these emissions) affected these nonattainment areas and the full
impact of this effect had not yet been determined. This phenomenon is called ozone transport.
On March 2, 1995, Mary D. Nichols, EPA's then Assistant Administrator for Air and
Radiation, issued a memorandum to EPA's Regional Administrators acknowledging the efforts
made by States but noting the remaining difficulties in making attainment demonstration SIP
HI-51
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
submittals.8 Recognizing the problems created by ozone transport, the March 2, 1995
memorandum called for a collaborative process among the States in the eastern half of the
country to evaluate and address transport of ozone and its precursors. This memorandum led to
the formation of the Ozone Transport Assessment Group (OTAG)9 and provided for the States to
submit the attainment demonstration SIPs based on the expected time frames for OTAG to
complete its evaluation of ozone transport.
In June 1997, OTAG concluded and provided EPA with recommendations regarding
ozone transport. The OTAG generally concluded that transport of ozone and the precursor NOx
is significant and should be reduced regionally to enable States in the eastern half of the country
to attain the ozone NAAQS.
In recognition of the length of the OTAG process, in a December 29, 1997 memorandum,
Richard Wilson, EPA's then Acting Assistant Administrator for Air and Radiation, provided
until April 1998 for States to submit for attainment demonstration SIPs and clarified that, by
April 1998, States with serious and higher classified nonattainment areas additionally needed to
submit (1) evidence that the applicable control measures in subpart 2 of part D of title I of the
CAA were adopted and implemented or were on an expeditious course to being adopted and
implemented; (2) a list of measures needed to meet the remaining ROP emissions reduction
requirement and to reach attainment; (3) for severe areas only, a commitment to adopt the control
measures necessary for attainment and ROP plans through the attainment year by the end of
2000; (4) a commitment to implement the SIP control programs in a timely manner and to meet
ROP emissions reductions and attainment; and (5) evidence of a public hearing on the State
submittal.10
Building upon the OTAG recommendations and technical analyses, in November 1997,
EPA proposed action addressing the ozone transport problem. In its proposal, the EPA found
that current SIPs in 22 States and the District of Columbia (23 jurisdictions) were insufficient to
provide for attainment and maintenance of the 1-hour standard because they did not regulate
NOx emissions that significantly contribute to ozone transport. 62 FR 60318 (Nov. 7, 1997).
The EPA finalized that rule in September 1998, calling on the 23 jurisdictions to revise their SIPs
to require NOx emissions reductions within the State to a level consistent with a NOx emissions
budget identified in the final rule. 63 FR 57356 (Oct. 27, 1998). This final rule is commonly
referred to as the NOx SIP Call or the Regional Ozone Transport Rule.
On July 16, 1998, EPA's then Acting Assistant Administrator, Richard Wilson, issued a
guidance memorandum intended to provide further relief to areas affected by ozone transport.11
The memorandum recognized that many moderate and serious areas are affected by transported
pollution from either an upwind area in the same State with a higher classification and later
attainment date, and/or from an upwind area in another State that is significantly contributing to
the downwind area's nonattainment problem. The policy recognized that some downwind areas
may be unable to meet their own attainment dates, despite doing all that was required in their
HI-52
-------
Chapter III: Environmental Impact
local area, because an upwind area may not have adopted and implemented all of the controls
that would benefit the downwind area through control of transported ozone before the downwind
area's attainment date. Thus, the policy provided that upon a successful demonstration that an
upwind area has interfered with attainment and that the downwind area is adopting all measures
required for its local area for attainment but for this interference, EPA may grant an extension of
the downwind area's attainment date. Local area measures would include all of the measures
within the modeling domain that were relied on for purposes of the modeled attainment
demonstration. Once an area receives an extension of its attainment date based on transport, the
area would no longer be subject to reclassification to a higher classification and subject to
additional requirements for failure to attain by its original attainment date provided it was doing
all that was necessary locally. The policy provides that the area must meet four criteria to receive
an attainment date extension. In summary, the area must: (1) be identified as a downwind area
affected by transport from either an upwind area in the same State with a later attainment date or
an upwind area in another State that significantly contributes to downwind nonattainment; (2)
submit an approvable attainment demonstration with any necessary, adopted local measures and
with an attainment date that reflects when the upwind reductions will occur; (3) adopt all local
measures required under the area's current classification and any additional measures necessary
to demonstrate attainment; and (4) provide that it will implement all adopted measures as
expeditiously as practicable, but no later than the date by which the upwind reductions needed for
attainment will be achieved.
The States generally submitted the SIPs between April and October of 1998; some States
are still submitting additional revisions. Under the CAA, EPA is required to approve or
disapprove a State's submission no later than 18 months following submission. (The statute
provides up to 6 months for a completeness determination and an additional 12 months for
approval or disapproval.)
c. Local Ozone Modeling in SIP Submissions
The EPA provides that States may rely on a modeled attainment demonstration
supplemented with additional evidence to demonstrate attainment. In order to have a complete
modeling demonstration submission, States have submitted the required modeling analysis and
identified any additional evidence that EPA should consider in evaluating whether the area will
attain the standard.
For purposes of demonstrating attainment, the CAA requires serious and severe areas to
use photochemical grid modeling or an analytical method EPA determines to be as effective.
The EPA has issued guidance on the air quality modeling that is used to demonstrate attainment
with the 1-hour ozone NAAQS.12 The photochemical grid model is set up using meteorological
conditions conducive to the formation of ozone. Emissions for a base year are used to evaluate
the model's ability to reproduce actual monitored air quality values and to predict air quality
changes in the attainment year due to the emission changes which include growth up to and
HI-53
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
controls implemented by the attainment year. A modeling domain is chosen that encompasses
the nonattainment area. Attainment is demonstrated when all predicted concentrations inside the
modeling domain are at or below the NAAQS or at an acceptable upper limit above the NAAQS
permitted under certain conditions by EPA's guidance. When the predicted concentrations are
above the NAAQS, an optional weight of evidence determination which incorporates but not is
limited to other analyses such as air quality and emissions trends may be used to address
uncertainty inherent in the application of photochemical grid models.
The EPA guidance identifies the features of a modeling analysis that are essential to
obtain credible results. First, the State must develop and implement a modeling protocol. The
modeling protocol describes the methods and procedures to be used in conducting the modeling
analyses and provides for policy oversight and technical review by individuals responsible for
developing or assessing the attainment demonstration (State and local agencies, EPA Regional
offices, the regulated community, and public interest groups). Second, for purposes of
developing the information to put into the model, the State must select air pollution days, i.e.,
days in the past with bad air quality, that are representative of the ozone pollution problem for the
nonattainment area. Third, the State needs to identify the appropriate dimensions of the area to
be modeled, i.e., the domain size. The domain should be larger than the designated
nonattainment area to reduce uncertainty in the boundary conditions and should include large
upwind sources just outside the nonattainment area. In general, the domain is considered the
local area where control measures are most beneficial to bring the area into attainment. Fourth,
the State needs to determine the grid resolution. The horizontal and vertical resolutions in the
model affect the dispersion and transport of emission plumes. Artificially large grid cells (too
few vertical layers and horizontal grids) may dilute concentrations and may not properly consider
impacts of complex terrain, complex meteorology, and land/water interfaces. Fifth, the State
needs to generate meteorological that describe atmospheric conditions and emissions inputs.
Finally, the State needs to verify the model is properly simulating the chemistry and atmospheric
conditions through diagnostic analyses and model performance tests. Once these steps are
satisfactorily completed, the model is ready to be used to generate air quality estimates to support
an attainment demonstration.
The modeled attainment test compares model predicted 1-hour daily maximum
concentrations in all grid cells for the attainment year to the level of the NAAQS. A predicted
concentration above 0.124 ppm ozone indicates that the area is expected to exceed the standard
in the attainment year and a prediction below 0.124 ppm indicates that the area is expected to
attain the standard. This type of test is often referred to as an exceedance test. The EPA's
guidance recommends that States use either of two modeled attainment or exceedance tests for
the 1-hour ozone NAAQS: a deterministic test or a statistical test.
The deterministic test requires the State to compare predicted 1-hour daily maximum
HI-54
-------
Chapter III: Environmental Impact
ozone concentrations for each modeled day18 to the attainment level of 0.124 ppm. If none of the
predictions exceed 0.124 ppm, the test is passed.
The statistical test takes into account the fact that the form of the 1-hour ozone standard
allows exceedances. If, over a three-year period, the area has an average of one or fewer
exceedances per year, the area is not violating the standard. Thus, if the State models a very
extreme day, the statistical test provides that a prediction above 0.124 ppm up to a certain upper
limit may be consistent with attainment of the standard. (The form of the 1-hour standard allows
for up to three readings above the standard over a three-year period before an area is considered
to be in violation.)
The acceptable upper limit above 0.124 ppm is determined by examining the size of
exceedances at monitoring sites which meet or attain the 1-hour NAAQS. For example, a
monitoring site for which the four highest 1-hour average concentrations over a three-year period
are 0.136 ppm, 0.130 ppm, 0.128 ppm and 0.122 ppm is attaining the standard. To identify an
acceptable upper limit, the statistical likelihood of observing ozone air quality exceedances of the
standard of various concentrations is equated to severity of the modeled day. The upper limit
generally represents the maximum ozone concentration level observed at a location on a single
day and it would be the only level above the standard that would be expected to occur no more
than an average of once a year over a three-year period. Therefore, if the maximum ozone
concentration predicted by the model is below the acceptable upper limit, in this case 0.136 ppm,
then EPA might conclude that the modeled attainment test is passed. Generally, exceedances well
above 0.124 ppm are very unusual at monitoring sites meeting the NAAQS. Thus, these upper
limits are rarely significantly higher than the attainment level of 0.124 ppm.
When the modeling does not conclusively demonstrate that the area will attain, additional
analyses may be presented to help determine whether the area will attain the standard. As with
other predictive tools, there are inherent uncertainties associated with modeling and its results.
For example, there are uncertainties in some of the modeling inputs, such as the meteorological
and emissions data bases for individual days and in the methodology used to assess the severity
of an exceedance at individual sites. The EPA's guidance recognizes these limitations, and
provides a means for considering other evidence to help assess whether attainment of the
NAAQS is likely. The process by which this is done is called a weight of evidence (WOE)
determination.
Under a WOE determination, the State can rely on and EPA will consider factors such as
other modeled attainment tests, e.g., a rollback analysis; other modeled outputs, e.g., changes in
the predicted frequency and pervasiveness of exceedances and predicted changes in the design
value; actual observed air quality trends; estimated emissions trends; analyses of air quality
18The initial, "ramp-up" days for each episode are excluded from this determination.
m-55
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
monitored data; the responsiveness of the model predictions to further controls; and, whether
there are additional control measures that are or will be approved into the SIP but were not
included in the modeling analysis. This list is not an exclusive list of factors that may be
considered and these factors could vary from case to case. The EPA's guidance contains no limit
on how close a modeled attainment test must be to passing to conclude that other evidence
besides an attainment test is sufficiently compelling to suggest attainment. However, the further
a modeled attainment test is from being passed, the more compelling the WOE needs to be.
Special explanation is necessary on the issue of how the NOx SIP Call/Regional Ozone
Transport Rule has been handled by states in their local ozone modeling. For purposes of CAA
section 202(i) we consider the emission reductions that will be achieved by the NOx SIP
Call/Regional Ozone Transport Rule to be previous emission reductions rather than "further"
reductions, since the 22 affected states and the District of Columbia are currently under an
enforceable requirement to obtain these reductions. In most of the local ozone modeling in these
SIP revisions, upwind NOx reductions have been assumed to occur through implementation of
the the NOx SIP Call/Regional Ozone Transport Rule in some or all of the states subject to that
rule, even though all states' rules to implement those reductions have not yet been adopted.
Where upwind and local implementation of the NOx SIP Call is assumed, our conclusion that the
modeling shows that an area cannot attain the NAAQS means that it cannot attain even with the
prior implementation of the NOx SIP Call, and thus the area requires further emission reductions
under section 202(i).19
d. Conclusions from the Local Modeling in SIP Submittals
All of the states have made use of the weight of evidence concept in their attainment
demonstrations. EPA has proposed to find that some of the demonstrations are adequate, while
for others additional reductions are needed to attain. There is an important distinction between
our proposed finding on these SIPs and the determination required by CAA section 202(i). We
are in some cases proposing to approve demonstrations that depend on emission reductions from
measures that the state has not yet adopted and has not yet made a legally enforceable
commitment to adopt and implement. Such emission reductions are "further" reductions under
CAA section 202(i). In some cases, therefore, we are considering an area to need further
reductions in order to attain and maintain and also proposing to approve its attainment
demonstration. Before we take final and unconditional action on an attainment demonstration in
such a case, the state will have to adopt all the necessary rules or make enforceable commitments
to adopt them.
19 Our recent proposals on the SIPs explain how we propose to approach the approval of 1-hour attainment
SIPs themselves with respect to the NOx SIP Call. To summarize, we have proposed to approve a SIP which
assumes implementation of the NOx SIP Call provided that the state is committed to implementing the NOx
reductions within the in-state portion of the modeling domain of the subject nonattainment area. Reductions outside
the domain and in other states may be assumed even if a commitment is currently lacking for those areas.
HI-56
-------
Chapter III: Environmental Impact
These state-specific findings are not final and we are not making them final via the Tier 2
rulemaking. In our final actions on these SIP revisions we may deviate from our proposal for one
or more areas, based on the full record of the rulemaking for each, including any comments
received after today. However, we have used the ozone attainment assessments as described
summarized below in making our Tier 2 determination on the need for additional emission
reductions in these areas.
As result of EPA's review of the states' SIP submittals, EPA believes that the ozone
modeling submitted by the applicable states for the Chicago, IL; Milwaukee, WI; Greater CT
(Hartford and New London metropolitan areas); and Western MA areas demonstrate attainment
through the control measures contained in the submitted attainment strategy. We expect that
Illinois, Wisconsin, and Indiana will submit further SIP revisions for Chicago and Milwaukee
prior to our taking final action on our recent proposals regarding the submissions they made
earlier. These new revisions will be based on a new round of modeling conducted by the Lake
Michigan Air Directors Consortium (LADCO) on behalf of the states. While we have not
received this modeling, we have received a progress report on it.13
While Michigan was not required to submit attainment demonstrations for the Benton
Harbor and Grand Rapids-Muskegon areas, the ozone modeling submitted and weight of
evidence analysis performed for the attainment demonstrations submitted in 1998 for Chicago
and Milwaukee indicates that these two areas will also be in attainment in 2007 based only on
emission reductions which come from measures which are already adopted and legally
enforceable.
For the New York Metro area, Philadelphia, Baltimore, District of Columbia, and
Houston nonattainment areas, the EPA has proposed to determine that additional emission
reduction beyond those provided by the SIP submission are necessary for attainment. A portion
of that reduction will be achieved by EPA's Tier 2/Sulfur program. In the case of Washington
DC, our analysis indicates that the Tier 2/Sulfur program will provide all of the additional
emission reductions needed to attain.
Atlanta's statutory attainment date as a serious 1-hour ozone nonattainment area was
November 1999, which it has not met. Georgia has requested an attainment date extension for
Atlanta to November 15, 2003 and has proposed an emission control program to achieve
attainment by that date. The EPA has proposed to assign Atlanta an attainment date of
November 2003 based on a successful demonstration by the State that the control strategy
described in the SIP will achieve attainment by this date. However, many of the measures in that
strategy are not yet adopted or fully committed, and are therefore "further" reductions under
202(i). It is clear from the amount of emission reductions from these measures that the
nonattainment status of Atlanta would extend into the 2004 and later period if only "previous"
emission reductions were considered. The modeling for Atlanta assumed implementation of the
NOx SIP Call outside the local modeling domain, but lesser NOx reductions within the domain.
HI-57
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
This is an issue, since the full NOx reductions from the SIP call are considered "previous" in
connection with the Tier 2 rulemaking. However, the difference in NOx reduction within the
modeling domain is small, and it is apparent that even if the full NOx reductions from the SIP
call had been assumed attainment would still be impossible without reductions from measures
which are "further" reductions for purposes of the Tier 2 determinations.
The specific reasons for reaching these conclusions are explained in the individual
Federal Register notices.
e. Other Local Ozone Modeling
We have received ozone modeling for the Beaumont-Port Arthur nonattainment area.14
Beaumont-Port Arthur is a moderate ozone nonattainment area which failed to attain by its
November 15, 1996 deadline. Presently, the state of Texas is seeking our approval for a
demonstration that Beaumont-Port Arthur is impacted by ozone transport from the Houston area,
in order to support a request that we extend its attainment deadline to 2007 which would be the
same as the deadline for Houston. We proposed action on this request on April 16, 1999 (64 FR
18864) and extended the comment period on June 3, 1999 (64 FR 29822). The modeling
analysis indicates nonattainment in 2007 even under an emissions scenario that includes
reductions that must be considered "further emission reductions" under CAA section 202(i).
We have also recently become aware of recent modeling by the state of Texas for the
Dallas-Fort Worth metropolitan area.15 Dallas has failed to meet its 1999 deadline for attainment
Texas has made known its intent to seek an attainment date extension for this area. We have
recently indicated to Texas that we will propose to approve its request for an attainment date
extension to 2007, provided that the state can meet several necessary conditions one of which is
to demonstrate attainment by that date. The state is conducting modeling analysis to identify its
options for reaching attainment in Dallas by 2007. This modeling has been made public, and
summaries of it have been put in the docket for this rulemaking. The modeling results to date
indicate that even with the emission reductions from the Tier 2/Sulfur program, Dallas will be in
nonattainment in 2007. This clearly demonstrates further reductions in emissions under CAA
section 202(i) are needed to attain and maintain in this area.
We have not received any recent ozone modeling from California, because California
submitted and we approved the SIPs for nonattainment areas in California some time ago, before
our proposal for the Tier 2/Gasoline Sulfur program. However, the air quality situation and a
recent SIP revision for one area in California support the conclusion that there is an overall need
for further reductions in order to attain and maintain.
It is appropriate for us to consider the need for further emission reductions in order for
areas in California to attain and maintain. California contains many of the most ozone-impacted
areas in the nation. According to California, about 7 to 10 percent of all car and light truck travel
HI-58
-------
Chapter III: Environmental Impact
in California takes place in vehicles originally sold outside of California. Nine areas in
California currently designated as nonattainment (and two counties currently designated as being
in attainment) with a population of approximately 30 million have 1996-1998 design values
above the 1-hour ozone NAAQS. Seven of the nonattainment areas have approved SIPs,
including demonstrations of attainment for their required dates. The approved demonstrations
did not depend on any reductions in emissions from more stringent standards for cars sold
outside California, having been prepared prior to our proposal for these standards.
However, the state of California has recently filed an update to its State Implementation
Plan for the South Coast Air Basin that expressly claims that the Tier 2 program will lead to four
tons of reduced NOx emissions per day in the South Coast area in 2010, and includes this
reduction in the attainment strategy for the area. 16 The four tons per day NOx reductions cited
represents only a small fraction of the emission reductions needed in the South Coast to attain the
NAAQS.
The state is developing yet another revision to the South Coast plan, which we understand
will also depend in part on the emission reductions from Tier 2 vehicles originally sold out of
state. We expect that California will be submitting one or more similar revisions including
emission reductions from Tier 2 standards for some other areas, since it appears that some
serious classification nonattainment areas in California with an attainment deadline of 1999 have
not met that date. These areas are San Diego and the San Joaquin Valley. San Joaquin has had
too many exceedances to be eligible for an extension other than through reclassification to severe
or through an attainment extension based on overwhelming transport from an area with a later
attainment date, while San Diego appears to be eligible for a 1-year attainment date extension
under the provisions of CAA section 181(a)(5). We have not yet received an indication of
California's intention in this regard, or any modeling which assesses whether these areas can
attain before 2004 relying only on baseline measures with respect to CAA section 202(i)
Attainment of the 1-hour standard in the South Coast Air Basin, Southeast Desert,
Sacramento, and Ventura nonattainment areas by their future attainment dates (2010 for the
South Coast, 2007 for Southeast Desert, and 2005 for Sacramento and Ventura) remains the goal
of California and EPA, but will be a challenging task. The difficulty of the task is reflected in
ongoing litigation and settlement negotiations over both the design and the implementation of the
attainment plans in the South Coast, for example. We believe that there is a possibility that some
of these areas would not attain on schedule if we were not adopting the new standards for cars
and light trucks, and the sulfur limits for gasoline sold outside of California.
f. Need for Further Reductions in Emissions in Order to Attain and Maintain
After considering the results of the exceedance method applied to the ozone
concentrations predicted by the EPA ozone modeling described in Sections TUB. 1,2, and 3, and
HI-59
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
the conclusions from our review of the local ozone modeling described above, we determined
which areas are certain or highly likely to require further reductions to attain and maintain, under
the meaning of CAA section 202(i). Table in.B-1 in the preamble lists these areas.
We first considered all areas with predicted 2007 nonattainment according to the
exceedance method applied to the EPA ozone modeling.
Areas which did not have a 1-hour ozone design value above the NAAQS in at least one
of the 1995 to 1997 and 1996 to 1998 periods are not considered to be certain or highly likely to
need further reductions, regardless of predictions of exceedances in our modeling. However,
there were six areas predicted in the EPA modeling to have exceedances in 2007 which had
design values of 90 to 100 percent of the NAAQS in at least one of these two periods. We
consider these to have moderate risk of failing to attain and maintain without further emission
reductions because meteorological conditions may be more severe in the future. These are
Biloxi, Cleveland, Detroit, New Orleans, Pensacola, and Tampa.
Next, for Chicago, Milwaukee, and Greater Connecticut (Hartford and New London) we
considered the SIP's successful demonstration of attainment relying only on baseline measures to
be a sufficient reason not to consider these areas to be certain or highly likely to need further
reductions. However, because of modeling uncertainties and the fact that these SIPs did not
consider attainment and maintenance beyond the attainment date, we consider these areas to have
a significant individual risk of failing to attain and maintain. Benton Harbor and Grand Rapids-
Muskegon are included in this characterization also.
Because the SIP modeling analysis and the EPA exceedance method agreed on the
nonattainment prospects for Houston, New York, Philadelphia, Baltimore, Atlanta, and
Washington DC , excluding measures which would provide "further" reductions, we consider
these to be certain or highly likely to require further emission reductions in order to attain and
maintain the 1-hour ozone NAAQS.
Because the EPA modeling for Dallas did not successfully reproduce exceedances which
actually occurred in 1995 we considered it at best inconclusive. The local modeling supports our
conclusion that Dallas is certain or highly likely to need further reductions to attain and maintain.
Because the episode days in the EPA modeling did not include any on which Beaumont-
Port Arthur experiences exceedances, yet exceedances have been observed, we considered the
EPA modeling at best inconclusive. The local modeling led us to conclude that Beaumont-Port
Arthur is certain or highly likely to need further reductions to attain and maintain. We included
the Los Angeles-Riverside-San Bernardino CMSA (South Coast Air Basin) in this category as
well, based on its reliance on Tier 2 reductions in its most recent SIP submittal.
HI-60
-------
Chapter III: Environmental Impact
C. Particulate Matter and Visibility/Regional Haze
1. Particulate Matter
a. Background on Particulate Matter
Particulate matter (PM) represents a broad class of chemically and physically diverse
substances that exist as discrete particles (liquid droplets or solids) over a wide range of sizes.
The NAAQS that regulates PM addresses only PM with a diameter less than 10 microns, or
PM10. The coarse fraction of PM10 are those particles which have a diameter in the range of 2.5
to 10 microns, and the fine fraction being those particles which have a diameter less than 2.5
microns, or PM25. These particles and droplets are produced as a direct result of human activity
and natural processes, and they are also formed as secondary particles from the atmospheric
transformation of emissions of SOx, NOx, ammonia, and VOCs.
Natural sources of particles in the coarse fraction of PM10 include windblown dust, salt
from dried sea spray, fires, biogenic emanation (e.g., pollen from plants, fungal spores), and
volcanoes. Fugitive dust and crustal material (geogenic materials) comprise approximately 80
percent of the coarse fraction of the PM10 inventory as estimated by methods in use today.20
Manmade sources of these coarser particles arise predominantly from combustion of fossil fuel
by large and small industrial sources (including power generating plants, manufacturing plants,
quarries, and kilns), wind erosion from crop land, roads, and construction, dust from industrial
and agricultural grinding and handling operations, metals processing, and burning of firewood
and solid waste. Coarse-fraction PM10 remains suspended in the atmosphere a relatively short
period of time.
Most of the emission sources listed for coarse particles also have a substantial fine
particle fraction. Their share of the PM25 inventory is reduced somewhat, however, because of
the role of other sources that give rise to primarily PM25. The other sources of PM25 include
carbon-based particles emitted directly from gasoline and diesel internal combustion engines, and
a large component of secondary sulfate-based particles (formed from SOx and ammonia), nitrate-
based particles (formed from NOx and ammonia), and carbonaceous secondary particles (formed
through transformation of VOC emissions). PM25 from fugitive dust and crustal sources
(geogenic materials) comprise approximately one-half of the directly emitted PM25 inventory,
substantially less than their share of coarse PM emissions. The presence and magnitude of
20 There is evidence from ambient studies that emissions of these materials may be overestimated and/or
that once emitted they have less of an influence on monitored PM concentrations than this inventory share would
suggest.
ni-6i
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
crustal PM2 5 in the ambient air is much lower even than suggested by this smaller inventory
share, due to the additional presence of secondary PM from non-crustal sources and the removal
of a large portion of crustal emissions close to their source. This near-source removal is a result
of the lack of inherent thermal buoyancy, low release height, and interaction with their
surroundings (impaction and filtration by vegetation).17
Secondary PM is dominated by sulfate particles in the eastern U.S. and parts of the
western U.S., with nitrate particles and carbonaceous particles dominant in some western areas.
Mobile sources can reasonably be estimated to contribute to ambient secondary nitrate and
sulfate PM in proportion to their contribution to total NOx and SOx emissions. (EPA 1998, p.
44-45)
The sources, ambient concentration, and chemical and physical properties of PM10 vary
greatly with time, region, meteorology, and source category. A first step in developing a plan to
attain the PM10 NAAQS is to disaggregate ambient PM10 into the basic categories of sulfate,
nitrate, carbonaceous, and crustal, and then determine the major contributors to each category
based on knowledge of local and upwind emission sources. Following this approach, SIP
strategies to reduce ambient PM concentrations have generally focused on controlling fugitive
dust from natural soil and soil disturbed by human activity, paving dirt roads and controlling of
soil on paved roads, reducing emissions from residential wood combustion, and controlling
major stationary sources of PM10 where applicable. The control programs to reduce stationary,
area, and mobile source emissions of sulfur dioxide, oxides of nitrogen, and volatile organic
compounds in order to achieve attainment with the sulfur dioxide and ozone NAAQS also have
contributed to reductions in the fine fraction of PM10 concentrations. In addition, the EPA
standards for PM emissions from highway and nonroad engines are contributing to reducing
PM10 concentrations. As result of all these efforts, in the last ten years, there has been a
downward trend in PM10 concentrations, with a leveling off in the later years. (EPA 1998, p. 38)
Scientific studies have linked particulate matter (alone or in combination with other air
pollutants) with a series of health effects.18 Coarse particles can accumulate in the respiratory
system and aggravate health problems such as asthma. Particles in motor vehicle exhaust
emissions and the particles formed by the transformation of motor vehicle gaseous emissions
tend to be in the fine particle range. 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. PM also
causes damage to materials and soiling. It is a major cause of substantial visibility impairment in
many parts of the U.S.
These effects are discussed further in EPA's "Staff Paper" and "Air Quality Criteria
HI-62
-------
Chapter III: Environmental Impact
Document" for particulate matter
19
There is additional concern regarding the health effects of PM from diesel vehicles, apart
from the health effects which were considered in setting the NAAQS for PM10 and PM2 5 Diesel
PM contains small quantities of chemical species that are known carcinogens, and diesel PM as a
whole has been implicated in occupational epidemiology studies. We have considered these
studies, and EPA's Office of Research and Development has recently submitted to a committee
of our Science Advisory Board a draft assessment document which contains a proposed
conclusion that diesel exhaust is a highly likely human cancer hazard and is a potential cause of
other nonmalignant respiratory effects. The scientific advisory committee has met to discuss this
document, and we are awaiting written review comments from the committee. We expect to
submit a further revision of the document to the advisory committee before we make the
document final.
b. PM10 Role of Cars and Light Trucks
Section A of this Chapter presents the estimates of PM emissions that were used for this
rulemaking. PM emissions from mobile sources were estimated with newly developed models.
PM emissions from other source categories were those from the most recent National Emissions
Trends inventory process. Estimates of emissions were prepared at the county level, and then
aggregated to higher levels for purposes of presentation.
The contribution of cars and light trucks to ambient PM10 concentrations can be assessed
in simple fashion by comparing the estimates of direct PM10 emissions and of PM10 precursors
emissions. This approach is subject to the uncertainties in those emission estimates, which can
be large for many types of natural and stationary sources, and to uncertainties attributable to
disregard for the locations and temporal patterns of emissions. An alternative approach to
assessing role is to begin with measurements of the quantity and chemical identity of PM10
material collected on ambient filters, and apportion that material to sources based on the
chemical identity of their emissions. The organic portion of PM10 can be analyzed for dozens of
specific tracer compounds, allowing it to be apportioned with more sophistication. It is necessary
to use estimates of inventory shares to apportion a single chemical class which has no tracer back
to source type, for example sulfate PM10.
Because of the current interest in PM2 5 most recent ambient-based source apportionment
studies have focused on PM25. Since virtually all of the PM10 attributable to cars and light
trucks is also PM25 estimates of the mass concentration of PM2 5 due to cars and trucks from
these studies can be treated as estimates of PM10 as well. The 1997 Air Quality Trends Report
presents a summary assessment based on studies across the country. Most of these studies
identified PM composition only down to the nitrate/sulfate/crustal/carbonaceous level, requiring
apportionment within each to be based on emission inventory estimates. In contrast, a recent
HI-63
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
study in Denver made use of detailed analysis of the organic compounds within the carbonaceous
fraction.
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 1997.20 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 ni.C-1.
Table III.C-1. NFRAQS Compositional Analysis of PM25 Samples
Site
Welby
Brighton
Carbon-based
PM7,
49%
42%
Sulfate-based
PM7,
10%
15%
Nitrate-based
PM7,
25%
32%
Crustal Matter
PM7,
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 PM2 5 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 ni.C-2.
Table III.C-2. 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 PM25, as shown in the middle three columns of Table ni.C-3. The
results can be summed to derive the contribution of gasoline vehicles to total PM25, as shown in
the last column in Table ni.C-3.
HI-64
-------
Chapter III: Environmental Impact
Table III.C-3. 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%
These results shown here within each PM2 5 chemical fraction may be typical of urban
areas, while the ratios between the chemical fractions may vary by area. Virtually all direct and
secondary PM from gasoline vehicles is PM25. The percentage contribution of gasoline vehicles
to PM10 concentrations would be lower than for PM2 5, after accounting for the crustal material
PM and coarse-fraction PM from all sources.
A summary of several studies of the PM contribution, including the NFRAQS study just
summarized, is given in Table ni.C-4. The table also shows the researchers' estimate of the
absolute contribution of gasoline vehicles to ambient PM.. This ranges to as high as about 8
//g/m3' The annual PM10 NAAQS is 50 //g/m3 and the 24-hour NAAQS is 150 //g/m3. On a
percentage basis, gasoline vehicles can therefore contribute up to about 5 to 15 percent of the
ambient loading allowed by the PM10 NAAQS. These findings for the most part apply to urban
areas. In rural areas with less vehicle travel and/or better dispersion conditions the contribution
would be less.
Source apportionment studies of the type summarized here have also addressed the
contribution of diesel vehicles to ambient PM, but at the present time virtually all diesel PM is
from vehicle classes other than cars and light trucks. The draft Health Assessment Document for
Diesel Emissions contains a table similar to Table ni.C-4 but for diesel PM contributions to
ambient PM concentrations. This table is reproduced here as Table ni.C-9, and discussed later in
the context of the possibility of increased sales of diesel cars and light trucks.
HI-65
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table III.C-4. Gasoline Vehicle/Engine Contribution to Ambient PM - from source apportionment reports
Author
(Reference)
Friedlander, 1973 21
Gartrell and Friedlander,
1974 22
Schaueretal., 1996
Southern California 23
Lowenthal et al., 1992 24
Wittorff,199425
NFRAQS, 1998
Year of
Sampling, No.
days
1969, 1 day
Sept/Oct 1972
1 day each site
1982, 60 days
(one every sixth
day)
1989, 59 days
(one every sixth
day)
Spring,
1993,
3 days
Winter, 1996-97,
60 days
Location
Type
Urban
Urban
Urban
Suburban
Urban
Urban Bus
Stop
Urban
Suburban
Source Profile Used
EC, OC total, Elements
EC, OC total, Elements,
Major Ions
OC Species, EC,
Elements
EC, OC total, Elements,
Major Ions
EC, OC total, Elements,
Major Ions
OC Species, EC,
Elements, Major ions
Location
Pasadena
Pasadena
Pomona
Riverside
Fresno
San Jose
West LA
Pasadena
Rubidoux
Downtown
LA
Santa Barbara,
CA
Manhattan,
NY
Welby, CO
Brighton, CO
Total PM2.5 (stdev),
^g/m3
nr
64(7)
180 (20)
125 (14)
207 (23)
189(21)
24.5 (2.0)
28.2 (1.9)
42.1 (3.3)
32.5 (2.8)
36.5*
35.8-83.0*
16.7
12.4
GasPM2.5
(stdev),
/jg/m3
8. 2% of aerosol
mass**
5.1 (0.15)
7.2 (0.3)
3.9(0.15)
2.2(0.1)
8.3 (0.33)
1.44(0.16)
1.63 (0.20)
0.34 (0.05)
2.12(0.23)
4.0 (2.2)*
4.2 (avg)
7% of total PM
on average
6.51 (39%)
3.35 (27%)
*PM10, **TSP?
f Not available
nr=not reported
OC: Organic Carbon EC: Elemental Carbon
Major Ions: nitrate, sulfate, chloride and in some cases ammonium, sodium, potassium
HI-66
-------
Chapter III: Environmental Impact
c. Current PM10 Nonattainment
PM10 attainment status for the period 1996-1998 was developed through an analysis of
data obtained from the Aerometric Information Retrieval System (AIRS) on OctoberlS, 1999.
The attainment determination was based on the concepts outlined for the pre-existing National
Ambient Air Quality Standards (NAAQS), and represent an exceedance-based form of the
standard.26 The PM10 NAAQS has a requirement that an area conducting 1 in 6 day sampling
increase sampling frequency to daily frequency upon recording a violation. If this procedure is
not followed, then an "estimate" of the number of exceedances expected is calculated for that
year. Some areas considered nonattainment in this recent analysis have fewer than four actual
monitored 24-hour exceedances. The analysis was not subjected to a review by the EPA
Regional Offices, who have an important role in verifying and validating these data. The
analysis therefore represents an upper bound of the number of areas that could potentially be in
violation of the 1987 PM10 NAAQS.
The most recent PM10 monitoring data indicates that 15 designated PM10 nonattainment
counties, with a population of almost 9 million in 1996, violated the PM10 NAAQS in the period
1996-1998. The areas that are violating do so because of exceedances of the 24-hour PM10
NAAQS. No areas had monitored violations of the annual standard in this period. Table in.C-5
lists the 15 counties. The table also indicates the classification for each area and the status of our
review of the State Implementation Plan. SIP status was obtained from an EPA data base based
on Federal Register actions. Maricopa County, AZ, did have an approved SIP as a moderate
PM10 nonattainment area. It has been reclassified as a serious area, and has not received approval
for serious area SIP requirements.
Although we do not believe that we are limited to considering only designated
nonattainment areas in implementing CAA section 202(i), we have focused on the designated
areas in the case of PM10. An official designation of PM10 nonattainment indicates the existence
of a confirmed PM10 problem that is more than a result of a one-time monitoring upset or a
results of PM10 exceedances attributable to natural events. In addition to these designated
nonattainment areas, there are 15 unclassified counties in 12 geographically spread out states,
with a 1996 population of over 4 million, for which the state has reported PM10 monitoring data
for this period indicating a PM10 NAAQS violation. We have not yet excluded the possibility
that a one-time monitoring upset or a natural event(s) is responsible for the monitored violations
in 1996-1998 in the 15 unclassified counties. We adopted a policy in 1996 that allows areas
whose PM10 exceedances are attributable to natural events to remain unclassified if the state is
taking all reasonable measures to safeguard public health regardless of the source of PM10
emissions. The policy was reiterated after the PM NAAQS were revised.27 Areas that remain
unclassified areas are not required to submit attainment plans, but we work with each of these
areas to understand the nature of the PM10 problem and to determine what best can be done to
reduce it. The Tier 2/Gasoline Sulfur program will reduce PM10 concentrations in these 15
unclassified counties, because all have car and light truck travel that contributes to PM10 and
HI-67
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
precursor emissions loadings. This reduction will assist these areas in reducing their PM10
nonattainment problem, if a problem is confirmed upon closer examination of each local
situation.
Boise, ID, had also been classified as a PM10 nonattainment area at one time and was
monitored to have a PM10 NAAQS violation in 1996-1998. However, the pre-existing PM10
NAAQS does not presently apply in Boise, ID, because in the period between our revision of the
old PM10 NAAQS and the Court's decision to vacate the revised PM10 NAAQS, we determined
that Boise was in attainment with the old PM10 NAAQS and that it therefore no longer applied in
that area.
Table III.C-5. Fifteen PM10 Nonattainment Areas Violating the PM10 NAAQS in 1996-1998
Area
Clark Co., NV
El Paso, TX
Gila, AZ
Imperial Co., CA
Inyo Co., CA
Kern Co., CA
Mono Co., CA
Kings Co., CA
Maricopa Co., AZ
Power Co., ID
Riverside Co., CA
San Bernardino Co., CA
Santa Cruz Co., AZ
Tulare Co., CA
Walla Walla Co., WA
Classification
Serious
Moderate
Moderate
Moderate
Moderate
Serious
Moderate
Serious
Serious
Moderate
Serious
Serious
Moderate
Serious
Moderate
SIP Approved?
No
Yes
No
No
No
No
No
No
No
No
No
No
No
No
Yes
TOTAL POPULATION
1996 Population
(millions)
0.93
0.67
0.05
0.14
0.02
0.62
0.01
0.11
2.61
0.01
1.41
1.59
0.04
0.35
0.05
8.61
d.
Future Nonattainment
Because the types and sources of PM10 are complex and vary from area to area, the best
projections of future PM10 concentrations are the local emission inventory and air quality
modeling analyses that states have developed or are still in the process of developing for their
PM10 attainment plans. We do employ a modeling approach, known as the source-receptor
matrix approach, for relating emission reductions to PM10 reductions on a national scale. This
HI-68
-------
Chapter III: Environmental Impact
approach is one of our established air quality models for purposes of quantifying the health and
welfare related economic benefits of PM reductions from major regulatory actions. One
application of this modeling approach was for the Regulatory Impact Analysis for the
establishment of the new PM NAAQS.28 This model is also the basis for the estimates of PM10
(and PM25) concentrations reductions we have used to estimate the economic benefits of the Tier
2/Gasoline Sulfur program in 2030. Its use for this purpose is described in the final RIA for the
Tier 2/Gasosline Sulfur rule. In both applications, we modeled an emissions scenario
corresponding to controls currently in place or committed to by states. As such, this scenario is
an appropriate baseline for determining if further reductions in emissions are needed in order to
attain and maintain the PM10 NAAQS.
In the RIA for the establishment of the PM NAAQS, we projected that in 2010 there will
be 45 counties not in attainment with the original PM10 NAAQS . We cited these modeling
results in our proposal for the Tier 2/Gasoline Sulfur program and in our first supplemental
notice. After reviewing public comments on our presentation of these modeling results, we have
concluded that while the source-receptor matrix approach is a suitable model for estimating PM
concentration reductions for economic benefits estimation, it is not a tool we can use with high
confidence for predicting that individual areas that are now in attainment will become
nonattainment in the future. However, we believe the source-receptor matrix approach is
appropriate for, and is a suitable tool for, determining that a current designated nonattainment
area has a high risk of remaining in PM10 nonattainment at a future date. Therefore, we have
cross-matched the results for 2030 from our final RIA for Tier 2 and the list of current PM10
nonattainment areas with monitored violations in 1996 to 1998 shown in Table IHC-5.
Based on this modeling, we conclude that the 8 classified nonattainment areas shown in
Table in.C-6 have a high risk of failing to attain and maintain without further emission
reductions. These areas have a population of nearly 8 million. Included in the group are the
counties that are part of the Los Angeles, Phoenix, and Las Vegas metropolitan areas, where
traffic from cars and light trucks is substantial. California areas will benefit from the Tier
2/Gasoline Sulfur program because of travel within California by vehicles originally sold outside
the state, and by reduced poisoning of catalysts from fuel purchased outside of California.
We used the more recent modeling for 2030 rather than the earlier modeling for 2010,
because the modeling for 2030 incorporates more recent estimates of emissions inventories. Our
emission estimates in our final RIA indicate that PM10 emissions under the baseline scenario
increase steadily between 1996 and 2030, for 47 states combined and for four specific cities,
suggesting that areas in nonattainment in both 1996-1998 and 2030 will be in nonattainment in
the intermediate years as well assuming no further emission reductions. A factor tending to make
Table in.C-6 shorter is that we have not relied on the source-receptor matrix model's prediction
of 24-hour nonattainment, as those predictions on an individual area basis are less reliable than
the predictions of annual average nonattainment.
HI-69
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table III.C-6. Eight areas with a high risk of failing to attain and maintain
the PM10 NAAQS without further reductions in emissions.
Area 1996 Population
(millions)
Clark Co., NV 0.93
Imperial Co., CA 0.14
Kern Co., CA 0.62
Kings Co., CA 0.11
Maricopa Co., AZ 2.61
Riverside Co., CA 1.41
San Bernardino Co., CA 1.59
Tulare Co., CA 0.35
TOTAL POPULATION 7.76
Taken together and considering their number, size, and geographic distribution,
these 8 areas are sufficient to establish the case that additional reductions are needed in order to
attain and maintain the PM10 NAAQS. This determination provides additional support for the
NOx and VOC standards and for the limits on gasoline sulfur, which are also fully supported on
ozone attainment and health effects considerations. The sulfate particulate, sulfur dioxide, NOx,
and VOC emission reductions from the Tier 2/Gasoline Sulfur program will help the 8 areas in
Table IHC-2 to attain and maintain the PM10 NAAQS. The new PM standards for gasoline and
diesel vehicles is also supported by this PM10 determination. We are also establishing the new
PM standard today to avoid the possibility that PM10 concentrations in these and other areas do
not actually get worse due to an increase in sales of diesel vehicles, which could create a need for
further reductions which would be larger and would affect more areas of the country.
Table in.C-6 is limited to designated PM10 nonattainment areas which both had
monitored violations of the PM10 NAAQs in 1996-1998 and are predicted to be in nonattainment
in 2030 in our PM10 air quality modeling. This gives us high confidence that these areas require
further emission reductions to attain and maintain, but does not fully consider the possibility that
there are other areas which are now meeting the PM10 NAAQS which have at least a significant
probability of requiring further reductions to continue to maintain it. Our air quality modeling
predicted 2030 violations of the annual average PM10 NAAQS in five additional counties that in
either 1997 or 1998 had single-year annual average monitored PM10 levels of at least 90 percent
of the NAAQS, but did not exceed the formal definition of the NAAQS over the three-year
period ending in 1998. These areas were identified from the most recent published EPA Air
Quality Trends Report (USEPA 1998) and from tables that have been prepared for inclusion in
the next update of the Air Quality Trends report.29 In two of these areas, New York Co., NY
and Harris Co., TX, the average PM10 level in 1998 was above the 50 //g/m3 value of the
HI-70
-------
Chapter III: Environmental Impact
NAAQS. These two areas are not included in the Table in.C-6 list of 15 areas with a high risk of
failing to attain and maintain because lower PM10 levels in 1996 and 1997 caused their three-year
average PM10 level to be lower than the NAAQS.
These five areas were listed based on their second high 24-hour concentration and annual
average concentration in 1997 or 1998 only. Actual nonattainment determinations are made
based on three years of data, and on estimates of expected exceedances of the 24-hour standard.
The second-high 24-hour PM10 level in 1998 is only an approximate surrogate for the outcome of
an expected exceedances calculation since other years of data and the frequency of monitor
operation would also enter the calculation.
The five additional counties are shown in Table in.C-7. They have a combined
population of almost 17 million, and a broad geographic spread. Unlike the situation for ozone,
for which precursor emissions are generally declining over the next 10 years or so before
beginning to increase, we estimate that emissions of PM10 will rise steadily unless new controls
are implemented. The small margin of attainment which these areas currently enjoy will likely
erode; the PM air quality modeling suggests that it will be reversed. We therefore consider these
areas to each individually have a significant risk of failing to maintain the NAAQS without
further emission reductions. There is a substantial risk that at least some of them would fail to
maintain without further emission reductions. The emission reductions from the Tier 2/Gasoline
Sulfur program will help to keep them in attainment.
Table III.C-7. Five areas with a significant risk of failing to attain and maintain
the PM10 NAAQS without further reductions in emissions
Area 1996 Population
(millions)
New York Co., NY
Cuyahoga Co., OH
Harris, Co., TX
San Diego Co., CA
Los Angeles Co., CA
1.33
1.39
3.10
2.67
8.11
TOTAL POPULATION 16.6
In addition to the counties already listed in Tables in.C-5 and IHC-7, there are other
areas for which 1997 and 1998 data indicate that maintenance of the PM10 NAAQS is at risk,
particularly if diesel sales of cars and light truck increase as discussed below. Table ni.C-8 lists
additional counties for which either 1997 or 1998 monitoring data, or both, indicated a second-
high PM10 concentration for the single year within 10 percent of the PM10 24-hour NAAQS or an
annual average PM10 concentration within 10 percent of the annual average PM10 NAAQS. Our
HI-71
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Source-Receptor Matrix modeling of annual average PM10 concentrations in these areas did not
indicate nonattainment in 2030, but the margin of attainment they currently enjoy is small. Only
counties which are part of metropolitan statistical areas are listed in Table IHC-8, in order to
focus on those in which traffic densities are high. Considering both the annual and 24-hour
NAAQS, there were 13 areas within 10 percent of the standard.
Table III.C-8. Thirteen metropolitan statistical area counties with 1997 and/or 1998
ambient PM10 concentrations within 10 percent of the annual or 24-hour the PM10
NAAQS.*
1996 Population
(millions)
Areas within 10 percent of the annual PM10 NAAQS
Lexington Co., SC 0.20
Union Co., TN 0.02
Washoe Co., NV 0.30
Madison Co., IL 0.26
Dona Ana Co., NM 0.16
El Paso Co., TX 0.68
Ellis Co., TX 0.97
Fresno Co., CA 0.74
Philadelphia Co., PA 1.47
Areas within 10 percent of the 24-hour PM10 NAAQS
Lexington Co., SC 0.20
El Paso Co., TX 0.68
Union Co., TN 0.02
Mobile Co., AL 0.40
Dona Ana Co., NM 0.16
Lake Co., IN 0.48
Philadelphia Co., PA 1.47
Pennington Co., SD 0.09
Ventura Co., CA 0.71
TOTAL POPULATION OF ALL 13 AREAS 6.48
* These areas are listed based on their second high 24-hour concentration and annual average concentration in
1997, 1998, or both. Official nonattainment determinations are made based on three years of data, and on estimates
of expected exceedances of the 24-hour standard.
HI-72
-------
Chapter III: Environmental Impact
e. Diesel PM
At the present time, virtually all cars and light trucks being sold are gasoline fueled. The
ambient PM10 air quality data for 1996 to 1998 reflects that current situation, and the predictions
of future PM10 air quality are based on an assumption that this will continue to be true. However,
we are concerned over the possibility that diesels will become more prevalent in the car and
light-duty truck fleet, since automotive companies have announced their desire to increase their
sales of diesel cars and light trucks. As current diesel vehicles emit higher levels of PM10 than
gasoline vehicles, a larger number of diesel vehicles could dramatically increase levels of exhaust
PM10, especially if more stringent standards are not in place. The new PM standards will ensure
that an increase in the sales of diesel cars and light trucks will not increase PM emissions from
cars and light trucks so substantially as to endanger PM10 attainment and maintenance on a more
widespread basis. Given this potential, it is appropriate to establish the new PM standards now
on the basis of the increase in sales of diesel vehicles being a reasonable possibility without such
standards. Establishing the new PM standards now avoids the public health impact and industry
disruption that could result if we waited until an increase in sales of diesels with high PM
emissions had already occurred.
In order to assess the potential impact of increased diesel sales penetration on PM
emissions, we analyzed the increase in PM10 emissions from cars and trucks under a scenario in
which the use of diesel engines in cars and light trucks increases. We used projections developed
by A.D. Little, Inc. as part of a study conducted for the American Petroleum Institute. The "Most
Likely" case projected by A.D. Little forecasts that diesel engines' share of the light truck market
will grow to 24 percent by the 2015 model year. Diesel engines' share of the car market would
grow somewhat more slowly, reaching 9 percent by 2015. The A.D. Little forecasts did not
address the period after 2015; we have assumed that diesel sales stabilize at the level reached in
2015, with the fraction of in-use vehicles with diesel engines continuing to increase through
turnover. We believe these projections are more realistic than the scenario of even higher sales
of diesels described in the notice for the proposed Tier 2/Gasoline Sulfur program, though the
A.D. Little forecasts still show much higher percentages of diesel vehicles in the light-duty fleet
than have ever existed historically in the U.S.
The A.D. Little "most likely" scenario of increased diesels would result in dramatic
increases in direct PM10 emissions from cars and light trucks, if there were no change in these
vehicles' PM standards. The increase in diesel exhaust PM10 emissions would more than
overcome the reduction in direct PM10 attributable to the sulfur reduction in gasoline. With no
change in the existing PM standards for cars and light trucks, our analysis of this scenario shows
that direct PM10 emissions in 2020 would be approximately 98,000 tons per year, which is nearly
two times the 50,000 tons projected if diesel sales do not increase. The increase in diesel PM
would be somewhat larger than the 48,000 ton difference between these two values, since the
HI-73
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
difference also reflects the lower amount of PM from gasoline vehicles due to fewer gasoline
vehicles in operation. The portion of ambient PM10 concentrations attributable to cars and light
trucks would climb steadily.
The added PM10 emissions from cars and trucks due to an increase in diesel sales without
action to reduce PM10 from each diesel vehicle would exacerbate the PM10 nonattainment
problems of the areas listed in Tables ni.C-6 and IHC-7, for which our air quality modeling
predicted future nonattainment even without an increase in diesel sales. Moreover, it might
cause PM10 nonattainment in the additional areas listed in Table IHC-8. Increases in PM10
emissions from more diesel vehicles would put these areas in greater risk of violating the PM10
NAAQS, especially if growth in other sources is high or meteorological conditions are more
adverse than in the 1996 to 1998 period.
We have considered two approaches to quantifying the increased risk of PM10
nonattainment that could results from the additional PM emissions from more diesel cars and
trucks. First, Section D of this chapter presents estimates of the population exposure to diesel
PM that would result from higher diesel sales, with and without the more stringent PM emission
standards. Table in.D-9 and the text preceding it indicate that increased sales of cars and light
trucks without a more stringent PM emission standard could increase personal exposure to diesel
PM by about 0.2 |ig/m3. The exposure estimates in Section D estimates are based on an approach
in which estimated personal exposures to carbon monoxide are adjusted based on ratios of
emission inventories. They therefore incorporate the effect of personal movement among micro
environments with more or less direct exposure to emissions and ambient air, and relate only
indirectly to the monitored PM10 levels that determine attainment or nonattainment.
Our second approach to making a rough estimate of the possible contribution of diesel
cars and light trucks to ambient PM10, assuming the PM emission standard were not being made
more stringent, was to first estimate the contribution of heavy-duty diesel vehicle PM to ambient
PM10 for a historical period, and then adjust this by a PM emission inventory ratio of the quantity
of diesel emissions associated with the historical contribution to ambient PM and the quantity
that would be emitted from diesel cars and light trucks on a future date if sales of diesel cars and
trucks did increase. The following paragraphs present an analysis along these lines.
The draft Health Assessment Document for Diesel Emissions includes a summary of the
diesel PM findings of several source apportionment studies. The most commonly used receptor
model for quantifying concentrations of diesel PM at a receptor site is the chemical mass balance
model (CMB). Input to the CMB model includes PM measurements made at the receptor site as
well as measurements made of each of the source types suspected to impact the site. Due to
problems involving the elemental similarity between diesel and gasoline emission profiles and
their co-emission in time and space, it is necessary to carefully quantify chemical molecular
species which provide markers for separation of these sources (Lowenthal et al., 1992). Recent
advances in chemical analytical techniques have facilitated the development of sophisticated
HI-74
-------
Chapter III: Environmental Impact
molecular source profiles including detailed speciation of organic compounds which allow the
apportionment of PM to gasoline and diesel sources with increased certainty. Older studies which
made use of only elemental source profiles have been published and are summarized here, but are
subject to more uncertainty.
The CMB model has been used to assess the contribution of diesel PM to total PM mass
in areas of California, Denver, CO, Phoenix, AZ, and Manhattan, NY (Table ni.C-9). Diesel
PM concentrations reported by Schauer et al. (1996) for data collected in 1982 ranged from 4.4
//g/m3 in west Los Angeles to 11.6 //g/m3 in Downtown Los Angeles. The average contribution
of diesel PM to total PM mass ranged from 13 percent in Rubidoux to 36 percent in downtown
Los Angeles. It should be noted that this model accounts for primary emissions of diesel PM
only and the contribution of secondary aerosol formation (both acid and organic aerosols) is not
included. In sites downwind from urban areas, such as Rubidoux in this study, secondary nitrate
formation can account for a substantial fraction of the mass (25 percent of the fine mass
measured in Rubidoux was attributed to secondary nitrate), a portion of which comes from diesel
exhaust.30
HI-75
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table III.C-9. Ambient Diesel PM Concentrations Reported from Chemical Mass Balance Modeling
Author
Schauer et
al, 1996
Southern
California
Chow et
al., 1991
California
EPA, 1998
Wittorff,
1994
NFRAQS,
1998
Year of
Sampling, No.
days
1982, 60 days
(one every
sixth day)
Winter, 1989-
90 f
1988-92,
approx. 150
days
Spring,
1993,
3 days
Winter, 1996-
97, 60 days
Location
West LA
Pasadena
Rubidoux
Downtown LA
Phoenix, AZ
Area
15 Air Basins
Manhattan, NY
Welby, CO
Brighton, CO
Location
Type
Urban
Urban
Suburban
Urban
Urban
Rural-Urban
Urban Bus
Stop
Urban
Suburban
Source Profile
Used
OC Species, EC,
Elements
t
EC, OC total,
Elements, Major
Ions
EC, OC total,
Elements, Major
Ions
OC Species, EC,
Elements, Major
ions
TotalPM2.5
(stdev), ,ug/m3
24.5 (2.0)
28.2 (1.9)
42.1 (3.3)
32.5 (2.8)
t
t
35.8-83.0
16.7
12.4
Diesel PM2. 5
(stdev),
^/m3
4.4 (0.6)
5.3 (0.7)
5.4 (0.5)
11.6(1.2)
4-22*
0.2-3.6*
13.2-46.7*
1.7
1.2
*PM10
f Not available
OC: Organic Carbon EC: Elemental Carbon
Major Ions: nitrate, sulfate, chloride and in some cases ammonium, sodium, potassium
HI-76
-------
Chapter III: Environmental Impact
A wintertime study conducted in the Phoenix, AZ area by Chow et al.,31 indicated that
diesel PM levels on single days can range from 4 //g/m3 in west, and central Phoenix, to 14
//g/m3 in south Scottsdale and 22 //g/m3 in central Phoenix. This apportionment, like the Schauer
et al., (1996) data, reflects direct emissions only. These data relied on source profiles and
ambient data collected prior to the introduction of technology to reduce PM emissions from
diesel-powered vehicles. This study has not appeared in the peer reviewed literature.
A second CMB study reported ambient diesel PM concentrations for California and used
ambient measurements from the San Joaquin Valley (1988-89), South Coast (1986), and San
Jose (winters for 1991-92 and 1992-93).32 The incorporation of sampling data from later dates
provides information regarding exposures more relevant to current levels. The CMB in the
California study (1998a) indicated that on an annual basis, basin-wide levels of direct diesel PM
emissions may be as low as 0.2 //g/m3 in the Great Basin Valleys and as high as 3.6 //g/m3 in the
South Coast basin.
The most recent study reporting diesel PM concentrations is from winter 1996-1997
sampling conducted in the Denver, CO area as part of the Northern Front Range Air Quality
Study (NFRAQS, 1998). Ambient levels of diesel PM in the urban core site at Welby averaged
1.7 //g/m3 over a 60-day winter period and a slightly lower average concentration of 1.2 //g/m3
was measured at an urban downwind site in Brighton, CO. One of the major findings from this
study was a substantial contribution of elemental carbon from gasoline-powered vehicles. At the
Welby site, the contribution of diesel and gasoline emissions to elemental carbon measurements
was 52 percent and 42 percent, respectively. At the Brighton site, the contribution of diesel and
gasoline emissions to elemental carbon measurements was 71 percent and 26 percent,
respectively. The findings from the NFRAQS study are compelling and suggest the need for
further investigations of this type that specifically address high-emitting vehicles. Geographical
and other site-specific parameters that influence PM concentrations, such as altitude, must be
considered when extrapolating the NFRAQS findings to other locations
Limited data are available which allow a characterization of diesel PM concentrations in
'hot spots' such as near heavily traveled roadways, bus stations, train stations, and marinas. One
'hotspof study conducted in Manhattan, NY reported diesel PM concentrations of 13.0 to 46.7
//g/m3 during a three-day sampling period in the spring of 1993 (Wittorff,1994). This study
attributed, on average, 50 percent of the PM to diesel exhaust. The diesel PM concentrations
resulting from the source apportionment method used in this study require some caution. The
CMB model overpredicted PM10 concentrations by an average 30 percent, suggesting that
additional sources of the mass were not accounted for in the model. New advances in organic
carbon speciation, as has been noted above, are necessary to most appropriately characterize
gasoline and diesel PM sources to ambient PM measurements. The relevance of the Manhattan
bus stop exposure for large urban populations provides strong motivation for further studies in
the vicinity of such 'hotspots'. This study has not appeared in the peer reviewed literature.
HI-77
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
In summary, recent source apportionment studies (California EPA, 1998; NFRAQS,
1998) indicate that ambient diesel PM concentrations averaged over 2-12 month periods for
urban/suburban areas can range from approximately 1.2 //g/m3 to 3.6 //g/m3, while diesel PM
concentrations in more rural/remote areas are generally less than 1.0 //g/m3. In the vicinity of
'hot spots', or for short exposure times under episode-type conditions diesel PM concentrations
are expected to be substantially higher than these levels, as high as 22 and 47 //g/m3. However, a
thorough and replicated characterization of these situations is not yet available. Two studies
nearing completion by the South Coast Air Quality Management District will shed some light on
near-highway concentrations of diesel PM.33
To quantify the potential contribution that diesel cars and light trucks may make to
ambient PM10 concentrations, we have used the four numbers cited in the previous paragraph
(1.2, 3.6, 22, and 47 //g/m3) as starting points, recognizing that the latter two values have
considerable uncertainty for reasons explained above. We assume that these estimated ambient
concentrations are attributable to highway diesel vehicles, and specifically to heavy-duty vehicles
because of the near-zero use of diesel engines in other classes in the time frames of these
studies.21
Table in.C-10 summarizes this analysis, combining key findings from the studies just
summaried with emissions estimates from the Tier 2 analysis, to predict increases in PM10 in
2030. A needed detail from the 47-state inventories described in Section A of this chapter is that
47-state PM emissions from all diesel vehicles in 1996 (the year of NFRAQS) was 160,109
tons. This figure must be adjusted to the time frame for each of the other three estimates of
ambient diesel PM concentration. We have done this using the emissions trends given in the
1997 Emission Trends report, applying the ratio of two calendar year's 50-state emissions to the
47-state estimate for 1996.34 The 2030 PM emissions from diesel cars and light trucks under the
scenario of higher diesel car and light truck sales without more stringent PM emission standards
would be 77,421 tons.
The final columns of Table in.C-10 suggest that with higher sales of diesel cars and light
trucks, they could contribute between 0.6 to 20 //g/m3 to PM10 concentrations. This would
represent between one-half and 40 percent of the PM10 concentration allowed by the NAAQS,
with the upper end of this range based on studies conducted in roadside situations with heavy
traffic and using older and more simple approaches to source apportionment.
21 Non-road diesel engines also operate in urban areas, and may actually have contributed somewhat to the
ambient concentrations observed in the various studies. To the extent that they did, the estimates of the possible
future contribution to ambient PM from diesel cars and light trucks would be overestimates.
HI-78
-------
Chapter III: Environmental Impact
Table IILC-10. Estimation of Potential Contribution of Diesel Cars and Light Trucks to Ambient PM10 Concentrations in
2030
Study Used as
Star ting Point
NFRAQS, 1998
California EPA,
1998
Chowetal., 1991
Wittorff, 1994
Ambient
PM10 from
Diesels in
Study
/;g/m3
1.2
3.6
22
47
Assumed
Year of
Study
1996
1990
1990
1993
Highway
Heavy-Duty
Diesel
PM10
Emissions
at Time of
Study
(47 -state
annual
tons) *
160,109
213,479
213,479
181,175
2030 Light-Duty
Diesel PM10
Emissions
(4 7 -state annual
tons)**
77,421
77,421
77,421
77,421
Estimate of 2030
Ambient PM 10
from Light-Duty
Diesels
(//g/m3)***
0.6
1.3
8.0
20.1
Percentage of
24-hour
PM10
NAAQS
0.4 %
0.9 %
5.3 %
13.4%
Percentage
of Annual
PM10
NAAQS
1.2%
1.8%
16.0%
40.2 %
* Interpolated from EPA Emissions Trends Report.
** Derived from estimates in Section A of this chapter.
*** Concentration estimate from study multiplied by ratio of emissions estimates.
HI-79
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
The standards included in today's actions will result in a steady decrease in total direct
PM10 from cars and light trucks even if this increase in the use of diesel engines in these vehicles
were to occur. If the A.D. Little scenario for increased diesel engines in light trucks were to
occur, today's actions would reduce diesel PM10 from cars and light trucks by over 75 percent in
2020. Stated differently, by 2030 today's actions would reduce over 93,000 tons of the potential
increase in PM10 emissions from passenger cars and light trucks. The result would be less direct
PM10 than is emitted today, because the increase in diesel PM10 would be more than offset by the
reduction in gasoline PM10.
It should be noted that the analysis of the economic benefits of the Tier 2/Gasoline Sulfur
program does not include any effects related to the possible increase in sales of diesel cars and
light trucks or the to the control of the PM increase that would otherwise occur if the PM
standard were not being revised.
Fortunately, the standards included in today's actions will result in a steady decrease in
total direct PM10 from cars and light trucks even if this increase in the use of diesel engines in
these vehicles were to occur. If the A.D. Little scenario for increased diesel engines in light
trucks were to occur, today's actions would reduce diesel PM10 from cars and light trucks by over
75 percent in 2020. Stated differently, by 2030 today's actions would reduce over 93,000 tons of
the potential increase in PM10 emissions from passenger cars and light trucks. The result would
be less direct PM10 than is emitted today, because the increase in diesel PM10 would be more than
offset by the reduction in gasoline PM10.
We are establishing tighter PM standards for diesel vehicles because of the impact greater
diesel PM emissions would have on PM10 attainment and public health and welfare if diesel sales
increased in the future without the protection of the tighter standards. Because diesel vehicles
will essentially be performing the same functions as the gasoline vehicles they will replace, it is
appropriate for the new PM standards to also apply equally to gasoline and diesel vehicles. We
expect that gasoline vehicles will need little or no redesign to meet the new PM standards when
free of defects and properly operating. However, the new standards may achieve some reduction
in real world PM emissions from gasoline vehicles by encouraging more durable designs. The
new standards for PM will also prevent any changes in gasoline engine design which would
increase PM emissions. These changes would otherwise be possible because of the current PM
standard is so much higher than the current performance on the gasoline vehicles.
f. Reductions In Ambient PM
In general, we project that the Tier 2/Gasoline Sulfur program will reduce both direct and
secondary PM from cars and light trucks substantially, regardless of the future market share for
diesel engines in the light-duty fleet. The larger part of the reduction is due to large reductions in
VOC, NOx, and SOx emissions, with corresponding reductions in secondary PM formation.
HI-80
-------
Chapter III: Environmental Impact
Low sulfur fuel will greatly reduce direct PM emissions and sulfate-based secondary PM
formation from SOx emissions from gasoline vehicles, while tailpipe PM standards are projected
to mitigate excess PM emissions from diesel vehicles, even at very aggressive rates of diesel
vehicle sales growth. Substantial reductions in NOx emissions will carry over to reductions in
indirect PM. These reductions will help reduce the number of areas with PM10 and PM2 5 levels
in excess of national standards, reduce the severity of PM nonattainment in other areas, and help
areas facing PM maintenance challenges stay in attainment.
The magnitude of the PM reductions from today's actions in a given area depends on
conditions such as the contribution of light-duty vehicles to the local PM, SOx, NOx, and VOC
inventory; the contribution of light-duty vehicles to the PM, SOx, NOx, and VOC inventories in
upwind areas; local and upwind ammonia inventories (involved in secondary PM formation);
control measures being implemented on both local and upwind sources of PM and its precursors,
and local meteorology. We have incorporated these factors into the air quality modeling used to
develop the benefit/cost analysis presented in Chapter Vn, which includes the economic benefits
of the direct and secondary PM reductions expected to result from today's actions. The estimates
of annual average ambient PM2 5 reductions with full program implementation and phase-in in
2030 are presented in a contractor report that was part of the benefit/cost analysis. Reductions
are given as an average for each state, and in the form of a shaded map. Reductions are larger in
urban areas than rural, and larger in the east than in the sparsely populated areas of the west. The
PM25 reductions and PM10 reductions are essentially equal. State-wide average ambient PM
reductions range from 0.04 //g/m3 in Nevada, reflecting its large areas with low vehicle travel, to
0.45 //g/m3 in Washington, DC. Much of the eastern half of the U.S. is estimated to have a
reduction of at least 0.20 //g/m3, with the largest reduction in any county being 1.25 //g/m3.
2. Visibility/Regional Haze
Visibility is greatly affected by ambient PM2 5 concentration, with PM2 5 concentrations
below the NAAQS being sufficient to impair visibility. The reductions in ambient PM2 5 from the
Tier 2/Gasoline Sulfur program will contribute to visibility improvements across the U.S. The
geographical pattern of the improvement mirrors that of the PM25 reductions. Visibility
improvements have value to Americans in both recreational areas traditionally known for scenic
vistas, and in the urban areas where people spend most of their time.
The Grand Canyon Visibility Transport Commission examined visibility impairment on
the Colorado Plateau. Figures II-4 and II-5 in the Commission's contain estimates for the
contribution of 11 different sources to the man-made visibility impairment at Hopi Point35.
Figure II-4 is for annual average light extinction22 and Figure II-5 for the worst days. Each
22 Light extinction is a measure of visibility impairment.
ni-8i
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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. Emissions inventory estimates
generally suggest a greater role for road dust than is suggested by studies of ambient PM. This
discrepancy is thought to result from the generally shorter atmospheric residence time for road
dust due to its greater size and lack of inherent thermal buoyancy, as well as its greater filtration
by vegetation, than other sources of PM. Therefore, the contribution of road dust may be
overstated in the estimates described in the preceding paragraph. 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 the reductions in highway vehicle
emissions from the Tier 2/Gasoline Sulfur program can contribute significantly to improved
visibility on the Colorado Plateau.
The economic benefits analysis reported in Chapter Vn included modeling to determine
the degree of visibility improvement, and estimated the economic value of visibility
improvements in both recreational and residential settings.
D. Air Toxics
This section summarizes our analysis of the impact of the final Tier 2 standards on
emissions of and exposure to air toxics. Section D. 1 reviews the effects of selected air toxics
emissions on human health. Section D.2 describes our analysis of air toxics emissions and
exposure and the effect that the proposed Diesel 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
potency estimates that have or could have significant emissions from light-duty as well as heavy-
duty vehicles: 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. 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 methodology used to develop these estimates has recently undergone peer
review.
m-82
-------
Chapter III: Environmental Impact
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.36 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 nonlymphocytic23 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.3738 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
animals39 and increased proliferation of mouse bone marrow cells40. 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.41
The latest assessment by EPA places the excess risk of developing acute nonlymphocytic
leukemia at 2.2 x lO'6 to 7.7 x !CT6/|ig/m3 (EPA, 1998a). 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).
23 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.
HI-83
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 1991,42). 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 pancytopenia24 ,a condition characterized
by decreased numbers of circulating erythrocytes (red blood cells), leukocytes (white blood
cells), and thrombocytes (blood platelets).43'44 Individuals that develop pancytopenia and have
continued exposure to benzene may develop aplastic anemia,25 whereas others exhibit both
pancytopenia and bone marrow hyperplasia (excessive cell formation), a condition that may
indicate a preleukemic state.45'46The most sensitive noncancer effect observed in humans is the
depression of absolute lymphocyte counts in the circulating blood47. 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 draft 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.48 This RfC
is currently being revised.
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
TOG, depending on control technology and fuel composition.
1-3-Butadiene was classified by EPA as a Group B2 (probable human) carcinogen in
24 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.
25 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,
orincreased synthesis of blood cell elements, followed by hypoplasia, or decreased synthesis. As the disease progresses,
thebone marrow decreases functioning. This myeloplastic dysplasia (formation of abnormal tissue) without acute
leukemiais known as preleukemia. The aplastic anemia can progress to AML (acute mylogenous leukemia).
m-84
-------
Chapter III: Environmental Impact
1985.49 This classification was based on evidence from two species of rodents and ***
epidemiologic data. EPA recently prepared a draft assessment that would determine sufficient
evidence exists to propose that 1,3-butadiene be classified as a known human carcinogen.50
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.51 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.
The unit risk estimates presented in EPA's draft risk assessment were not accepted by the
SAB. The SAB panel recommended that EPA recalculate the lifetime cancer risk estimates
based on the human data from Delzell et al. 199552 and revise EPA's original calculations to
account for the highest exposure of "360 ppm-year" instead of "250+ ppm-year"and 70 years at
risk instead of 85 years. Based on these recalculations53 the MLE estimate of lifetime cancer
risk from continuous 1,3-butadiene exposure is 2.21 x 10"6/|ig/m3. This estimate implies that
approximately 2 persons in one million exposed to 1 |ig/m3 1,3-butadiene continuously for their
lifetime (70 years in this case) would develop cancer as a result of their exposure.
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.54 It is present in
emissions 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
HI-85
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 carcinogen55 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 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 this 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.56 In persons with bronchial asthma,
the upper respiratory irritation caused by formaldehyde can precipitate an acute asthmatic attack,
sometimes at concentrations below 5 ppm;57 formaldehyde exposure may also cause bronchial
asthma-like symptoms in nonasthmatics.58'59 However, it is unclear whether asthmatics are more
sensitive than nonasthmatics to formaldehyde's effects.60
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
HI-86
-------
Chapter III: Environmental Impact
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., 1991,61 ). 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
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.62'63'64 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.65
e. Diesel Particulate Matter
Diesel exhaust includes components in the gas and particle phases. 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 areas of the
human lung. Gaseous components of diesel exhaust include nitrates, sulfur compounds, organic
compounds, carbon monoxide, carbon dioxide, water vapor, and excess air (nitrogen and
oxygen). Among these gas-phase constituents, the components suspected to have carcinogenic
potential are the organic compounds (including benzene, formaldehyde, acetaldehyde, 1,3-
HI-87
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
butadiene). Current studies have not been designed to specifically discern a role for the gaseous
compounds in contributing to diesel exhaust carcinogenicity, but it is likely that some portion can
be attributed to gas phase organics.
While some of the cancer risk is likely associated with exposure to the gaseous
components of diesel exhaust, studies conducted suggest that the particulate component plays a
substantial role in carcinogenicity. Information that is currently available suggests that the
particulate fraction of diesel exhaust is carcinogenic independent of the gaseous component.
Specifically, (1) diesel particles (the elemental carbon core plus the adsorbed organics) induce
lung cancer at high doses, and the particles, independent of the gaseous compounds, elicit an
animal lung cancer response; (2) the presence of elemental carbon particles as well as the
organic-laden diesel particles correlate with an adverse inflammatory effect in the respiratory
system of animals, with some limited evidence in humans; (3) the extractible particle organics
taken collectively produce cancer and adverse mutagenic toxicity in experimental test systems;
and (4) many of the individual organic compounds adsorbed onto the particles are mutagenic or
carcinogenic in their own right (EPA, 1999b). This information suggests that the particle may be
playing a dual role in contributing to the carcinogenicity of diesel exhaust: both as a mechanism
of delivery for many of the organics into the respiratory system, and the role of the elemental
carbon core.
In two human studies on railroad workers and one on Teamster Union Truck Drivers and
attendant personnel who were occupationally exposed to diesel exhaust (EPA 1999b), it was
observed that long-term inhalation of diesel exhaust produced an excess risk of lung cancer.
Taken together, these and other human studies show a positive association between diesel
exhaust exposure and lung cancer. While some uncertainties remain about confounding from
smoking and possible coincident exposures to other agents, the totality of human evidence
provides a strong inference for a human lung cancer hazard. Inhalation studies in rats show a
lung cancer response at high doses, though the rat model is not deemed a satisfactory test system
for indicating a low exposure hazard for humans. Results from inhalation studies in mice range
from equivocal to suggestive but are not compelling (EPA 1999b). Lung implantation animal
studies do show the carcinogenic reactivity of diesel particulates and the extracted organics.
Extensive mutagenicity and genotoxicity studies show that the particle organics and the gaseous
fractions are reactive and it is also evident that a number of the organic constituents present on
the particles and in the gases are carcinogenic in their own right, though not necessarily in the
lung (EPA 1999b). EPA's draft Diesel Health Assessment also identifies several types of
adverse chronic respiratory effects including respiratory tract irritation/inflammation, changes in
lung function, and a suggestion of adverse immunological changes as concerns for long term
exposure to diesel exhaust. The evidence for chronic respiratory effects comes mainly from
animal studies (the rat being the most studied), given the limited availability of human studies.
The evidence for both cancer and chronic respiratory effects comes from the studies involving
occupational exposures and or high exposure animal studies. The Agency's draft assessment
(EPA, 1999b) stated that diesel exhaust is a highly likely human lung cancer hazard, but that the
HI-88
-------
Chapter III: Environmental Impact
data are currently unsuitable to make a confident quantitative statement of risk. The draft
assessment also states that this risk is applicable to ambient exposures and that the risk may be in
the range of regulatory interest (greater than one in a million over a lifetime). In addition, EPA
believes that keeping long term exposures to diesel particulate matter at or below 5 |ig/m3
provides an adequate margin of safety for the noncancer chronic respiratory hazards.66
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, ARB's
evaluation. Based on human epidemiology studies, the ARB's estimate of the range of a lifetime
upper confidence limit unit cancer risk from continuous diesel exhaust particulate exposure
ranges from 1.3 x 10"4 to 2.4 x 10"3 |ig/m3 (lifetime- jig/m3)"1. The geometric mean unit risk
obtained from these end points of the range is 6 x 10"4 (lifetime- jig/m3)"1. In other words, it is
estimated that approximately 130 to 2400 persons in one million Californians exposed to 1 |ig/m3
diesel exhaust particulate continuously for their lifetime (70 years) would develop cancer as a
result of their exposure.
Particulates (i.e, particulate matter, PM) are a prominent part of diesel exhaust and play a
role in contributing to total ambient PM, especially PM2 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, though the health concerns from exposure ambient PM vs diesel
have differences as well as similarities. As diesel particulates make up more and more of the
ambient PM mixture, the health concerns would overlap to a larger extent. Compared to a
typical ambient PM mixture from many sources, diesel exhaust particles probably have a higher
percentage of small particles which also have a higher surface area laden with adsorbed organics.
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 information67'68 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 nine urban areas and the
results were extrapolated nationwide. Results from these analyses were summarized in the draft
regulatory impact analysis for the proposal for this rulemaking. These analyses have been
updated and extended for the final rule.69 First, additional areas were modeled to encompass a
HI-89
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
broader selection of I/M programs, fuel parameters, and temperature regimes. This enabled us to
develop more accurate nationwide extrapolations. Second, model inputs were revised to reflect
more recent information on emission rates, and to reflect the standards being promulgated in the
final rule. As mentioned previously, EPA has assessed emissions and exposure from the
following air toxics: benzene, formaldehyde, acetaldehyde, 1,3-butadiene, and diesel paniculate
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 for the final rule. Subsection
D.2.a. discusses the emission modeling conducted for mobile source gaseous air toxics
(including both exhaust and nonexhaust air toxics) and diesel PM. Subsection D.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 D.2.c.
describes our analysis of air toxics exposure for our baseline scenario. Subsection D.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 cars and 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 total organic gas (TOG) estimates. The TOG basic emission
rates used in this modeling incorporated the available elements for MOBILE6 used to develop
the VOC inventory for this rule. 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.
Toxic emissions were modeled for 10 urban areas and 16 geographic regions. These
urban areas and geographic regions are presented in Table in.D-1. They were selected to
encompass a broad range of I/M programs, fuel parameters, and temperature regimes. The intent
of the selection process was to best characterize the different combinations needed to perform
accurate nationwide toxic emissions estimates. Each U. S. county was then mapped to a modeled
area or region. This approach was also used to develop the inventory estimates in the 1996
National Toxics Inventory.
m-90
-------
Chapter III: Environmental Impact
Modeling for these areas was done on a seasonal basis. Information on fuel properties for
1990 and 1996 was obtained from surveys conducted by the National Institute for Petroleum and
Energy Research (NIPER) and the American Automobile Manufacturers Association (AAMA).
Fuel parameters for 2007 and 2020 were projected from 1996 baseline values using information
from a February 26, 1999 report from Mathpro to the American Petroleum Institute. Data from
the EPA Emission Trends Database and other agency sources were used to develop appropriate
local modeling parameters for inspection maintenance programs, Stage n refueling controls, fuel
RVP, average ambient temperature, and other inputs.
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.70 This difference is especially significant for 1,3-
butadiene; its toxic/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. The target fuel is simply the fuel of
concern in the modeling analysis. These toxic emission rates were then weighted to come up with
a composite toxic emission factor, based on a distribution of 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.
HI-91
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table III.D-1. Areas Included in Toxic Emissions Modeling
Chicago, IL
Denver, CO
Houston, TX
Minneapolis, MN
New York, NY
Philadelphia, PA
Phoenix, AZ
Spokane, WA
St. Louis, MO
Atlanta, GA
Western WA/ OR
Northern CA
Southern CA
ID/ MT/ W Y
UT/NM/NV
West TX
ND/ SD/ NB/ IA/ KS/
Western MO
AR/ MS/ AL/ SC/ Northern
LA
Florida
Northeast States - non-I/M and
non-RFG
Northeast States - I/M and
non-RFG
Northeast States - non-I/M and
RFG
Ohio Valley - non-I/M and
non-RFG
Ohio Valley - I/M and non-
RFG
Ohio Valley - I/M and RFG
Northern MI/ WI
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
HI-92
-------
Chapter III: Environmental Impact
obtained from the unconsolidated version of the Complex Model. TOG-H was assumed to be the
same for all Tier 0 and later high emitting 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 ni.D-2. 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:
TOXHt Fuel AI FTP — A + B TOG
Baseline
(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:
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 III.D-2. 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
HI-93
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
These relationships can be thought of graphically, as illustrated in Figure IHD-1, below.
Hypothetical Benzene-TOG Curve
140
0.5
1 1.5
Baseline Fuel TOG (g/mi)
2.5
Figure III.D-1. 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
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/mi. 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
VOC emission rates below the Tier 0 normal emitter rate, for example, the toxic fraction stays
constant at the toxic fraction for Tier 0 normal emitters). In the example above, the benzene
emission rate for a baseline TOG value of 0.1 g/mi would be calculated as follows:
BZ
'(TOG=0.1g/mi)
0.1 g/mi * (16 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
HI-94
-------
Chapter III: Environmental Impact
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 regardless of the VOC emission level.
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 conditions71'72'73 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 EPA (1999d). 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-pj-p fLLJJ Aggressive Driving •rt-L/J TOX UC/FTP V V
where
TOXUC = Unified Cycle toxic emission rate
= FTP toxic emission rate
sive Driving = Adjustment to TOG emissions for aggressive driving
ADJTOXUC/FTP = Adjustment for difference in toxic mass fraction over the UC versus FTP
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 PMEmissions Modeling
HI-95
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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.
Diesel PM emission estimates are not presented in this section of the RIA since the
impact of this rulemaking on the diesel PM inventory is discussed in Section in.A.4. It should
be noted, however, that the diesel PM exposure estimates presented herein were based on
inventory numbers developed in a slightly different manner than the ones in Section HI. A.4. To
develop these inventory estimates, we modeled emission factors in 26 areas explicitly and then
mapped the remaining U. S. counties to these modeled areas, as described in Section in.D.2.a.i.
The resulting emission factors for each county were multiplied by VMT estimates from EPA's
Trends database to obtain total mass emissions estimates. The 2007 and 2020 light-duty diesel
PM emission estimates obtained using this approach were within about 15 percent of the
estimates presented in HI. A.4 for all scenarios modeled.
b. Nationwide Toxic Emissions Estimates - Baseline Scenario
Toxic emission factor estimates for each county in the United States were developed by
mapping them to one of the modeled areas (EPA, 1999d). The resulting county level emission
factors were multiplied by VMT estimates from EPA's Emission Trends database to come up
with nationwide emissions in tons. Forty-seven state estimates for gaseous toxics under baseline
scenarios in 1990, 1996, 2007, and 2020 are given in Table ffl.D-3;26 the diesel particulate
estimates can be found in or inferred from Table HI. A-l 1 for current diesel penetration rates and
Table ILL A-15 for increased diesel penetration rates. The baseline scenario 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.
26Diesel PM inventory estimates are presented elsewhere in the Regulatory Impact Analysis.
HI-96
-------
Chapter III: Environmental Impact
Table III.D-3. 47 State Highway Vehicle Toxic Emissions (tons)
In 1990,1996, 2007, and 2020, for Baseline Scenarios.
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3 -Butadiene
CY 1990
228,000
37,000
126,000
33,000
CY 1996
156,000
25,000
72,000
21,000
CY2007
89,000
14,000
34,000
12,000
CY2020
81,000
14,000
34,000
12,000
c. Exposure - Baseline Scenario
Exposure modeling was done for 1990 using the Hazardous Air Pollutant Exposure
Model for Mobile Sources, Version-3, or HAPEM-MS3.74'75 Data from 10 urban areas were
used. These areas were Atlanta, GA, Denver, CO, Houston, TX, Minneapolis, MN, New York,
NY, Philadelphia, PA, Phoenix, AZ, Spokane, WA, and St. Louis, MO. HAPEM-MS3 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
total ambient CO) by the fraction of the 1990 CO emissions inventory from on-road motor
vehicles, determined from the EPA Emission Trends database.76'77 Nationwide urban CO
exposure from on-road motor vehicles was estimated by first calculating a population-weighted
average CO exposure for the ten 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 1993 a).
HI-97
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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:
TO^T = rr^o ic^c\ i x TO^T (f\\
1 v~'yvExposure(ng/m3) L*-xV-'Exposure(ng/m3) EF(g/mi)Jl990 wyvEF(g/mi) Vu/
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. In the baseline scenario, we did not assume any increased penetration of
diesel engines into the light-duty fleet. 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.78 These factors account for the difference in reactivity between relatively
inert CO, which is being used as the tracer for toxics exposure, and 1,3-butadiene. 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 ni.D-4 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 III.D-4. Average 47 State 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
0.99
0.16
0.54
0.11
0.78
CY 1996
0.68
0.11
0.32
0.07
0.44
CY2007
0.39
0.06
0.15
0.04
0.25
CY2020
0.35
0.06
0.15
0.04
0.27
m-98
-------
Chapter III: Environmental Impact
It should be noted that recent California-EPA studies estimated a population-weighted
average outdoor diesel exhaust PM10 (particulate matter < 10 jim) exposure for 1995.79
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 47 State highway diesel
PM10 1996 exposure estimate of 0.44 |ig/m3 in Table in.D-4. One significant reason for the
difference is that the California estimate is for diesel PM10 from all sources, including nonroad
while the estimate in Table ni.D-4 is only for highway vehicles. Other reasons may be
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 30 ppm sulfur standard.
• NLEV, 30 ppm sulfur, and Tier 2 tailpipe standards
NLEV, 30 ppm sulfur, Tier 2 tailpipe standards, and increased diesel penetration
Estimates of the impact of VOC reductions from Tier 2 tailpipe standards for the full
useful life of the vehicle, combined with a 30 ppm sulfur standard, on toxics emissions and
exposure, are provided in Tables ni.D-6 through IHD-9.
The current updated assessment also evaluated the potential increase in diesel PM
emissions and exposure due to increased use of diesel engines in cars and 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 cars and light trucks. Some manufacturers have announced capital
investments to build such engines. The impact of this increase in light of Tier 2 standards were
evaluated. For our projections through 2015, we assumed the most likely level of increased
diesel penetration modeled in a draft report prepared by Arthur D. Little, Inc. for the American
Petroleum Institute.80 For years subsequent to 2015, we assumed that diesel engines' share of
vehicle sales would continue to grow at the rate projected for 2010 through 2015 in the Arthur D.
Little most likely scenario.
This assumption differs from that used to project the potential impact of greater diesel
sales on PM emissions from cars and light trucks found in Section A.4.b of this chapter, where
diesel engines' share of car and light truck sales were held constant at 2015 levels for subsequent
years. The difference in these projections stems from their different purposes. The projections in
Section A.4.b represent our efforts to project the impact of a likely diesel engine sales scenario
on PM emissions. The projections used for the diesel PM analysis described in this section are
HI-99
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
designed to illustrate the potential for greater diesel engine use in cars and light trucks to increase
the health risks associated with diesel PM; they represent a more speculative scenario and should
be used to evaluate the potential (rather than the most likely) impact of greater diesel engine use
on the unique health risks associated with diesel PM. The sales penetration rates for this
cautionary scenario are presented in Table IHD-5.
Table III.D-5. Percentage of sales fleet expected to be diesel in each respective year under
"most likely" scenario, as estimated in A. D. Little, Inc. report.
Vehicle Class
LDV
LDT1
LDT2
LDT3
LDT4
2000
0
1
0
2
3
2005
0
3
0
13
14
2010
2
17
12
23
23
2075
9
22
17
30
31
2020
16
27
22
37
39
The impact of such increased diesel penetration on exposure to diesel PM are provided in
Tables ni.D-8 and in.D-9. Based on the exposure estimates for 2020 under Tier 2 controls, the
potential 47 State cancer risk from diesel particulate matter would increase by about 8 percent
under this scenario. Beyond 2020, the health risks would be even greater for two reasons. First,
the proportion of cars and light trucks equipped with diesel engines would continue to increase as
the older, gasoline-powered cars and light trucks are replaced by a mix of gasoline and diesel cars
and light trucks. Second, continued growth in the total number of miles driven would increase
diesel PM emissions. It should be noted that without Tier 2 controls, we estimate that the
increased presence of diesel-powered cars and light trucks on the nation's roads could increase
the potential cancer risks associated with PM emissions from all diesel-powered highway
vehicles (including heavy-duty diesel trucks, diesel buses, and light-duty diesel vehicles) by
approximately 80 percent as of 2020. This estimate is based on an inventory of about 69,000
tons of diesel PM from all highway vehicles in 2020 without increased diesel penetration, versus
about 125,000 tons with increased diesel penetration. The 80 percent increase in inventory likely
would translate into a similar increase in cancer risk.
m-ioo
-------
Chapter III: Environmental Impact
Table IILD- 6. 47 State Highway Vehicle Toxic Emissions (tons)
in 2007, for Various Scenarios
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3 -Butadiene
No New
Controls
Scenario
89,000
14,000
34,000
12,000
30 ppm Sulfur
Scenario
82,000
13,000
33,000
11,000
Tier 2 Standard
w/ 30 ppm Sulfur
Scenario
80,000
13,000
32,000
11,000
Tier 2 Standard,
30 ppm Gasoline
Sulfur, &
Increased Diesel
Sales Scenario
78,000
13,000
33,000
10,000
Table III.D-7. 47 State Highway Vehicle Toxic Emissions (tons)
in 2020, for Various Scenarios
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3 -Butadiene
No New
Controls
Scenario
81,000
14,000
34,000
12,000
30 ppm Sulfur
Scenario
74,000
13,000
33,000
11,000
Tier 2 Standard
w/ 30 ppm Sulfur
Scenario
63,000
12,000
30,000
10,000
Tier 2 Standard,
30 ppm
Gasoline Sulfur,
& Increased
Diesel Sales
Scenario
55,000
12,000
30,000
9,000
m-ioi
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table III.D-8. Average 47 State Highway Vehicle Toxic Exposures for the Entire
Population (jig/m3) in 2007, for Various Scenarios
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3 -Butadiene
Diesel PM
No New
Controls
Scenario
0.39
0.06
0.15
0.04
0.25
30 ppm Sulfur
Scenario
0.39
0.06
0.15
0.03
0.25
Tier 2 Standard
w/ 30 ppm Sulfur
Scenario
0.35
0.06
0.14
0.03
0.25
Tier 2 Standard,
30 ppm
Gasoline Sulfur,
& Increased
Diesel Sales
Scenario
0.34
0.06
0.15
0.03
0.28
Table III.D-9. Average 47 State Highway Vehicle Toxic Exposures for the Entire
Population (jig/m3) in 2020, for Various Scenarios
Toxic
Benzene
Acetaldehyde
Formaldehyde
1,3 -Butadiene
Diesel PM
No New
Controls
Scenario
0.35
0.06
0.15
0.04
0.27
30 ppm Sulfur
Scenario
0.32
0.06
0.14
0.03
0.27
Tier 2 Standard
w/ 30 ppm Sulfur
Scenario
0.27
0.05
0.13
0.03
0.26
Tier 2 Standard,
30 ppm
Gasoline Sulfur,
& Increased
Diesel Sales
Scenario
0.24
0.05
0.13
0.03
0.29
m-102
-------
Chapter III: Environmental Impact
E. Carbon Monoxide
The standards being promulgated today will help reduce levels of carbon monoxide (CO).
Twenty areas, with a combined population of 33 million, are designated as being in
nonattainment with the CO NAAQS. An additional 24 areas with a combined population of 22
million are designated as CO maintenance areas. In 1997, 6 of 537 monitoring sites reported
ambient CO levels in excess of the CO NAAQS.
As discussed in Section in.A, the Tier 2/Sulfur standards will require light trucks to meet
more stringent CO standards. These more stringent standards will help extend the trend towards
lower CO emissions from motor vehicles and thereby help the remaining CO nonattainment areas
reach attainment while helping other areas remain in attainment with the CO NAAQS. The
analysis of economic benefits and costs found in Section IV.D.-5. does not account for the
economic benefits of the CO reductions expected to result from today's proposal.
m-103
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Chapter HI References
1. "Development of On-Highway Inventory Adjustment Factors Used in the Tier 2 Final
Rule Air Quality Analysis", Memorandum from John Koupal and Gary Dolce to Docket
A-97-10, October 18, 1999
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, March 1999
3. E.H. Pechan and Associates, "Procedures for Developing Base Year and Future Year
Mass and Modeling Inventories for the Tier 2 Final Rulemaking", Report to U.S. EPA,
September 1999.
4. "Development of Light-Duty Emission Inventory Estimates in the Final Rulemaking for
Tier 2 and Sulfur Standards", Memorandum from John Koupal to Docket No. A-97-10
5. Ibid.
6. "Determination of Tier 2/Sulfur Emission Reductions in Terms of Equivalent Baseline
Vehicles for the Tier 2 Final Rule", Memorandum from John Koupal to Docket No. A-
97-10
7. A.D. Little, Inc., "U.S. Light-Duty Dieselization Scenarios - Preliminary Study", Report
to the American Petroluem Institute, July 1999
8. Memorandum, "Ozone Attainment Demonstrations," issued March 2, 1995. A copy of
the memorandum may be found on EPA's web site at
http://www.epa.gOv/ttn/oarpg/t 1 pgm. html
9. Letter from Mary A. Gade, Director, State of Illinois Environmental Protection Agency to
Environmental Council of States (ECOS) Members, dated April 13, 1995.
10. Memorandum, "Guidance for Implementing the 1-Hour Ozone and Pre-Existing PM 10
NAAQS," issues December 29, 1997. A copy of this memorandum may be found on
EPA's web site at http://www.epa.gov/ttn/oarpg/tlpgm.html.
11. Memorandum, "Extension of Attainment Dates for Downwind Transport Areas," issued
July 16, 1998. This memorandum is applicable to both moderate and serious ozone
nonattainment areas. A copy of this policy may be found on EPA's web site at
http://www.epa.gov/ttn/oarpg/tlpgm.html .
m-104
-------
Chapter III: Environmental Impact
12. U.S. EPA, (1991), Guideline for Regulatory Application of the Urban Airshed Model,
EPA-450/4-91-013, (July 1991). A copy may be found on EPA's web site at
http://www.epa.gov/ttn/scram/ (file name: "UAMREG"). See also U.S. EPA, (1996),
Guidance on Use of Modeled Results to Demonstrate Attainment of the Ozone NAAQS,
EPA-454/B-95-007, (June 1996). A copy may be found on EPA's web site at
http://www.epa.gov/ttn/scram/ (file name: "O3TEST").
13. Lake Michigan Air Directors Consortium. Midwest Subregional Modeling: 1-Hour
Attainment Demonstration - Tier IJ/Low S Controls. November 8, 1999.
14. Evaluation of Control Strategies For The Beaumont-Port Arthur Area. Environmental
Programs MCNC-North Carolina Supercomputing Center for the Texas Natural
Resources Conservation Commission. June 30, 1999.
15. Web pages operated by the North Central Texas Council of Governments.
http://dfwinfo.com/envir/aq/ozone/dfwsip/html
http://dfwinfo.com/envir/aq/ozone/100799/tnrcc/index.html
http://dfwinfo.com/envir/aq/ozone/100799/tnrcc/sld008.htm
Downloaded and placed in Docket A-97-10. November 10, 1999.
16. California Air Resources Board, Executive Order G-99-037, May 20, 1999, Attachment
A, p.6-7, 10.
17. National Air Quality and Emissions Trend Report, 1997. EPA454/R-98-016. December
1998.
18. U.S. EPA, 1996, Air Quality Criteria for Particulate Matter, EPA/600/P-95/001aF.
18. U.S. EPA, 1996, Air Quality Criteria for Particulate Matter, EPA/600/P-95/001aF.
Review of the National Ambient Air Quality Standards for Particulate Matter: Policy
Assessment of Scientific and Technical Information, OAQPS Staff Paper, EPA-452 \R-
96-013, July 1996.
20. (NFRAQS) Northern Front Range Air Quality Study, Colorado. January 1998, Volume I.
http://charon/cira/colostate/edu/
21. Friedlander, S.K. (1973) Chemical Element Balances and Identification of Air Pollution
Sources Environ. Sci. & Technol. 7:235-240.
22. Gartrell, G., and S.K. Friedlander (1975) Relating particulate pollution to sources: The
1972 California Aerosol Characterization Study. Atmos. Environ. 9:279-299.
m-105
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
23. Schauer J.J., Rogge W.F., Hildemann L.M., Mazurek M.A., Cass G.R. and B.R.T.
Simoneit (1996) Source apportionment of airborne particulate matter using organic
compounds as tracers. Atmos. Environ. 38:3837-3855.
24. Lowenthal D.H., Chow J.C., Watson J.G., Neuroth G.R., Robbins R.B., Shafritz B.P.,
RJ. Countess (1992) The Effects of Collinearity on the Ability to Determine Aerosol
Contributions from Diesel- and Gasoline-powered Vehicles using the Chemical Mass
Balance Model. Atmos. Environ. 26A:2341-2351.
25. Wittorff, D.N.; Gertler, A.W.; Chow, J.C.; Barnard, W.R.; Jongedyk, H.A. The Impact of
Diesel Particulate Emissions on Ambient Paniculate Loadings. Presented at the 87th
Annual Meeting of the Air & Waste Management Assoc., Cincinnati, OH June 19-24,
1994.
26. EPA Memo to Docket A-97-10, "1996-1998 PM10 Attainment Status Determination."
Robert J. Wayland. Integrated Policy and Strategies Group, Office of Air Quality
Planning and Standards, November 12, 1999.
27. EPA Memo, "Guidance for Implementing the 1-Hour Ozone and Pre-Existing PM10
NAAQS." Richard D. Wilson, Acting Assistant Administrator for Air and Radiation.
January 8, 1998.
28. 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, N.C., July 16, 1997.
29. EPA Memo, Tables from the Draft 1998 National Air Quality and Emissions Trends
Report, December 9, 1999, David Mintz, Air Quality Trends and Analysis Group,
EMAD, OAQPS.
30. Gray, H.A. and A. Kuklin (1996) Benefits of Mobile Source NOx Related Paticulate
Matter Reductions. Systems Applications International. Final Report prepared for U.S.
Environmental Protection Agency. SYSAPP
31. Chow, J.C., Watson, J.G., Richards, L.W., Haase, D.L., McDade, C., Dietrich, D.L.,
Moon, D., and C. Sloane (1991) The 1989-1990 Phoenix PM10 Study. Volume H: Source
Apportionment. Final Report. DRI Document No. 8931.6F1, prepared for Arizona
Department of Environmental Air Quality, Phoenix, AZ, by Desert Research Institute,
Reno, NV.
m-106
-------
Chapter III: Environmental Impact
32. California Environmental Protection Agency (1998) Report to the Air Resources Board
on the Proposed Identification of Diesel Exhaust as a Toxic Air Contaminant. Appendix
IE, Part A: Exposure Assessment. April 1998.
33. South Coast Air Quality Management District (SCAQMD) (1999) "Hot-Spot"
Monitoring and Multiple Air Toxics Exposure Study (MATES II).
http://www.aqmd.gov/news/mates.html. Accessed on 20 July 1999.
34. National Air Pollutant Emission Trends Update, 19770 - 1997. U.S. EPA, Office of Air
Quality Planning and Standards, EPA-454/E-98-007. December 1998.
35. Grand Canyon Visibility Transport Commission, Recommendations for Improving
Western Vistas. June 10, 1996.
36. EPA 1998a. Environmental Protection Agency, Carcinogenic Effects of Benzene: An
Update, National Center for Environmental Assessment, Washington, DC. 1998.
37. 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.
38. Clement Associates, Inc., Motor vehicle air toxics health information, for U.S. EPA
Office of Mobile Sources, Ann Arbor, MI, September 1991.
39. 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.
40. 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
granulocyte/macrophage colony-stimulating factor in vitro, Proc. Natl. Acad. Sci.
89:3691-3695, 1992.
41. Lumley, M., H. Barker, and J.A. Murray, Benzene in petrol, Lancet 336:1318-1319, 1990.
42. 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.
43. Aksoy, M. 1991. Hematotoxicity, leukemogenicity and carcinogenicity of chronic
exposure to benzene. In: Arinc, E.; Schenkman, J.B.; Hodgson, E., Eds. Molecular
m-107
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Aspects of Monooxygenases and Bioactivation of Toxic Compounds. New York:
Plenum Press, pp. 415-434.
44. Goldstein, B.D. 1988. Benzene toxicity. Occupational medicine. State of the Art
Reviews. 3: 541-554.
45. Aksoy, M., S. Erdem, and G. Dincol. 1974. Leukemia in shoe-workers exposed
chronically to benzene. Blood 44:837.
46. 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.
47. 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. Hematotoxicity among Chinese workers heavily exposed to benzene.
Am. J. Ind. Med. 29: 236-246.
48. EPA 1998b. Environmental Protection Agency, Toxicological Review of Benzene (Non-
Cancer Effects), July 1998 draft. National Center for Environmental Assessment,
Washington, DC.
49. EPA, 1985. Mutagenicity and carcinogenicity assessment of 1,3-butadiene. EPA/600/8-
85/004F. U.S. Environmental Protection Agency, Office of Health and Environmental
Assessment. Washington, DC.
50. EPA 1998c. Environmental Protection Agency, Health Risk Assessment of 1,3-
Butadiene. EPA/600/P-98/001A, February 1998.
51. Scientific Advisory Board. 1998. An SAB Report: Review of the Health Risk Assessment
of 1,3-Butadiene. EPA-SAB-EHC-98, August, 1998.
52. Denzell, E., N. Sathiakumar, M. Macaluso, M. Hovinga, R. Larson, F. Barbone, C. Beall,
and P. Cole, 1995. A follow-up study of synthetic rubber workers. Final report prepared
under contract to International Institute of Synthetic Rubber Producers, October 2, 1995.
53. EPA 1999a. Memo from Dr. Aparna Koppikar, ORD to Laura McKlevey, OAQPS and
Pamela Brodowicz, QMS. Slope Factor for 1,3-Butadiene, April 26, 1999.
54. 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.
m-108
-------
Chapter III: Environmental Impact
55. EPA 1987. Environmental Protection Agency, Assessment of health risks to garment
workers and certain home residents from exposure to formaldehyde, Office of Pesticides
and Toxic Substances, April 1987.
56. Wilhelmsson, B. and M. Holmstrom. 1987. Positive formaldehyde PAST after prolonged
formaldehyde exposure by inhalation. The Lancet: 164.
57. 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.
58. Hendrick, D.J., RJ. Rando, DJ. Lane, and MJ. Morris. 1982. Formaldehyde asthma:
Challenge exposure levels and fate after five years. J. Occup. Med. 893-897.
59. Nordman, H., H. Keskinen, and M. Tuppurainen. 1985. Formaldehyde asthma - rare or
overlooked? J. Allergy Clin. Immunol. 75:91-99.
60. 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.
61. Ligocki, M.P., and G.Z. Whitten, Atmospheric transformation of air toxics:
acetaldehydeand polycyclic organic matter, Systems Applications International, San
Rafael, CA, (SYSAPP-91/113), 1991.
62. 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.
63. 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.
64. EPA 1997b. Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria
andAssessment Office, Cincinnati, OH, 1997.
65. EPA 1991b. Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office, Cincinnati, OH.
66. EPA 1993b. Environmental Protection Agency, Integrated Risk Information System
(IRIS), Office of Health and Environmental Assessment, Environmental Criteria and
Assessment Office, Cincinnati, OH.
m-109
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
67. EPA, 1999c. Estimation of Motor Vehicle Toxic Emissions and Exposure in Selected
Urban Areas. Prepared by Sierra Research, Inc., Radian International Corp., and Energy
& Environmental Analysis, Inc. for U. S. EPA, Office of Mobile Sources, Assessment
and Modeling Division, Ann Arbor, MI, Report No. EPA420-D-99-002, March 1999.
68. 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.
69. EPA. 1999d. Analysis of the Impacts of Control Programs on Motor Vehicle Toxic
Emissions and Exposure in Urban Areas and Nationwide: Volumes I and II. Prepared for
EPA by Sierra Research, Inc. and Radian International Corporation/Eastern Research
Group, November 30, 1999. Report Nos. EPA420-R-99-029, andEPA420-R-99-030.
70. EPA. 1994. Regulatory Impact Analysis for the Final Rule on Reformulated and
Conventional Gasoline, February, 1994.
71. Auto/Oil Air Quality Improvement Research Program. Technical Bulletin No. 19:
Dynamometer Study of Off-Cycle Exhaust Emissions; April, 1996.
72. 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.
73. CARB. 1998. Unpublished data.
74. 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.
75. Glen, G. and Shadwick, D. 1999. HAPEM-MS3 Exposure Modeling data for Atlanta.
Prepared by Mantech Environmental Technology, Inc. for the U.S. Environmental
Protection Agency.
76. 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.
77. E. H. Pechan and Associates, Inc. 1999. CO Inventories and Mobile Source Contribution
for Atlanta. Prepared for U. S. EPA, Office of Mobile Sources, Assessment and
Modeling Division.
m-no
-------
Chapter III: Environmental Impact
78. Systems Applications International. 1994. Projected Emission Trends and Exposure
Issues for 1,3-Butadiene. Prepared for the American Automobile Manufacturers
Association, March, 1994.
79. California-EPA and the California Air Resources Board, Proposed Identification of
Diesel Exhaust as a Toxic Air Contaminant, Appendix HI, Part A, Exposure Assessment,
April 22, 1998.
80. Arthur D. Little, Inc. 1999. U. S. Light Duty Dieselizati on Scenarios-Preliminary
Study. Prepared for the American Petroleum Institute, July 2, 1999.
m-iii
-------
Chapter IV: Technological Feasibility
Chapter IV: Technological Feasibility
A. Feasibility of Tier 2 Exhaust Emission Standards for Vehicles
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, might 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. Perhaps most important of
all, emission control calibrations continue to become more refined and sophisticated.
Table IV-1 below lists specific types of emission controls which EPA projects will be
needed in order for the affected vehicles to meet the final Tier 2 standards. It is important to
point out that all of the following technologies would not necessarily be needed to meet the 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 the need for increases in catalyst
volume and precious metal loading, EPA believes that most, if not all, cars and trucks will use
the specified emission control technique.
IV-1
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table IV-1. Emission Control Hardware and Techniques
Projected to Meet 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 PGM Loading
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.
rv-2
-------
Chapter IV: Technological Feasibility
Combustion Chamber Design
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
IV-3
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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
as 20 to 251 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. Some manufacturers use a mechanical back-pressure system
that measure EGR flow (via delta pressure across an orifice) rather than inferring flow from the
EGR pintle position. This system uses electronic control of the vacuum actuation and has very
precise control. Many manufacturers are now using 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, in some cases, 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
1 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
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.
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. ARE also projects that in order to
meet LEV II and ULEV n 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.
IV-5
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
/'/'. 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 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 (FIEGO)
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 HEGO 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 n 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
IV-6
-------
Chapter IV: Technological Feasibility
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
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 FIEGO sensors. Planar FIEGO 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
IV-7
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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.
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
IV-8
-------
Chapter IV: Technological Feasibility
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
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
IV-9
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 Pd/Rh, Pt/Rh 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. In 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 an increased reliance on Pd, 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 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 and 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.
IV-10
-------
Chapter IV: Technological Feasibility
SAE 960802: 1.8 liter 4 cyl; 100 h aged; Pd/Rh=5/l @ 50 g/cu. ft.
| Single layer Pd/Rh
]Two layer - Pdtop
]Two 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 and 1200
cpsi catalysts has been progressing. Typical cell densities for conventional catalysts are 400 cpsi.
We have projected that in order to meet the 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 light and medium duty trucks
only have ratios of 0.6 or less. We believe that in order to comply with Tier 2 standards, most
vehicles 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 Tier 2 standards. Typical catalyst loadings 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 - 250 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
IV-11
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 be 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 results
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,
providing a greater quantity of heat for heating the exhaust manifold, headpipe, and catalyst.
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
IV-12
-------
Chapter IV: Technological Feasibility
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.
One manufacturer who incorporates a zeolite HC adsorber in its California SULEV
vehicle found that an electrically heated catalyst was necessary after the adsorber because the
zeolite acts as a heat sink and nearly negates the cold start advantage of the HC adsorber.
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
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.
IV-13
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 more quickly, 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.
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 after the catalytic converters have been thermally aged to
the equivalent of both the intermediate useful life (50,000 miles) and full useful life (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 Tier 2 standards at intermediate useful would be approximately 0.035
to 0.050 g/mi NMOG and 0.025 to 0.035 g/mi NOx. At full useful life, the design targets for the
Tier 2 standards would be approximately 0.045-0.063 g/mi NMOG and 0.035-0.050 g/mi NOx at
IV-14
-------
Chapter IV: Technological Feasibility
full useful life.
With this in mind, we will present data from several sources that establish our Tier 2
standards to be feasible. The data ranges from certification emission levels to feasibility
evaluation programs undertaken in the last year by EPA, ARB and MECA. Even though theARB
and MECA 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 Tier 2 standards are the same as the LEV II standards. We will also present
the results of an EPA test program that demonstrates the feasibility of the Tier 2 emissions
standards for the largest sport utility vehicles and pickup trucks regulated under this final rule.
/'. 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 1999 model year 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 1999 Model Year 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
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
IV-15
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
engine families certified to nationwide Tier 1 standards, NLEV program standards, and the
California program standards.
IV-16
-------
Chapter IV: Technological Feasibility
Table IV-3. 1999 MY 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
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
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*
0.05
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
LEV
Tierl
Tierl
TLEV
TLEV
LEV
Tierl
LEV
LEV
LEV
TLEV
Tierl
TLEV
Tierl
TLEV
TLEV
LEV
Tierl
TLEV
Tierl
LEV
Tierl
TLEV
TLEV
IV-17
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Volkswagen
Mazda
Mitsubishi
Volvo
Volvo
Daimler Chrysler
Honda
Honda
Honda
Infiniti
New Beetle, New Golf, New Jetta
MX-5 Miata
Mirage
S80
S80
C230 Kompressor
Accord
Civic HX
Civic
Q45
0.06*
0.07
0.07
0.06-0.07*
0.04-0.05
0.07
0.07*
0.07*
0.07*
0.07*
0.06
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
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
IV-18
-------
Chapter IV: Technological Feasibility
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.
Table IV-4 provides a listing of 2000 model year 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. As for the 1999 data, where a range is shown, there is more than
one configuration within the engine 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. We assembled this list by reviewing
2000 model year certification data for engine families certified to nationwide Tier 1 standards,
NLEV program standards, and the California program standards. At the time this document was
published, approximately 90 percent of the 2000 model year certification data had been
submitted.
Table IV-4. 2000 MY 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
Daimler Chrysler
Ford
Volvo
Ford
Daimler Chrysler
Mitsubishi
Mitsubishi
Daewoo
SAAB
Daimler Chrysler
Hyundai
Ford
Tiburon, Elantra
Neon
Mystique, Contour
S80
Mystique
Neon
Eclipse, Gallant
Mirage
Lanos
9-5
Stratus
Tiburon, Elantra
LS
0.01-0.02+
0.01
0.01-0.04*
0.01
0.02-0.03+
0.02
0.02*+
0.02-0.03*+
0.02-0.07
0.03
0.03
0.03-0.04*+
0.03 -0.05
0.05
0.05
0.03-0.05
0.05
0.03 -0.04
0.04
0.02 - 0.04
0.03-0.04
0.06 - 0.07
0.03
0.06
0.04-0.05
0.06-0.07
LEV
ULEV
TLEV
LEV
Tierl
ULEV
LEV
LEV
LEV
LEV
LEV
Tierl
LEV
IV-19
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Volvo
Toyota
Jaguar
Mazda
Honda
Volvo
Daimler Chrysler
GM
GM
Honda
Mazda
Volvo
Hyundai
Daimler Chrysler
Nissan
Kia
Honda
Infiniti
Ford
Toyota
Volkswagen
Daewoo
Honda
Daewoo
Honda
Daimler Chrysler
BMW
Nissan
S70, V70
Avalon, Lexus ES300
X200
Protege
Accord
S80
SLK230 Kompressor, C230
Kompressor
Metro
Park Avenue
Accord
Protege
S40
Sonata
ML320
Infiniti G20
Sephia
Accord
130
Contour, Cougar
Lexus GS300/GS400
Jetta
Nubira, Lanos
Insight
Leganza, Nubira
Accord
E430, S500
X5
Altima
0.03*
0.03 -0.06
0.03-0.05*
0.04+
0.04-0.06
0.04-0.06*
0.04-0.05
0.04+
0.04+
0.04-0.06
0.04
0.05-0.06*
0.05
0.05*
0.06*
0.06-0.07*
0.06-0.07*
0.06
0.06*
0.06*
0.06*
0.06*
0.06*
0.07*
0.07*
0.07*
0.07
0.07
0.03-0.04
0.05
0.05-0.07
0.03
0.04-0.05
0.09
0.04-0.05
0.03
0.04
0.05-0.06
0.03
0.06 - 0.07
0.05
0.04-0.05
0.04
0.04-0.08
0.06
0.05
0.16
0.05-0.06
0.06
0.08
0.04
0.05-0.07
0.03
0.02
0.04
0.06
LEV
LEV
TLEV
LEV
LEV
TLEV
TLEV
TLEV
LEV
LEV
LEV
LEV
TLEV
Tierl
LEV
LEV
LEV
LEV
Tierl
LEV
LEV
TLEV
ULEV/LEV
LEV
ULEV
LEV
LEV
LEV
IV-20
-------
Chapter IV: Technological Feasibility
SAAB
Hyundai
9-5
Accent/Brio
0.07
0.07*
0.02-0.03
0.03
LEV
LEV
LDT1
Toyota
Kia
Ford
Mazda
Daimler Chrysler
Tacoma
Sportage
Ranger Pickup
B2000
Dakota Pickup
0.01-0.02
0.02-0.05
0.04-0.07
0.06*
0.07
0.05-0.07
0.04-0.06
0.09-0.18
0.05
0.08
LEV
LEV
Tierl
Tierl
TLEV
LDT2
Ford
Ford
Mazda
GM
Daimler Chrysler
Daimler Chrysler
F 150 Pickup
Ranger Pickup
B3000
Montana
Grand Cherokee
Caravan
0.03*
0.04*
0.04-0.05*
0.05+
0.05
0.06
0.13-0.16
0.13-0.15
0.06-0.10
0.05-0.06
0.10-0.11
0.09
Tierl
Tierl
Tierl
LEV
LEV
LEV
LDT3
Ford
Daimler Chrysler
Ford
Daimler Chrysler
Land Rover
F 150 Pickup
Durango
F 150 Pickup
ML55
Range Rover, Discovery
0.03-0.04
0.05
0.05*
0.06 - 0.07
0.07*
0.16-0.20
0.08
0.14-0.16
0.04-0.05
0.09-0.17
Tierl
ULEV
Tierl
LEV
Tierl
LDT4
Ford
F250 Pickup
0.04-0.05
0.13-0.21
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 levels.
A review of the Tables above show that most of the engine families with configurations
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
IV-21
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 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 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 California LEV II and our
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 the equivalent of 50,000
miles 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 Tier 2 standard with a headroom of 23
percent. The actual test results are summarized in Table IV-5 below.
IV-22
-------
Chapter IV: Technological Feasibility
Table IV-5. MECA Test Program: Emissions with Catalysts Aged to 100,000 Miles (g/mi)
Tier 2 Design Targets
Crown Victoria (LDV)
Buick LeSabre (LDV)
Toyota Tl 00 (LDT2)
NMOG
0.045-0.063
0.049
0.038
0.052
NOx
0.035-0.049
0.057
0.037
0.014
iv. CARB Test Program
CARB tested five different 1997-98 model year production LEV LDV models. Two of
the five models met the 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-6. CARB 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 LEV n 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
Tier 2 NMOG and NOx emission standards. While these results are not with catalysts aged to
IV-23
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
full useful life, we believe these results are still very promising, since in-use deterioration rates
have been steadily declining. Even if these emissions were to double, they would still be very
close to or below the Tier 2 standards. Table IV-7 lists the modified passenger car emission
results.
Table IV-7. ARE 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-8 lists the
baseline emission results for the two Expeditions.
Table IV-8. CARB 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. CARB 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-9 lists the
emission results of the Expeditions with advanced catalyst systems.
IV-24
-------
Chapter IV: Technological Feasibility
Table IV-9. CARB 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 large 1999 LDT3 and LDT4
heavy-light-duty trucks to levels at or below those of the Tier 2 Standards at intermediate life
(50,000 miles). All of the vehicles tested had large displacement (greater than 5.3 liter), high
horsepower (230-270 hp) engines; four wheel drive; curb weights of 4,500 to 5500 pounds; and
gross vehicle weights of greater than 6,000 Ibs. Specifications of the trucks tested are included in
table IV-10.
IV-25
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table IV-10: EPA Test Vehicle Specifications
Trucks Tested
1999 Ford
Expedition LEV
1999 GM
Chevrolet
Silverado LEV
Test
Weight
(Ibs.)
5876
4818
Engine
5.3LV8,
230 bhp
5.4 LV8,
270 bhp
Drivetrain
4-speed
Auto., 4-WD
4-speed
Auto., 4-WD
Intermediate Useful Life
(50, 000 mile) Certification Levels
NMOG
(g/mi)
0.09
0.11
CO
(g/mi)
1.7
2.4
NOx
(g/mi)
.07
0.3
A key element of the test was the alteration of engine calibration parameters of the
powertrain control module (PCM), which included modification of spark timing, EGR, and fuel
control. During testing at EPA-NVFEL, flash-reprogramming of the PCM, off-board ROM, and
ROM emulation were used to accomplish PCM calibration changes. All of the catalysts/exhaust
systems evaluated were thermally aged to an equivalent 50,000 miles using the vehicle
manufacturers' specific catalyst aging cycle.
Specifications of some of the exhaust catalyst systems tested in this program can be found
in table IV-11.
IV-26
-------
Chapter IV: Technological Feasibility
Table IV-11: Catalyst Specifications
Catalyst Configuration
Total Catalyst Volume (L)
Total No. of Catalyst Bricks
Total Pd Loading (g)
Total Rh Loading (g)
Front Bricks
Middle Bricks
Rear Bricks
Volume (L)
Pd Loading (g)
Cell Density (cells/in2)
Wall Thickness (mil)
Volume (L)
Pd Loading (g)
Rh Loading (g)
Cell Density (cells/in2)
Wall Thickness (mil)
Volume (L)
Pd Loading (g)
Rh Loading (g)
Cell Density (cells/in2)
Wall Thickness (mil)
Ford Expedition
EXP1
(stock
OEM)
5.3
6
17.2
0.82
0.69
4.87
400
6
1.06
2.02
0.22
400
6
0.9
1.72
0.19
400
6
EXP3
5.9
6
26.4
4.18
0.69
4.87
400
6
1.26
4.45
1.12
600
4
1.01
3.88
0.97
600
4
GM Chevrolet
Silverado
SILV1
(stock
OEM)
4.8
4
4.6
0.28
1.2
1.29
0
400
6
1.2
1.02
0.14
400
6
SILV2
3.9
6
69
4.94
0.345
12.17
600
4
0.695
13.14
1.45
400
6
0.907
9.18
1.02
400
6
Only minor changes were made to the Chevrolet Silverado PCM calibration. These included:
• 4 to 6 degree spark timing retard under cold-start conditions to improve catalyst light-off
times
IV-27
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
• Earlier enablement of EGR after cold-start, using the original EGR map
The majority of the emissions improvement on the tested Silverado configurations are due to
increased precious metal loading of the exhaust catalysts tested. No measurable differences in
fuel economy were noted after the changes. Driveability was not affected. The final tested
configuration of the Chevrolet Silverado achieved NOx emissions of 0.05 g/mi and NMHC
emissions of 0.06 g/mi, meeting the Tier 2 standard.
The availability of ROM emulation for the Ford Expedition allowed considerable PCM
calibration tuning to be performed both on the chassis dynamometer and while driving on the
highway. The tested Expedition configurations relied considerably more on calibration tuning
than the Silverado.
We found that more than 80 percent of the NOx and NMHC emissions from the Expedition
occurred during the first 30 seconds after a cold start. Therefore, most of the calibration tuning
focused on reducing NOx and NMHC emissions from the cold start portion, or "phase 1", of the
light-duty FTP. Some of the calibration changes included:
• 15 to 20 degree spark timing retard under cold-start conditions to improve catalyst light-
off times
• Minor spark timing retard to increase catalyst temperatures at lighter load, lower speed
conditions
• Earlier enablement of EGR (enabled after -30 seconds under typical FTP cold-start
conditions)
• Increased EGR rates, particularly at part load conditions
• Extension of the EGR map to cover higher-speed, higher load driving conditions
The retarded spark timing initially after cold start resulted in increased front catalyst brick
temperatures, which were increased from 425 °C to 550 °C at 30 seconds after cold start.
Considerable EGR tuning approximately halved engine-out (pre-catalyst) NOx emissions.
Maximum EGR rates did not exceed 14 percent, and were considerably less for most operating
conditions. Engine-out CO was unchanged by the additional EGR. Engine-out HC was
increased by 5 to 15 percent. The engine-out HC increase due to addtional EGR was more than
offset by higher catalyst efficiency due to the higher PGM loading and volume of the underfloor
catalyst, and due to the increased catalyst temperatures immediately after cold-start from the cold
spark retard.
Cold start NOx performance was further improved by the use of low-mass, sealed-air-gap,
tubular-steel exhaust manifolds. The prototype manifolds further increased front catalyst brick
temperatures from 550 °C to 630 °C at 30 seconds after cold start. Catalyst brick temperatures
did not exceed 850 °C for any of the tested configurations, even over the US06 cycle. The
reliance on Pd and Pd/Rh formulations, and stabilized cerium oxide, allowed a safe margin with
respect to catalyst brick temperatures. Catalyst manufacturers have indicated to us that current
catalyst formulations can typically withstand temperatures of 950 °C to 1000 °C without
IV-28
-------
Chapter IV: Technological Feasibility
damage.
Considerable tuning of the PCM calibration was also used to minimize impacts of the calibration
changes on driveability and fuel economy. The final calibrations achieved considerable
improvements in emissions performance with no measurable impact on fuel economy and no
perceptible change in driveability. The Expedition achieved NOx emissions of 0.04 g/mi and
NMHC emissions 0.07 g/mi with the OEM cast exhaust manifolds, and 0.02 g/mi NOx, 0.07
g/mi NMHC with the sealed-air-gap exhaust manifolds. The final tested configurations easily
met the Tier 2 NOx standard. The NMHC emissions meet the 50,000 mile standard. Use of
close-loop controlled secondary-air-injection (similar to that used by SwRI for the MECA test
program) would further reduce cold-start NMHC emissions with only a minor degradation in
NOx performance.
In addition to testing at EPA, a virtually identical 1999 Ford Expedition was tested under
an EPA contract at Southwest Research Institute (SwRI) using the ERIC system to facilitate
calibration changes. The exhaust catalyst system tested was identical to the system tested by
ARB with a 1998 Ford Expedition (see section iv). Data from a 1999 Chevrolet Silverado
similar to the vehicle tested at NVFEL was provided by MECA. This vehicle was also tested at
SwRI using the ERIC system to provide engine calibration changes. Emissions from the trucks
tested at NVFEL and at SwRI for a number of the tested exhaust catalyst and engine calibration
configurations are compared in figures IV-2 and IV-3. The trucks tested at SwRI achieved
approximately the same low emissions levels as those tested at NVFEL, even though their mix of
hardware and calibration changes were relatively different. The low emissions levels achieved
essentially demonstrate the feasibility of the Tier 2 standards for heavy-light-duty trucks and the
ability to achieve those standards using a variety of logical engineering paths.
IV-29
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
0.400
0.300
E
5
in
.2 0.200
VI
V)
I
LJJ
X
O
z
0.100
MDV-2 LEV-I Standard
0.000
11999 GM Chevrolet Silverado Certification
Data
EPA Silverado, advanced catalysts
EPA Silverado, advanced catalysts, sealed-air-
gap exh. manifolds, minor calibration changes
QSwRI/MECA Silverado, advanced catalysts
<>SwRI/MECA Silverado, advanced catalysts
secondary air, calibration changes
oA
O
0.000
0.050 0.100
NMHC Emissions (g/mi)
0.150
Figure IV-2: 50,000 mile equivalent NOx vs. NMHC levels for a number of hardware and
engine calibration configurations tested with a 1999 GM Chevrolet Silverado Pickup (5.3L
V8) originally certified to the LEV MDV-2 standard (0.4 g/mi NOx, 0.16 g/mi NMOG).
IV-30
-------
Chapter IV: Technological Feasibility
0.600
MDV-3 LEV-I Standard
0.500 -
1 0.400 -
s
I/I
.2 0.300 -
U)
U)
I
LJJ
^ 0.200 -
0.100 -
0.000
O SwRI/CARB/EPA Expedition, advanced
catalyst, secondary air, calibration changes
11999 LEV Expedition (certification data)
rj EPA Expedition, stock/OEM configuration
EPA Expedition, advanced catalyst,
calibration changes
0.000
0.050 0.100 0.150
NMHC/NMOG Emissions (g/mi)
Figure IV-3: 50,000 mile equivalent NOx vs. NMHC emissions levels for a number of
hardware and engine calibration configurations tested with a 1999 Ford Expedition (5.4L
V8) originally certified to the LEV MDV-3 standard (0.6 g/mi NOx, 0.195 g/mi NMOG).
The technologies and emission control strategies that will be used for LDT3 and LDT4
vehicles should also apply directly to medium-duty passenger vehicles (MDPVs), which have a
GVWR greater than 8,500 pounds. In our LDT technology demonstration program discussed
above, we found that a combination of calibration changes and improvements to the catalyst
system resulted in emission levels for NOx well below and NMHC/NMOG approximately at the
Tier 2 intermediate useful life standards. The catalyst improvements consisted of increases in
volume and precious metal loading, and higher cell-densities than those found in the original
hardware. We are confident that the use of secondary-air-injection will greatly help cold-start
hydrocarbon control, making the NMOG standards achievable.
The most significant difference between LDT4s and MDPVs is that MDPVs have a
vehicle weight up to 800 pounds more than LDT4s. MDPVs will also be typically equipped
with larger displacement engines. The potential impact of these differences is higher engine-out
IV-31
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
emissions than LDT4s due to the larger engine displacement and greater load that the engine will
be operated under due to the extra weight. However, neither of these preclude manufacturers
from applying the same basic emission control technologies and strategies as used by LDVs and
LDTs. The only difference will likely be the need for larger catalysts with higher precious metal
loading than found in LDT4s. We are confident that MDPVs will be capable of meeting the Tier
2 standards.
We are currently testing a Ford Excursion as part of our LDT technology demonstration
program. Preliminary baseline results with a green (i.e., "new") catalyst indicate that emission
levels are higher than baseline emissions for the Ford Expedition. These results, although with a
green catalyst, are well below our interim Tier 2 upper bin standards. In fact, the majority of
these vehicles certified on the chassis dynamometer in California, have certification levels well
below our interim upper bin standards. We have also tested the Excursion at loaded vehicle test
weight (curb + 300 Ib) and again at adjusted loaded vehicle weight (half payload) and found that
the engine-out and tailpipe emission results for NMHC and NOx were the same for the two test
weights. In other words, the additional weight (approximately 700 Ibs) had no impact on
emission performance. This suggests that challenge for MDPVs in meeting Tier 2 standards may
not be as difficult as originally believed.
While this testing is still ongoing, we feel that the preliminary results are encouraging
since they suggest that the additional weight for these vehicles may not be as significant as
originally thought, and the difference in emission results between the Excursion and Expedition
suggest that the strategies used on the Expedition can be successfully employed with the
Excursion. Therefore, we believe that by using technologies and control strategies similar to
what will be used by LDVs and LDTs, combined with larger catalysts, MDPVs will be able to
meet our Tier 2 emission standards.
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.
IV-32
-------
Chapter IV: Technological Feasibility
A number of potential techniques are being developed to control NOx emissions when
excess air is 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.
As part of the ongoing efforts in developing GDI technology, substantial progress is being made
in the application of aftertreatment controls. While much remains to be done both in lowering
engine-out emission levels and in aftertreatment development, we believe that the bin structure in
the Tier 2 standards is sufficient to allow the introduction of GDI engines.
2. CO Emissions from Gasoline Fueled Vehicles
EPA is only requiring tighter CO emission standards for LDT2s, LDT3s, LDT4s and
MDPVs. Basically, CO emissions from these vehicles must be reduced to the levels now
required for LDVs and LDTIs under the NLEV program. Also, LDVs and LDTls must comply
with the NLEV CO standards over a slightly longer useful life of 120,000 miles instead of the
current useful life of 100,000 miles.
Compliance with the Tier 2 CO emission standards should not be difficult given
compliance with the Tier 2 NMOG standards. The control of both pollutants utilizes much of
the same technology and the Tier 2 NMOG standards are the more stringent of the two sets of
standards. In addition, the change in test weight from "adjusted loaded vehicle weight" to
"loaded vehicle weight" will make it easier to meet the Tier 2 CO emission standards. The
following table IV-12 summarizes CO emissions from vehicles certified to the LEV standards in
California.
Table IV-12. 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 Tier 2 CO
standard of 4.2 g/mi. While CO emissions from LDT3s and LDT4s are more than half the 4.2
g/mi standard, the current LEV standards for these vehicles is more than twice the 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 Tier 2 CO standard should not add any additional burden to
manufacturers relative to compliance with the NMOG and NOx standards.
3. Formaldehyde Emissions from Gasoline Fueled Vehicles
IV-33
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
EPA is only requiring tighter formaldehyde emission standards for LDT2s, LDT3s,
LDT4s and MDPVs. 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 Tier 2 formaldehyde emission
standards should not be difficult given compliance with the Tier 2 NMOG standards. The
control of both pollutants utilizes the same technology and the Tier 2 NMOG standards are the
more stringent of the two sets of standards. Table IV-13, below, summarizes formaldehyde
emissions from vehicles certified to the LEV standards in California.
Table IV-13. 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 Tier 2 formaldehyde standard of 0.018 g/mi. Thus, compliance with the
Tier 2 formaldehyde standard should not add any additional burden to manufacturers relative to
compliance with the NMOG and NOx standards.
4. Evaporative Emissions
The Tier 2 standards 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 Tier 2 standards 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
IV-34
-------
Chapter IV: Technological Feasibility
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 of alcohols in
gasoline on permeability of evaporative components, hoses and seals.
5. Diesel Vehicles
As discussed earlier, the Tier 2 standards are intended to be "fuel neutral." In this
document, we establish that the Tier 2 standards are technologically feasible and cost-effective
for LDVs and LDTs overall. Under the principal of fuel neutrality, all cars and light trucks,
including those using diesel engines, will be required to meet the Tier 2 standards. Contrary to
some of the comments received on our proposal, given that the overwhelming majority of
vehicles in these classes are gasoline-fueled, we do not believe it is appropriate to provide less
stringent standards for diesel-fueled vehicles. Manufacturers of LDVs and LDTs today provide
consumers with a wide choice of vehicles that are overwhelmingly gasoline-fueled. Less
stringent standards for diesels would create provisions that could undermine the emission
reductions expected from this program, especially given the expectation that some manufacturers
are hoping to greatly increase their diesel sales.
As with gasoline engines, manufacturers of diesels have made abundant progress 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 actually met with only engine modifications. Indeed, emissions and performance
of lighter diesel engine are rapidly approaching the characteristics of gasoline engines, while
retaining the durability and fuel economy advantages that diesels enjoy. Against this background
of continuing progress, we believe that the technological improvements that would be needed
could be made in the time that would be available before diesels would have to meet the new
Tier 2 standards.
Manufacturers may take advantage of the flexibilities in our Tier 2 emission standards to
IV-35
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
delay the need for diesel LDVs and LDTs to meet the final Tier 2 levels until late in the phase-in
period (as late as 2007 for LDVs/LLDTs and 2009 for HLDTs), giving manufacturers a relatively
large amount of leadtime. In a recent public statement, Cummins Engine Company has indicated
that the interim Tier 2 standards in effect for vehicles and trucks in the early years of the Tier 2
program are feasible for diesel equipped models through further development of currently
available engine technology.2 We also believe that standards can be met through the use of
existing technologies, such as cooled EGR where it currently is not used, moderate amounts of
fuel injection timing retard, and perhaps limited use of lean NOx catalysts and/or diesel oxidation
catalysts, as well as particulate traps.
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 Tier
2 standards for these pollutants. Engine-out emissions continue to be reduced. The following
are some examples of technologies and strategies that can be used to reduce engine-out
emissions.
One of the most important control strategies for the reduction of engine-out NOx
emissions is the the addition of cooled exhaust gas recirculation (cooled EGR). This method
recirculates a portion of the exhaust back to the intake manifold where it is drawn into the
combustion chamber. The resulting mixture of fresh air and exhaust products has a lower
concentration of O2 than fresh air alone. The lower concentration of O2 in the combustion
chamber results in lower O2 partial pressure which lowers its propensity to oxidize N2 to NO and
NO2 (NOx) during the combustion process.
More sophisticated electronic control systems will be necessary to control the EGR
system. EGR control algorithms will require additional engine condition information, possibly
including mass air flow, oxygen, NOx, or EGR valve position sensors. These inputs will be
necessary to control the EGR rate via an EGR valve or possibly a variable geometry turbocharger
(VGT). These turbochargers will also require a sophisticated control algorithm to take advantage
of their transient response, EGR pumping, and air flow control characteristics. In addition, the
turbomachinery used with EGR will likely be pushed near the limits of its capability, and the
engine's electronic control module (ECM — the engine's control computer) will need to ensure
the limits of the hardware are not exceeded.
While reductions in "engine-out" emissions may continue to be made, increasing
emphasis is being placed on various aftertreatment devices for diesels. We believe that the use of
aftertreatment devices alone will allow diesels to comply with the Tier 2 standards for NOx and
PM.
2"Cummins Sees Diesel Feasible for Early Years of Tier 2". Hart Diesel Fuel News, Sept. 20, 1999, p.2.
IV-36
-------
Chapter IV: Technological Feasibility
ForNOx emissions, potential aftertreatment technologies include leanNOx catalysts,
NOx adsorbers and selective catalytic reduction (SCR). Lean NOx catalysts are still under
development, but generally appear capable of reducing NOx emissions by about 15-30 percent.
This efficiency is not likely to be sufficient to enable compliance with the final Tier 2 standards,
but it could be used to meet the interim standards that would begin in 2004.
NOx adsorbers appear capable of reaching efficiency levels as high as 90 percent.
Efficiency in this range is likely to be sufficient to enable compliance with the proposed Tier 2
standards. NOx adsorbers temporarily store the NOx and thus the engine must be run
periodically for a brief time with excess fuel, so that the stored NOx can be released and
converted to nitrogen and oxygen using a conventional three-way catalyst, like that used on
current gasoline vehicles.
There is currently a substantial amount of development work being directed at NOx
adsorber technology. While there are technical hurdles to be overcome, progress is continuing
and it is our judgement that the technology should be available by the time it would be needed for
the final Tier 2 standards.
One serious concern with current NOx adsorbers is that they are quickly poisoned by
sulfur in the fuel. Some manufacturers have strongly emphasized their belief that, in order to
meet the final Tier 2 levels, low sulfur diesel fuel would also be required to mitigate or prevent
this poisoning problem. In its comments on the NPRM, Navistar indicated that the Tier 2
standards may be achievable given low sulfur fuel and other programmatic changes such as those
included in this Final Rule. Navistar has also been quoted publically as describing the Tier 2
standards as "challenging but achievable" given appropriate low sulfur fuel.3
One solution would be to reduce sulfur to very low levels. Another solution would be to
reduce sulfur somewhere below current levels and develop a way to periodically remove the
sulfur from the adsorber. In any event, this technique, if used, would also require low sulfur
diesel fuel. We will be issuing a Notice of Proposed Rulemaking in the near future intended to
reduce sulfur in highway diesel fuel on a parallel path with today's final rule as a step to enable
the technology most likely to be used to meet the Tier 2 standards.
SCR has been demonstrated commercially on stationary diesel engines and can reduce
NOx emissions by 80-90 percent. This efficiency would be sufficient to enable compliance with
the proposed Tier 2 standards. However, SCR requires that the chemical urea be injected into the
exhaust before the catalyst to assist in the destruction of NOx. The urea must be injected at very
precise rates, which is difficult to achieve with an on-highway engine, because of widely varying
engine operating conditions. Otherwise, emissions of ammonia, which have a very objectionable
3Harts Diesel Fuel News, August 9, 1999, p4.
IV-37
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
odor, can occur. Substantial amounts of urea are required, meaning that vehicle owners would
have to replenish their vehicles' supply of urea frequently, possibly as often as every fill-up of
fuel. As the engine and vehicle will operate satisfactorily without the urea (only NOx emissions
would be affected), some mechanism would be needed to ensure that vehicle owners maintained
their supply of urea. Otherwise, little NOx emission reduction would be expected in-use.
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 15-30 percent. By itself, the oxidation catalyst is not likely to be
sufficient to enable compliance with the 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 a 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 Tier 2 PM standards.
In summary, we believe that the structure of our final program, including the available
bins and phase-in periods, will allow the orderly development of clean diesel engine
technologies. We believe that the interim standards are feasible for diesel LDV/LDTs, within the
bin structure of this rule and without further reductions in diesel fuel sulfur levels. And, as
indicated earlier, at least one major diesel engine manufacturer (Cummins) has publically agreed
with this assessment. We further believe that in the long-term, the final standards will be within
reach for diesel-fueled vehicles in combination with appropriate changes to diesel fuel to
facilitate aftertreatment technologies. Manufacturers have argued that low sulfur diesel fuel will
be required to permit diesels to meet the final Tier 2 standards, and we agree. Once again, at
least one major manufacturer (Navistar) has indicated its belief that the final Tier 2 standards
IV-38
-------
Chapter IV: Technological Feasibility
may be achievable for diesel engines with low sulfur diesel fuel.
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
compounds are broken up into smaller compounds, or cracked, and the embedded sulfur ends up
in gasoline. Thus, the refinery units which convert the heavy parts of crude oil into gasoline are
the units most responsible for putting sulfur into gasoline.
The fluidized catalytic cracker (FCC) unit is 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 streams produced in 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 (much of it is hydrotreated and processed further before blending
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 hydrocrackate, and
isomerate. Oxygenates which are blended into gasoline usually have very little or no sulfur,
IV-39
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 simplistic
conclusion which could be reached would be that refiners could simply shut down these units in
their refineries 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
the volume of crude oil is composed of heavy compounds which have no end use, and thus is not
usable without processing by these units. 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 amount of sulfur in gasoline for determining the most cost-
effective sulfur removal methods for our cost analysis, and for developing the gasoline sulfur
phase-in requirements. For the NPRM, we used a mixture of gasoline sulfur data from the
American Petroleum Institute (API) and the National Petrochemical Refiners Association
(NPRA) survey which was conducted during the Summer of 1996,6 and 1995, 1996 and 1997
gasoline sulfur data from the RFG data base. To enable our cost analysis, we compiled the data
by 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)
IV-40
-------
Chapter IV: Technological Feasibility
Petroleum Administration for Defense (PAD) Districts
Figure IV-4. Map of U.S. Petroleum Administrative Districts for Defense
The API/NPRA study showed that the gasoline sulfur, outside of California, averaged 340
ppm during the Summer of 1996. 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. Another possible reason for the discrepancy is a difference in the specific
refiners included in the two sets of data. Some refiners did not participate in the API/NPRA
survey (especially in PADDs 1 and 5), while data handling complications precluded the inclusion
of gasoline sulfur data from some refiners from being reported in the RFG data base. 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 in lieu of the API/NPRA survey information.
For the Final Rule, we analyzed the 1998 RFG fuel quality reports to determine the
gasoline sulfur levels for 1998. The analysis revealed that during 1998 national gasoline sulfur
levels were significantly lower than the sulfur levels in 1997 and previous years. The most likely
IV-41
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
reason why the sulfur levels are lower in 1998 is related to the certification requirements for both
RFG and conventional gasoline which changed in 1998. Prior to 1998, RFG was certified using
the EPA Simple Model which only required that sulfur not increase relative to each refiner's
1990 baseline level. Regarding conventional gasoline, sulfur levels were simply prevented from
increasing by more than 25 percent over the refiner's 1990 baseline level. Starting in 1998,
refiners had to use the EPA Complex Model to certify both fuels, which included sulfur's impact
on emissions from Tier 0 vehicles. RFG sulfur levels were also capped at 500 ppm starting in
1998. Finally, RFG NOx emission performance began to be determined relative to the Clean Air
Act baseline fuel, which is much cleaner than many refiners' baseline levels.
Since the 1998 data is the best estimate of where refiners will start from in meeting the
new sulfur standards, we recalculated our PADD and national average gasoline pool sulfur levels
for estimating gasoline desulfurization cost and the phase-in of the low sulfur program based on
the new data. Table IV-14 below summarizes the U.S. sulfur levels by PADD, and for the
country as a whole used in the NPRM and for this analysis. Because California has its own low
sulfur gasoline program, gasoline produced there was excluded from consideration in this
analysis.
Table IV-14. Estimated Average Sulfur Levels by PADD and for the Nation.
NPRM
FRM
PADD 1
215
189
PADD 2
338
276
PADD 3
307
288
PADD 4
265
282
PADD 5
OC*
506
301
U.S.
Avg. *
305
268
* Outside of California
It is important to note that the gasoline sulfur values reported in Table IV-14 are
estimates 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 Final Gasoline Sulfur Standards
The feasibility of meeting the final 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
IV-42
-------
Chapter IV: Technological Feasibility
the final standard. The second way is to determine if refiners are already demonstrating that they
can meet a low sulfur gasoline standard similar to that contained in this final rule. Evidence that
a large number of refineries having various configurations are already meeting a stringent
gasoline sulfur program is a more compelling example of feasibility since the technology is
clearly already 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 in the U.S. today. Furthermore, the State of
California has established tighter gasoline sulfur standards. The average sulfur standard is 15
ppm, with a 60 ppm cap, which takes effect starting January 1, 2003. The cap decreases to 30
ppm starting January 1, 2005.10 Thus, the 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 30 ppm sulfur.11
4. Meeting a Low Sulfur Gasoline Standard
a. Background
The methodology that refiners would use 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-5,
below.)
The amount of excess refinery desulfurization equipment on hand
IV-43
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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-44
-------
Chapter IV: Technological Feasibility
Natural
Gas
Vacuum Tower
Coker
*" Fuel Gas
LPG
Gasoline
Aromatics
Kerosine
Jet Fuel
Onroad
Diesel
Off-roadDiesel
Fuel Oil
*- Resid
Coke
Figure IV-5. Diagram of a Typical Complex Refinery
IV-45
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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, 50 ppm) will probably not
need to invest in expensive capital. These refineries have very low sulfur levels due to one or
more of a number of possible reasons. For example, some of these refiners may not have certain
refining units, such as a fluidized catalytic cracker (FCC) unit, or a coker, which convert heavy
boiling stocks to gasoline (Figure IV-5 shows 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 average sulfur levels below 30
ppm would not have to do anything. On the other hand, those refineries with sulfur levels above
30 ppm but below some level, such as 50 ppm, could probably meet the standard employing
operational changes only and avoid making capital investments. There are only 3 refineries in
this category, representing a total of 2.8% of non-California gasoline production. One such
refinery does not have a FCC unit. However, it does have a coker, which produces less gasoline
volume than a FCC unit, but the sulfur level of this gasoline can be quite high. This refinery
also has extensive hydrotreating capacity. The other two refineries in this situation have FCC
units, but also have utilize FCC feed hydrotreating. Thus, all 3 refineries have the capability to
desulfurize nearly all of the gasoline components being produced in their refineries. These
refiners should be able to meet the 30 ppm sulfur standard by running existing desulfurization
units more severely, or by increasing the volume of blendstock sent to these units. If necessary,
more active desulfurizing catalysts could be utilized. Refiners also have FCC additives available
to them which could allow them to reduce their FCC gasoline sulfur level by 15 to 35 percent,
which should be more than sufficient for the two refineries with FCC units.12 Two of these
refineries have average sulfur levels of 40 ppm or less. Thus, they should be able to meet the 80
ppm cap with no change in average sulfur level. They could buy credits (or transfer them from
other refineries in their corporation, in the case of multi-refinery refiners) to meet the 30 ppm
average standard. Because of their low current sulfur levels, these refineries have until at least
2006, and possibly later in order to implement these strategies. Given the wide range of options
available, these 3 refineries should be able to meet the 30 ppm standard without building a
desulfurization unit.
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
IV-46
-------
Chapter IV: Technological Feasibility
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.
b. FCC Feed Hydrotreating
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 the highest 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 (i.e., 150,000-200,000 BPCD), the capital costs may exceed $100
million. Because of the higher temperatures and pressures involved, utility costs are expensive
relative to other forms of hydrotreating explained below. Another justification for this approach
is that it allows refiners to switch to a heavier, more sour crude oil. These crude oils are less
expensive per barrel and can offset the increased utility cost of the FCC desulfurization unit,
providing that the combination 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. 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 FCC naphtha, light
cycle oil 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.
c. FCC Gasoline Hydrotreating
A less expensive alternative for reducing FCC gasoline sulfur levels 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 50 to 60 percent of the feed to the
FCC unit ends up as gasoline. The unit is often smaller than that as refiners, which choose to use
a fixed bed hydrotreater, could choose to only treat the heavier, higher sulfur portion of that
stream with hydrotreating, and then treat the lighter fraction with catalytic extractive
desulfurization. FCC gasoline hydrotreaters operate at lower temperatures and pressures than
FCC feed hydrotreating 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.
IV-47
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
One drawback of this desufurization methodology is that the octane value and/or some of
the gasoline yield may be lost depending on the process used for desulfurization. Octane loss can
occur by the saturation of high octane olefms which are produced by the FCC unit. Most of the
olefins are contained in the lighter fraction of FCC naphtha. With increased olefin saturation
comes increased hydrogen consumed. There can be a loss in the gasoline yield caused by mild
cracking which breaks some of the gasoline components into smaller fractions which are too
light for blending into gasoline. If there is octane loss, it must be made up by additional octane
production by other units in the refinery or by oxygenate addition, and any 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 by technologies which were recently developed. These processes preserve much of the
octane and gasoline yield because they were designed for treating gasoline blendstocks. Octane
is preserved because their catalysts are specially designed to either avoid saturating olefins, or if
the process does saturate olefins, it causes other reactions to occur which improves the octane of
the gasoline lost through desulfurization. These processes may also operate at less severe
conditions than conventional hydrotreaters which preserves yield compared to conventional
hydrotreating processes. The less severe conditions lowers the capital and operating costs for
this process. Typical capital cost for these newer desulfurization technologies ranges from $20 to
$40 million for a medium to large sized refinery. The lower operating costs arise out of the
reduced utility requirements (e.g., process heat, electricity), octane losses and hydrogen
consumption. For example, because these processes are less severe, there is much less or no
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.
The lower capital and operating costs of these newer FCC gasoline hydrotreaters are
important incentives for refiners to choose this desulfurization methodology over conventional
FCC gasoline or FCC feed hydrotreating. That some refiners would use this newer
desulfurization technology is reinforced by conversations with refiners and licensors of
hydrotreating hardware. For this reason, we are assuming that many refiners will choose to use
the more recently developed FCC gasoline hydrotreating technologies for meeting the gasoline
sulfur standard.
For the NPRM we presumed that refiners would choose either of two of the more
improved FCC gasoline desulfurization processes, CDTech or Mobil Oil Octgain 220. However,
we received a number of comments from the oil industry that they generally require that refining
processes be commercially demonstrated, for at least two years, before choosing to use these
technologies. Since the CDTech and Mobil Oil have not been commercially demonstrated that
length of time yet, we expanded our list of technolgies upon which we we are basing our rule to
currently proven FCC gasoline desulfurization technologies. Furthermore, we learned of another
class of FCC gasoline desulfurization technologies which are now commercially available.
IV-48
-------
Chapter IV: Technological Feasibility
These newest desulfurization technologies work by adsorption and work more efficiently than
hydrotreating desulfurization technologies. We are including these technologies in our analysis
as well.
i. Proven Desulfurization Technologies
We know of three commercially proven FCC gasoline desulfurization technologies.
These are Mobil Oil Octgain 125, Exxon Scanfming, and IFF Prime G. These are all fixed bed
desulfurization technologies, so they function similar to each other. These processes are called
fixed bed because the catalyst resides in a fixed bed reactor.13 The high sulfur gasoline
blendstock is heated to a high temperature (on the order of 600 degrees Fahrenheit) and pumped
to a high pressure, to maintain the stream as a liquid, 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, depending on the process, some, most or
all of the olefin compounds which are present in the cracked stream are saturated which increases
the amount of octane lost and hydrogen consumed. The difference between these and
conventional hydrotreating processes is that these technologies have a way for either minimizing
the loss in octane or compensating for it, either by minimizing the loss of olefins, or by
recovering the loss octane through octane producing reactions. The catalyst may cause yield loss
through cracking of some of the petroleum compounds. After the reactor, the gaseous
compounds, which include unreacted hydrogen, hydrogen sulfide, and any light end petroleum
compounds which may have been produced in the reactor by cracking reactions, are separated
from the liquid compounds. The hydrogen sulfide must be stripped out from the other
compounds and then converted to elemental sulfur in a separate sulfur recovery unit, and the
recovered sulfur is then sold. If there is enough hydrogen and it can be economically recovered,
it is separated from the remaining hydrocarbon stream and recycled. Otherwise, it would
probably be burned with the light hydrocarbons as fuel gas.
Each of these commercially proven desulfurization technologies are a little different. The
Octgain 125 process saturates all the olefins, but recovers the lost octane through isomerization
and alkylation.14 It needs to be run at fairly severe conditions for it to recover octane, so this
process is more appropriate for refiners with higher sulfur levels which requires severe
hydrotreating to reach the sulfur target. While octane loss can be eliminated with the proper
operating conditions, yield loss can be significant. It has been commercially demonstrated at
Mobil's refinery in Joliet, Illinois.
Exxon's Scanfming process preserves octane by saturating very few olefins, however, at
severe operating conditions for higher levels of desulfurization, octane loss can be high. The
Scanfming catalyst causes very little yield loss. This process has been demonstrated for a total of
over 4 years in two of Exxon's refineries.15
IV-49
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
IFP's (Intitule Francais du Petrole) Prime G desulfurization process largely preserves
olefins as its strategy for diminishing octane loss.1617 Like Scanfining, Prime G is less severe
and cracks the petroleum compounds less resulting in less yield loss. Prime G has been
commercially demonstrated for over 7 years in two U.S. refineries, and in an Asian refinery.
ii. Improved Gasoline Desulfurization Technology
Consistent with the NPRM, we are placing CDTech and Mobil Oil Octgain 220 processes
in the same category called improved desulfurization technologies, and these technologies have
not yet had significant commercial experience. Mobil Octgain will be discussed first since the
process is similar to the commercially proven hydrotreating technologies discussed above. Like
the commercially proven desulfurization technologies, the Mobil 220 process uses a fixed bed for
its catalyst.18 Octgain 220 preserves most of the olefins and recovers lost octane through
isomerization reactions.19 20 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. For high levels of desulfurization, yield and octane loss increase
significantly for this process so Mobil recommends that refiners use the 125 process for these
desulfurization cases. Mobil loaded the 220 catalyst into their Joliet hydrotreater during March
of 1999, so the process has some commercial experience. In addition, Mobil Oil has signed a
license agreement with a refinery outside the U.S., so another Octgain unit will be installed soon.
The CDTECH process is significantly different from either conventional hydrotreating or
Octgain, and it is a little more complex to describe. The CDTECH process utilizes catalytic
distillation.2122 23 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
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 most important portion of the CDTech desulfurization process is a set of two
distillation columns loaded with desulfurization catalyst in a packed structure. The first vessel,
called CDHydro, treats the lighter compounds of FCC gasoline and separates the heavier portion
IV-50
-------
Chapter IV: Technological Feasibility
of the FCC gasoline for treatment in the second column. The second column, called CDHDS,
removes the sulfur from the heavier compounds of FCC gasoline. All of the FCC gasoline is fed
to the CDHydro column. The 5 and 6 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 are 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 diolefins). 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 some of the
olefms which increases the octane of this stream by about three octane numbers, and few of the
olefms are saturated to degrade this octane gain.
The seven-carbon and heavier petroleum compounds leave the bottom of the CDHydro
unit and are fed into the CDHDS column. There, the heavier compounds head down the column,
and the lighter compounds head up. Both sections of the CDHDS column have catalyst loaded
into them which serve as hydrotreating reaction zones. Similar to how hydrogen is fed to the
CDHydro column, hydrogen is fed to the bottom of the CDHDS column.
The temperature and pressure of the CDTech process columns are lower than fixed bed
hydrotreating processes, particularly in the upper section of the distillation column, which is
where most of the olefms end up. 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 is very familiar to refiners. Every refinery has distillation in its refinery, thus,
refiners are very skilled in its application.
CDTech has numerous CDHydro units in operation, but CDHDS units have not yet been
installed in refineries. Thus, one portion of the CDTech process is commercially proven, while
the other portion is not. A CDHDS unit is expected to be operational in the Motiva refinery in
Port Arthur, Texas starting March of 2000. Additionally, a combined CDHydro/HDS unit is
expected to be operational in North America in October of 2000, and another license aggrement
has been signed for an installation in Europe. An installation of an HDS unit is planned for the
IV-51
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Transamerican refinery in Louisiana, however, that refinery is currently shutdown and the startup
date of the refinery and the planned CDHDS unit is unclear.
The relative cost of the improved desulfurization technologies and the commercially
proven technologies depends on the specific situation faced by a refiner. For most refiners, the
more recent, improved desulfurization technologies are projected to be less expensive. However,
as we point out above, Mobil Oil recommends that the Octgain 125 process be used for treating
the FCC naphtha from heavy, high sulfur crude oils. In our analysis, we are estimating the
desulfurization cost of average refineries meeting the gasoline sulfur standard, and Mobil Oil
would probably recommend that their Octgain 220 process be used for this application.
However, Scanfming and Prime G processes are well suited desulfurizing technologies for
average refineries as well. Thus, when we use costs developed for improved technologies, these
costs could be representative for some of the proven technologies as well.
iii. Adsorption Desulfurization Technologies
Black and Veatch Pritchard Inc. and Phillips Petroleum Co. have announced the
commercial availability of adsorption desulfurization technologies (i.e., they are prepared to
design and license this technology to refiners). We believe that similar adsorption technologies
may be available soon from UOP and a major refiner as well. These technologies use the
chemical process of adsorption, instead of hydrotreating, as the principal methodology for the
removal of sulfur from gasoline. Adsorption has the benefit of operating at much lower pressure
and temperatures, which lowers operating costs, and potentially can lower capital costs as well.
Each of these desulfurization processes operates differently.
The Black and Veatch process, named IRVAD, adsorbs heteroatom-containing petroleum
compounds, which are sulfur, nitrogen and oxygen containing petroleum compounds, onto their
adsorption catalyst.24 The catalyst is alumina-based and manufactured by Alcoa Industrial
Chemicals. The catalyst is fluidized and continuesly removed and regenerated, using hydrogen,
in a second column. The regenerated catalyst is then recycled back into the reactor vessel at the
rate which it is being removed. In the regeneration column, the adsorbed heteroatom containing
petroleum compounds, which is about 4 percent of the petroleum stream being treated, are
removed from the catalyst. Since the hydrogen used in the regeneration column is for scavenging
the petroleum compounds off of the catalyst and it is not reacting with the petroleum, hydrogen
loss is considered by Black and Veatch to be negligable. According to Black and Veatch
process operations information, the treated FCC gasoline is 2 octane numbers higher than the
untreated FCC gasoline.
This high sulfur stream, which contains about 1 percent by weight of sulfur (10,000 ppm),
must then be treated for reblending with gasoline. This stream cannot just be blended to offroad
diesel since it contains many volatile petroleum compounds. Black and Veatch surmises that
IV-52
-------
Chapter IV: Technological Feasibility
most refiners would treat the stream in an existing diesel hydrotreater. After hydrotreating, the
gasoline substream, which is now blended in with desulfurized diesel, would have to be
separated either by the existing stripping column which hydrotreating processes have for
separating the light ends produced in these facilities, or in a splitting column which may already
be installed after the hydrotreater. If the stripping column is insufficient for the volume of
gasoline which would have to be hydrotreated, and there is not an existing splitting column, then
a splitting column would have to be added, or the stripping section would have to be enlarged.
At the time that this document was being drafted, Black and Veatch had not yet signed a license
agreement to install one of their units, although two refiners who own small refineries have given
Black and Veatch verbal agreements that they will install IRVAD units in their refineries.
The adsorption process by Phillips, called S-Zorb, is similar to the Black and Veatch
process in that two separate columns are needed and the catalyst is constantly moving from the
reactor vessel to the regeneration column, and back again.25 However, beyond that, the processes
are very different. The untreated FCC naphtha and hydrogen are fed to the reaction vessel where
the Phillips catalyst adsorbs the sulfur-containing petroleum compounds onto the catalyst.
However, the catalyst also catalytically removes the sulfur from the petroleum compound so the
petroleum compound which contained the sulfur never leaves the reaction vessel. Instead, the
catalyst which begins to accumulate the removed sulfur, is transfered over to the regeneration
column on a continual basis where the sulfur is removed from the catalyst using hydrogen as the
scavenging compound. Then the hydrogen disulfude is converted to sulfur dioxide and sent to
the sulfur recovery unit. Since, the petroleum compounds are desulfurized in the main reactor,
there is no need to hydrotreat any high sulfur stream. However, because the process still relies
upon catalytic processing in the presence of hydrogen, there is some saturation of olefms, with a
commensurate reduction in octane. Currently, there are no S-Zorb units operating, however,
Phillips is working rapidly to install a 6000 barrel per day unit at its Borger, Texas refinery, and
plans to have it operating by the first quarter of 2001.
5. Expected Desulfurization Technology to be Used by Refiners
With the promulgation of the Federal sulfur control program, which begins to phase-in in
2004, 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 install
capital to meet the sulfur standard. Arguably, refiners would try to minimize the cost to their
business. As stated above, the adsoprtion gasoline desulfurization technologies seem to be the
lowest cost technologies based on our analysis of average refineries in each PADD, followed by
the improved desulfurization technologies, such as CDTECH and Mobil Oil Octgain. However,
several 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
IV-53
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
away from using the lower cost adsorption and improved desulfurization technologies, to
applying proven desulfurization technologies.
While there is a concern now on the part of some refiners about using the adsorption and
improved, but not commercially tested desulfurization technologies, we believe that much of this
concern will dissipate in the near future. All these processes are expected to be operating
commercially in the next two years. Mobil Oil has already installed its Octgain 220 catalyst in
the hydrotreater at its Joliet, Illinois refinery and is accruing commercial experience with that
technology. CDTech has years of accumulated service with its CDHydro unit, but not with the
CDHDS unit. The CDHDS unit and the adsorption technologies are expected to be installed in
refineries in the year 2000 and 2001.
We have clear examples specific to these technologies that refiners do not need to
observe a certain technology operating in a refinery before they will choose to use that
technology. For example, no CDHDS units, nor complete CDHydro/CDHDS units, are up and
operating, however, a handful of refiners have committed to installing the process in their
refineries. Thus, these refiners have been willing to commit to that technology without observing
it operating commercially in a refinery. We believe that it may be more difficult to commit to the
adsorption technologies as easily, as they are somewhat different from conventional
desulfurization technologies. For developing the cost of the desulfurization program, we need to
project the types of desulfurization units which will be installed during the years that the gasoline
sulfur program is phased-in. Our projections are summarized below in Table IV-15.
Table IV-15. Projected Use of Desulfurization Technology Types by
Refiners During the Phase-in Period
Year
2004
2005
2006
2007 & 2008
Mix of Technology Types Used
1/2 Proven, 1/2 Improved
3/4 Improved, 1/4 Adsorbent
1/2 Improved, 1/2 Adsorbent
1/4 Improved, 3/4 Adsorbent
Prior to 2004, we project that new desulfurization units will fall into two broad
categories: early units being installed by refiners who desire to generate credits and possibly use
low sulfur as a marketing factor and demonstration units. We project that the former will
primarily utilize proven technology (90 percent proven, 10 percent improved). On the other
hand, the demonstration units will all utilize improved or adsorbent technology, as there is no
need to demonstrate the proven technology. On a volumetric basis, we project a breakdown of
IV-54
-------
Chapter IV: Technological Feasibility
50 percent proven, 25 percent improved, and 25 percent adsorbent.
6. Feasibility for a Low Gasoline Sulfur Standard in 2004
The final gasoline sulfur control program provides a full four years before the first sulfur
standard must be met starting on January 1, 2004. As discussed below, not all refiners will have
to modify their operations on this date. Thus, more than four years leadtime is available for
many refiners (i.e., those with low current sulfur levels). This is consistent with requests
received from API and NPRA, as well as from a number of refiners, in their comments to the
proposal, that at least four years be provided prior to the start date of the program.
The following table breaks down the steps involved in constructing new refining
equipment and our projection of the time necessary for each step and the entire process. The
reader is also referred to the Draft RIA for some additional detail in the development of these
estimates.
Table IV-16. 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
Naphtha/Gasoline Hydrotreating
Time for
Individual Step
0.5-1.0*
0.5
0.25-1.0
0.5-0.75
0.75-1.0
0.25
Cumulative
Time
0.5
1.0
1.25-2.0
1.5-2.25
2.0-3.0
2.25-3.25
More Major Refinery
Modification (e.g., FCC Feed
Hydrotreating)
Time for
Individual Step
0.5-1.0*
0.5-0.75
0.25-1.0
0.5-1.0
1.0-1.5
0.25
Cumulative
Time
0.5
1.0-1.25
1.25-2.0
1.5-2.25
2.5-3.5
2.75-3.75
* 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
IV-55
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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.).
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. For
example, refiners have been sending samples of their FCC naphtha to the vendors of
desulfurization technology for some time to determine exactly how well each technology will
perform in their specific refinery. 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 not have been demonstrated in actual refinery applications by the end of 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. Thus, the fact that less than four
years are actually needed to design and build gasoline desulfurization equipment once the
technology is selected and some refiners do not need this equipment until 2005 or beyond allows
refiners to observe the performance of these new, potentially lower cost technologies and make a
more informed selection in a year or two. We believe that at a minimum, refiners should have 6
months after the final rule to assess their situation with respect to the final sulfur control program
and select their technological solution.
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.
All of this equipment is already common to refineries. Aside from the new adsorption
technologies, all gasoline desulfurzation units are 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.26
IV-56
-------
Chapter IV: Technological Feasibility
For example, CDTECH estimates that 10-12 weeks are needed for the basic process design of
their equipment. Discussions with other contractors have indicated that 4-5 months is usually
more than sufficient to complete the process design.
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
significantly reduce a variety of emissions from the FCC unit, particularly SOx and toxics.
However, FCC hydrotreating can increase NOx emissions relative to naphtha hydrotreating due
to processing more hydrocarbons and the greater temperature and pressure involved. The
emission reductions should ease permitting and compliance with future MACT standards for
toxics, while the NOx emission increase would have to be addressed in ozone nonattainment
areas subject to NOx offsets. We have allocated up to 3 months more for the process design of
these more major modifications.
Regarding permitting, EPA has held a number of discussions with state/local permitting
agencies, environmental organizations and refiners. EPA is committed to streamlining the
process of obtaining permits. One step in this process would be to identify soon after the final
rule the technologies that EPA believes would constitute Best Available Control Technology
(BACT) and the Lowest Achievable Emission Rate (LAER). This would inform both states and
refiners of the types of refinery emission control technology EPA believes would meet the BACT
and LAER requirements The lower limit of 3 months is typical for obtaining a minor source
permit. States such as Texas typically issue permits in four months on average, including major
NSR permits. One year for a permit would represent a very protracted process which should be
avoidable if refineries are working closely with the states to resolve any issues that may arise
during the permitting process. Nonetheless, this time period was included above in order to
identify the worse case situation which may occur. EPA's permit streamlining approaches
should provide opportunities to shorten this time period even further.
Based on discussions with contractors, design and construction of naphtha hydrotreaters
typically requires about 18 to 20 months, while about two years is required for more major
equipment like FCC feed hydrotreaters. If all refiners attempted to construct their new
equipment at the same time, limited capacity of vendors who manufacture the pressure vessels
and compressors could extend these times to 24-30 and 36 months. However, as described in the
next section, we have explicitly designed the sulfur control program to spread out construction.
Thus, the manufacturing capability of these equipment vendors should not be over-taxed.
Several different fuel programs already in place suggest that a stringent gasoline
desulfurization program can be phased-in in four or less years. 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.27 Similarly, the Phase U California
Reformulated Gasoline Program was promulgated in November 1991 and took effect about 41A>
IV-57
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
years after promulgation.28 However, in addition to a stringent sulfur control standard, refiners
also had to meet stringent controls for aromatics, olefms, Reid vapor pressure, and distillation
index. 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
1993.29 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 fixed bed gasoline desulfurization units, such as
Octgain. Refiners' volume of onroad diesel fuel is generally less than their FCC gasoline
volume. However, leadtime is generally not a strong function of unit capacity, at least for
capacity differences of this magnitude.
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.
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.30 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. Under one set of
provisions, the smallest refiners will be given until 2008 to meet the 30 ppm sulfur standard.
Under another set of provisions, refiners supplying fuel to a number of western states will be
given until 2007 to meet the 30 ppm standard. This provision most directly affects the relatively
small refineries located in these western states. This additional leadtime should provide not only
enough time for these small refiners to construct equipment, but to also allow these refiners more
time to select the most advantageous desulfurization technology. This additional time for
technology selection will help to compensate for the diseconomy of scale inherent with adding
IV-58
-------
Chapter IV: Technological Feasibility
equipment to a small refinery.
7. Phase In of Compliance with the Proposed Sulfur Standards and Early
Credit Generation
As stated earlier in this chapter, the sulfur content of gasoline in 1998 averaged about 268
ppm, well above the eventual refinery average standard of 30 ppm. The vast majority of the
sulfur in gasoline comes from a single gasoline blending component, FCC or cat naphtha. As
projected above, for most refiners, removing enough sulfur from FCC naphtha to meet a 30 ppm
standard on average will require the construction of a naphtha desulfurization unit.
There are two reasons to evaluate the timing of the construction of these desulfurization
units. One, the type of desulfurization technology employed must be determined at the beginning
of the design and construction process. A desulfurization unit whose design and construction
begins prior to the time when a specific desulfurization technology is deemed to be commercially
proven and ready for broad application will generally not be available for use in that specific
refinery. As discussed above, we have identified three groups of desulfurization technology
ranging from those already commercially proven to those which will generally not be available
for use (not design and construction) until 2005. Therefore, if the applicable gasoline sulfur
standards require a refinery to install a hydrotreating unit in 2004, a refiner may consider its
options for desulfurization technology to be more limited than if he could delay the installation
until 2005 or later.
The second reason to project the timing of new desulfurization units is to evaluate the
ability of the design and construction industry to fulfill the needs of the refining industry. Here,
the type of unit is less important (though not irrelevant), because all of the technologies use
pressure vessels of some type, compressors and process heaters. The vendors which manufacture
these items are limited in number, so the more refiners ordering this equipment at the same time,
the longer the leadtime for delivery.
Projecting the timing of this new equipment consisted of a two step process. In the first
step, EPA evaluated the sulfur levels of gasoline certified in 1997 and 1998 to estimate each
refinery's average gasoline sulfur level in these two years. This was done separately for RFG and
CG, as these two types of fuels are certified separately to different quality standards. There are
126 refineries in the U.S. that produced gasoline in 1998. This total includes 12 refineries
located in California, which produce gasoline primarily for the state of California. As only a
small portion of the gasoline produced by California refineries will be affected by this regulation,
these refineries will be addressed separately below. The total of 126 also includes 14 refineries
which are located in the 6 states which are included in the temporary geographical phase-in
program and 17 refineries which fall under the small refiner provisions. These two sets of
refineries are also addressed separately below.
IV-59
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
The following table shows the number of refineries whose average sulfur levels fall into
various ranges.
Table IV-17. Number of U.S. Refineries with 1998 Sulfur Averages Falling Into the
Specified Range of Sulfur Content (ppm)
Sulfur
Range
Number of
Refineries
<100
22
101-200
29
201-300
28
301-400
12
401-500
5
>500
14
As can be seen, refinery's current sulfur levels vary dramatically. This causes some refineries to
deeply desulfurize FCC naphtha earlier under EPA's sulfur control program than others.
The current national average pool sulfur levels were 293 ppm and 207 ppm for CG and
RFG, respectively, based on the CG and RFG batch certification reports. This is considerably
lower than the 1997 CG and RFG sulfur levels of 314 and 282, respectively. This significant
reduction in gasoline sulfur is likely due to the mandatory use of the Complex Model in the
certification of both CG and RFG beginning in 1998. Prior to 1998, the Simple Model was used
to certify RFG and sulfur levels were simply capped at 1990 levels. Prior to 1998, CG sulfur
levels could be as much as 25 percent higher than 1990 levels. The Phase II RFG specifications
begin in 2000. EPA projects that RFG sulfur levels will average roughly 150 ppm during the
summer months. In their comments to the Tier 2/sulfur rule, several refiners and refining
organizations indicated that refiners could meet the sulfur-related portion of the Phase II RFG
requirements without reducing their refinery average sulfur levels. They indicated that they
could shift sulfur from RFG to CG during the summer and reverse this shift during the winter.
As noted above, 1998 RFG sulfur levels are only 57 ppm above this target. No further reduction
in pool sulfur is expected to occur with Phase II RFG, only shifts in sulfur between CG and RFG
The second step in this process was to develop and apply criteria which indicate which
refineries must install and apply desulfurization equipment under various sulfur standards. In
any particular calendar year, refiners have to comply with up to three specifications: a per gallon
standard, a corporate average standard and a refinery average standard. The first applies to every
batch of gasoline introduced into the market. The second can be met by averaging across
refineries within a corporation, and also through the trading of allotments across refiners. Thus,
in essence, assuming allotments are actively traded, the corporate average standard essentially
becomes a national average standard. Finally, the refinery average standard can be met on
average by a refinery and through the trading of credits generated relative to this standard. In
2005 and 2006, the refinery average standard can also be met through the trading of early sulfur
reduction credits (i.e., reductions in sulfur relative to a refinery's baseline generated prior to
IV-60
-------
Chapter IV: Technological Feasibility
2004). Criteria were developed to represent each of these three specifications.
The first set of criteria apply to a refinery's ability to comply with a per gallon cap. For
sulfur levels of 100 ppm or more, EPA projects that a refinery will produce gasoline with an
average sulfur level about two-thirds of the level of the per gallon cap. For example, a refinery is
expected to average 200 ppm sulfur under a 300 ppm cap. There is no need to project an average
sulfur level under the 80 ppm cap, as compliance with the average standard of 30 ppm should be
sufficient for compliance with the cap (temporary equipment disruptions aside).
As indicated above, the corporate average standards are essentially national average
standards, intended to ensure that, on average, the national pool of gasoline does not exceed
certain sulfur levels. Thus, the primary criterion indicating compliance is the national average
sulfur level. Again assuming actively traded allotments, the national average sulfur level need be
just below the corporate average standard. However, it is unlikely that every refiner would
market every allotment generated relative to the standard. Refiners will likely keep some
allotments back from the market in order to provide themselves with a compliance cushion
towards the standard. Also, there could be some inefficiencies in the market. For example, some
refiners may be looking for allotments prior to other refiners deciding that they have excess
allotments to sell. Later, when allotments become available, potential buyers have already
decided to comply with the standards without the need for allotments. Thus, to be conservative,
we projected that refiners would have the capability of producing fuel with a national average
sulfur level 30 ppm below the applicable corporate average standard. This translates into the
capability to produce national average sulfur levels of 90 ppm in 2004 and 60 ppm in 2005.4
Finally, criteria were developed to represent compliance with the refinery average
standard of 30 ppm. In 2005, we assumed that refiners would use early credits to allow them to
produce fuel complying with the corporate average standard of 90 ppm. In 2006 and later, we
assumed that the refinery would have to average 30 ppm sulfur.
The third and final step was to determine when a refinery had to build a desulfurization
unit in order to comply with one of the above criteria. First, when complying with the temporary
cap of 300 ppm, as indicated above, a refinery is presumed to average 200 ppm sulfur or less.
4 While refiners are projected to be able to produce gasoline at these levels, this does not necessarily mean
that these sulfur levels will be achieved. We expect that most refiners will design their refineries to be capable of
producing 30 ppm sulfur gasoline, as this is the long-term standard. However, prior to 2006, some refiners may
utilize their equipment to produce gasoline at 30 ppm, while others will find that they can comply with the interim
standards while using this equipment to produce gasoline with higher sulfur levels. For example, a refiner may
produce 30 ppm gasoline for a portion of the year in order to generate allotments in case of operational difficulties
later in the year. However, as the year progresses, these allotments will build up and the likelihood of a disruption
will decrease. Therefore, they could operate this equipment at a lower sulfur removal efficiency in order to reduce
costs.
IV-61
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
We project that a refinery could reduce its maximum sulfur level (i.e., that which determines
compliance with the per gallon cap) by 20 percent for a year or so in order to delay construction
of new equipment. Therefore, refineries with average sulfur levels of 250 or less are projected to
be able to delay construction of new desulfurization units under the 300 ppm cap. In addition,
three refiners informed EPA that they had refineries with sulfur averages above 250 ppm which
could meet the 300 ppm cap without major construction due to more unique circumstances
(excess equipment on site, existing excess hydrodesulfurization capacity, etc.)
This 20 percent sulfur reduction due to operational modifications is based on an
assessment of the capability of a number of sulfur reducing techniques available to refiners.
First, refineries with FCC feed hydrotreating can usually increase the severity of their units to
remove more sulfur from the FCC naphtha. Second, many refiners can switch or shift to a lower
sulfur crude oil. Third, refiners can route some of their FCC naphtha to their reformer
hydrotreater and reformer. Fourth, refiners can change their FCC catalyst to one which reduces
the amount of sulfur in FCC naphtha. (Grace makes such a catalyst.) Finally, refiners can shift
the heaviest (and most sulfur laden) portion of their FCC naphtha to the distillate pool. The cost
associated with each of these techniques generally increases as one progresses down the list. The
cost which a refiner would be willing to pay will depend on the value of delaying the selection
and construction of new desulfurization equipment.
With regard to the final 30 ppm refinery average standard, we project that a refinery will
need to construct a new desulfurization unit if its current average is greater than 50 ppm. As
discussed in Section 4 above, refiners that currently have such a low sulfur level probably do not
have an FCC unit, hydrotreat the FCC feed, or utilize low sulfur crude. In this situation, the
refiner should be able to reduce their sulfur further through operational changes and avoid major
capital investment.
The same basic criteria were applied to those refineries which are covered by the
temporary geographical phase-in program and the small refiner provisions, with the exception
that the standards which these refineries must meet differ to some degree from those which are
generally applicable. For example, under the temporary geographical phase-in program, refiners
must meet a 300 ppm cap and a 150 ppm refinery average standard in 2004. Since credits or
allotments can be purchased and used in meeting the 150 ppm standard, the 300 ppm cap is the
primary controlling standard. As discussed above, refineries currently with an average sulfur
level of more than 250 ppm are projected to install a desulfurization unit in order to meet the 300
ppm cap, except for three refineries, as noted previously. These standards apply unchanged until
2007, when the 30 ppm average standard and 80 ppm cap take effect. As projected above for
other refineries, all refineries with a current average sulfur level above 50 ppm are projected to
install a desulfurization unit in order to meet the 2007 standards.
Refineries covered under the small refiner provisions which currently have an average
sulfur level of 200 ppm or less only need to maintain that level until 2008. Those with current
IV-62
-------
Chapter IV: Technological Feasibility
average sulfur levels between 200 and 300 ppm will have to modestly reduce their sulfur levels,
while those with higher sulfur levels will have to reduce their sulfur levels more significantly.
Only five small refiners currently average more than 300 ppm sulfur. Given that these refiners
will not have to meet the 30 ppm standard until 2008 (four years after the date of the initial
standard), these five refiners have a significant incentive to delay the installation of naphtha
desulfurization equipment until that time. Thus, we project that they will modify their refinery
operations to the fullest degree possible to reduce sulfur and avoid constructing their final
desulfurization equipment. This could involve the operational modifications described above. It
could also involve more modest investment in new equipment, which would reduce sulfur
modestly and immediately, while being useful in the long term, as well. An example of this
would be the installation of a distillation column to separate light and heavy FCC gasoline and a
Merox unit to remove sulfur from the light FCC naptha. Instead of projecting that all five will
require a naphtha desulfurization unit in 2004, we only project that two will require units.
Regarding California refiners, we analyzed the gasoline sold by these refiners outside of
California in 1997 and 1998. We also compared the volume of fuel sold outside of California to
that sold inside of California by each refiner. With two exceptions, the non-California fuel
produced by California refiners contains 50 ppm sulfur or less and represents a small fraction of
total gasoline production. Thus, compliance with the Federal sulfur program should not require
major new equipment at 10 of the 12 California refineries marketing gasoline outside of the state.
One of the two remaining refineries certified non-California fuel in 1998, but did not do so in
1997. The 1998 non-California fuel averaged 91 ppm and represented 10-15 percent of their
total gasoline production. Assuming that their California gasoline was at 30 ppm sulfur or lower,
this refinery's overall average sulfur level was 36-39 ppm, well under our 50 ppm criterion
mentioned above. Thus, this refinery should not require an additional desulfurization unit. The
one remaining refinery appears to be selling a significant fraction of its fuel outside of California
at an average sulfur level of about 150 ppm in 1998. We project that this refinery will add a
desulfurization unit in 2006 in order to comply with the Federal sulfur control program.
We applied these criteria to the each refinery's 1997 and 1998 two-year average sulfur
levels and projected both the number of new desulfurization units which would be required each
year, as well as the national average sulfur level. In determining the national average sulfur
level, we assumed that refineries with new desulfurization units would operate at 30 ppm sulfur
on average year round. The results are shown in Table IV-18 below.
IV-63
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table IV-18. Number of New Desulfurization Units Operating by January 1 of Indicated
Year and National Pool Average Sulfur Levels Under the Final Sulfur Standards
New Units
Sulfur (ppm)
2003 and
Earlier
7-10
—
2004
37-40
67
2005
6
60
2006
23-26
31
2007
9
30
2008
9
30
As shown in Table IV-16, 7-10 desulfurization units are projected to be built and
operating in 2003 or earlier. This is based on information received from both licensors of
desulfurization technology and refiners. Included in these units are a CDTech unit at Motiva's
Port Arthur refinery, a CDTech unit at another refinery not yet publicly identified, two Black and
Veatch units at smaller refineries, one Phillips S-Zorb unit, and 2-5 additional units at major
refineries desiring to reduce sulfur early. The two CDTech projects are expected to be
operational in 2000, while the Black and Veatch and Phillips projects are expected to be
operational in 2001 and 2002. The additional units are projected to be operational in 2002 and
2003.
These pre-2004 desulfurization units will also provide a source for early sulfur reduction
credits. Based on the production volumes and sulfur levels of the refineries projected to receive
these projects, we project that these new units will reduce national average sulfur levels by 1 ppm
in 2001, 29 ppm in 2002, and 29-51 ppm in 2003. These projections are based on:
Two refineries (representing about 1.0% of non-California gasoline production) installing
CDTech and Octgain 220 desulfurization units in 2000 which reduces half of its gasoline
to 30 ppm,
Three refineries (representing about 1% of non-California gasoline production) installing
absorbent desulfurization units in 2001 and 2002 to produce 30 ppm gasoline in order to
demonstrate this technology,
Two to five refineries (representing about 3-8% of non-California gasoline production)
installing absorbent desulfurization units in later 2002 and 2003 to produce 30 ppm
gasoline in order to generate early sulfur credits and allotments,
The baseline sulfur levels for these refineries range from 130-700 ppm; on average, the
baseline sulfur level is roughly 350-400 ppm.
On an annual national gasoline pool basis, these annual credits sum to a total of 59-81
ppm of credit. Ignoring the small increase in gasoline consumption annually, operating at 90
ppm in 2005 will require that refiners use 60 ppm of credits relative to the 30 ppm refinery
average standard. The credits which can be used by small refiners beyond 2005 are very low, due
IV-64
-------
Chapter IV: Technological Feasibility
to their small production volumes and amount to less than 5 ppm. Thus, these early units will
either provide all of the credits which are needed in 2005 and beyond, with roughly 15 ppm of
credit to spare, or will fall less than 5 ppm short.
In addition to credits from new plants, refiners can generate credits operationally, as
discussed above. Eliminating the 5 ppm shortfall would only require that annual average sulfur
levels from 2000-2003 be reduced by 1 ppm on average. Routing heavy FCC gasoline to
reforming hydrotreaters has far more potential to reduce sulfur than 1 ppm on average. Also, as
mentioned above, sulfur levels in 1998 were significantly below those of 1997. Given that
refinery's baselines will be based on their 1997-98 average, many of these refineries can generate
credits if they can continue producing gasoline at their 1998 levels. This approach could
generate more than 15 ppm of credit per year, well above that needed to complement credits
generated from new desulfurization units, even if refiners did not trade all of the credits that they
generated.
IV-65
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Chapter IV References
1. SAE paper 1999-01-0774, "Using Advanced Emission Control Systems to Demonstrate
LEV H ULEV 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.
4. Final Report, 1996 American Petroleum Institute/National Petroleum Refiners
Association Survey of Refining Operations and Product Quality, July 1997.
6. Final Report, 1996 APJ/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. "Summary of Revisions to the October 22, 1999 Proposed Phase 3 California Gasoline
Regulations And Other Actions that ARB Staff Intends to Propose at the December 9,
1999 Hearing Proposed California Phase 3 Reformulated Gasoline Regulations, Staff
Report: Initial Statement of Reasons, California Environmental Protection Agency,
October 22, 1999," California Environmental Protection Agency, Air Resources Board,
December 7, 1999..
11. Idemitsu Kosan Co., Ltd., "Clean Air Program in Japan," Presentation to the U.S. EPA,
December 1997.
12. Davey, Steven W., Haley, John T., FCC Additive Technology Update, 1996 O&G J
International Catalyst Conference and Exhibition, Houston Texas, February 1996.
13. Petroleum Refinery Process Economics, Maples, Robert E., PennWell Books, Tulsa,
Oklahoma, 1993.
14. Mobil Octgain Process, a Proven FCC Gasoline Desulfurization Process, Recent Process
Improvements, Presentation by Trig Tryjankowski to EPA staff, August 1998.
IV-66
-------
Chapter IV: Technological Feasibility
15. Greeley, J.P., Zaczepinski, S., Selective Cat Naphtha Hydrofming with Minimal Octane
Loss, NPRA 1999 Annual Meeting.
16. Nocca, J.L., et al, Cost-Effective Attainment of New European Gasoline Sulfur
Specifications within Existing Refineries, November 1998.
17. Prime G, A Sweet Little Process for Ultra-Low Sulfur FCC Gasoline without Heavy
Octane Penalty, IFF Industrial Division.
18. Podar, Syamal K., Hilbert, Timothy L., Octgain, Evaluation for the Manufacture of
Reformulated Gasoline via LP Modeling, NPRA 1995 Annual Meeting.
19. Mobil Octgain Process, a Proven FCC Gasoline Desulfurization Process, Recent Process
Improvements, Presentation by Trig Tryjankowski to EPA staff, August 1998.
20. Shih, S. S., Mobil's Octgain Process: FCC Gasoline Desulfurization Reaches a New
Performance Level, NPRA 1999 Annual Meeting.
21. CDTECH, FCC Gasoline Sulfur Reduction, CDTECH, Sulfur 2000, Hart's Fuel and
Technology Management, Summer 1998.
22. Rock, Kerry J., Putman, Hugh, Global Gasoline Reformulation Requires New
Technologies, Presented at Hart's World Fuels Conference, San Francisco, March 1998.
23. Rock, Kerry L., et al, Improvements in FCC Gasoline Desulfurization via Catalytic
Distillation, Presented at the 1998 NPRA Annual Meeting, March 1998.
24. Irvine, Robert L. et al, IRVAD Process-Low Cost Breakthrough for Low Sulfur Gasoline,
Paper presented at 1999 NPRA Annual Meeting.
25. Printed Literature by Phillips Petroleum Shared with EPA September 1999.
26. Shanley, A., Benchmarking for the Next Century, Chemical Engineering, April 1996.
27. California Code of Regulations, Title 13 §2252.
28. California Code of Regulations, Title 13 §2260 - §2272.
29. 55 FR 34138, August 21, 1990.
30. Refining Industry Profile Study; EPA contract 68-C5-0010, Work Assignment #2-15, ICF
Resources, September 30, 1998.
IV-67
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
IV-68
-------
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. Cost estimates have been prepared for all
categories of vehicles, LDVs through LDT4. The cost estimates for medium-duty passenger
vehicles (MDPVs) to meet Tier 2 exhaust and evaporative standards have been grouped with the
costs for LDT4S.1 We have taken this approach with MDPVs because they are grouped with
FtLDTs in the program for phase-in purposes and are required to meet essentially the same
requirements as vehicles in the LDT4 category. 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, LOT Is, and LDT2s, and Tier 1 technologies
for LDT3s and LDT4s. These are the standards that vehicles will be meeting in 2003.2 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
1 EPA has categorized passenger vehicles (primarily SUVs and passenger vans) between 8,500 pounds and
10,000 pounds GVWR as MDPVs and has included them in the Tier 2 program.
2 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. The MDPVs are required to meet
engine-based standards prior to 2004. The projected technologies likely to be used by manufacturers to meet the
2003 engine-based standards form the baseline for these vehicles.
V-l
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
discussions with manufacturers about trends in technology.
As described in detail below, we have projected costs for the final Tier 2 standards. We
have not projected specific incremental costs for the interim standards contained in the Tier 2
program.3 To account for the interim standards in the cost analysis, we have assumed that the
manufacturers would opt to accelerate the phase-in of Tier 2 vehicles rather than redesign
vehicles for the interim program. The Tier 2 program averaging flexibility allows manufacturers
to take this approach. We believe this approach by the manufacturers is likely because it allows
manufacturers to avoid significant R&D efforts to meet standards that are in effect for only a few
model years.
The following analysis projects a relatively uniform emission control strategy for various
LDV, LDT, and MDPV models. However, this should not suggest that a single combination of
technologies would be used by all manufacturers. Selecting technology packages requires
extensive engineering development work and EPA does not know future technology mixes and
costs with certainty for each vehicle model. 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 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
certification levels are well below current standards, also suggesting this approach makes sense.
We have made a best estimate of the combination of technologies that any manufacturer might
use to meet the standards at an acceptable cost and these technologies form the basis of the cost
estimates. In making our cost estimates, we have relied on our own technology assessment
including the results of our in-house testing, described in Chapter IV. Since California, in their
LEV n program, has adopted essentially the same standards and time-line as Tier 2, 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
3 We have assumed for purposes of our cost analysis that manufacturers will choose the Tier 2 program
option that brings all 2004 model year vehicles into the Tier 2 program. We believe manufacturers are very likely
to select this option due to the program flexibility it provides.
V-2
-------
Chapter V: Economic Impact
analyses. Most manufacturer input is considered confidential business information and therefore
is not described in detail.
As noted above, 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 LDTls, 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 LDTls 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 about 0.156 g/mile or less. Vehicles have
their emissions capped at 0.60 g/mile NOx and 0.23 g/mile NMHC prior to phase-in.4 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.
4 Manufacturers may select an option that provides an NMOG standard of 0.280 g/mile for LDT4s and
MDPVs for the 0.6 g/mile NOx bin. Manufacturers also may select an option that allows MDPVs to be placed in a
bin with a NOx level of 0.9 g/mile and a NMOG level of 0.280 g/mile during the interim program. Further, the
optional program provides that diesel vehicles in the MDPV category may be certified to heavy-duty engine-based
standards prior to 2008. The optional standards are equivalent to those that apply in the California LEV I program
in 2004-2006.
V-3
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
For the interim fleet average NOx standard, (average standard of 0.2 g/mile NOx with an
NMHC standard of about 0.156 g/mile or less), 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. In other words, manufacturers
would introduce Tier 2 vehicles early and use the averaging program to avoid redesigning
vehicles to the interim standards. We believe this approach is reasonable considering
manufacturers are likely to avoid significant R&D efforts to meet an interim standard that is in
effect for only a few model years. Essentially, a few such interim 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. We have projected an accelerated phase-in of LDT3s and LDT4s, as noted above. For
both phase-in periods (for LDVs, LOT Is, LDT2s, and for LDT3s, LDT4s, and MDPVs), 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/MDPVs 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 and 10-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
V-4
-------
Chapter V: Economic Impact
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
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, LDTs, and MDPVs will be equipped with advanced catalysts. Catalyst manufacturers are
currently working with engine manufacturers on improved 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
V-5
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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
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 $12.20 and $67.10 per vehicle. We derived the increased volume cost by
taking the baseline cost of the catalyst per liter ($6 I/liter) and multiplying by the increase in
catalyst volume.5 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 per vehicle cost from $2.36 for light-duty vehicles to $29.50 for
LDTs and MDPVs with the largest displacement engines. The details of the underfloor catalyst
cost estimates are provided in Table V -1.
5 We have updated the baseline per liter catalyst cost and other catalyst costs from the NPRM to reflect
changes in the spot prices of precious metals. The precious metals costs used in the cost analysis are shown in
Table V-l.
V-6
-------
Chapter V: Economic Impact
Table V-l. Main or Underfloor Catalyst Cost Breakdown
Vehicle Engine
Type Type
LDV 4-cylinder
6-cylinder
8-cylinder
LOT/ 4-cylinder
MDPV 6-cylinder
8-cylinder
8/10-cylinder
Sales wtd. Projected Projected Increased
Engine Baseline Cat. Tier 2 Cat. Volume
Displacement Volume Volume Cost (a)
(liter] (liter) (liter) (dollars)
2.0 1.8 2.0 12.20
3.2 2.8 3.2 24.40
4.5 4.0 4.5 30.50
2.3 2.3 2.3 0.00
3.7 2.6 3.7 67.10
5.4 4.7 5.4 42.70
6.0 4.7 5.4 42.70
Increased Increased Increased Added Added Added
Platinum Palladium Rhodium (b) Pt cost Pd cost Rh cost (b)
(Pt) (Pd) (Rh) (dollars) (dollars) (dollars)
(grams) (grams) (grams)
0.000 0.000 0.085 0 0 2.35
0.000 0.000 0.138 0 0 3.86
0.000 0.000 0.194 0 0 5.43
0.000 0.000 0.097 0 0 2.71
0.035 0.540 0.157 0.44 7.17 4.39
0.082 0.550 0.229 1.03 7.30 6.41
0.164 1.100 0.458 2.06 14.62 12.82
Higher
substrate
cost (e)
(dollars)
2.44
3.90
5.49
2.81
4.52
6.59
6.59
Total
Increased
Cost
(dollars)
16.99
32.16
41.42
5.52
83.62
64.03
78.79
Precious Metal Costs
$/troy ounce I/gram
Pt 412 12.58
Pd 390 13.29
Rh 868 28.00
(a) Baseline catalyst cost is |61/litar. Increased catalyst volume costs are the increase in catalyst volume multiplied by |61/liter.
(b) Increase in Rh of 1.2 g/cu t
V-7
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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, LDT4s, and MDPVs the use of close
coupled catalysts is currently low relative to the other classes, especially for MDPVs. For Tier 2
LDT3s, LDT4s and MDPVs, 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 $107.54 and $131.44, 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 available
for use in 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. We believe 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-8
-------
Chapter V: Economic Impact
/'/'/'. 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-powered air injection strategies for Tier 2 vehicles equipped with 6- and 8-
cylinder engines. The air injection systems consist of an electric-powered 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. In some cases, manufacturers may be able to use the combined
exhaust system improvements in lieu of adding close-coupled catalysts. However, we are not
projecting an increase in the use of low thermal capacity manifolds due to the Tier 2 standards.
For most vehicles, manufacturers using close-coupled catalysts are not likely to need the
improved manifolds as well.
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
V-9
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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
set of major changes to these engines. Therefore, EPA has not included an engine modification
cost for these vehicles. For LDT3s, LDT4s and MDPVs, which would be changing from Tier 1
to Tier 2 technology, we have included a hardware cost for engine modifications of $10 and $15
for six and eight/ten cylinder engines, respectively.
vi. Exhaust Gas Recirculation (EGR)
One of the most effective means of reducing engine-out NOx emissions is exhaust gas
recirculation. By recirculating spent exhaust gases into the combustion chamber, the overall air-
fuel mixture is diluted, lowering peak combustion temperatures and reducing NOx. Many EGR
systems in today's vehicles utilize a control valve that requires vacuum from the intake manifold
to regulate EGR flow. Some vehicles are being equipped with electronic EGR in place of
mechanical back-pressure designs. By using electronic solenoids to open and close the EGR
valve, the flow of EGR can be more precisely controlled. EPA projects that the use of full
electronic EGR systems will increase due to Tier 2 standards. We estimate that about 50 percent
of Tier 2 vehicles will be equipped with electronic EGR at an incremental cost often dollars per
vehicle.
vii. Total Hardware Costs for Exhaust Emissions Control
Table V-2 provides a summary of the total hardware costs for each vehicle and engine
type. Tables V-3 through V-7 present detailed estimated manufacturer costs itemized for each
vehicle and engine type. The tables indicate EPA's estimate of the percentage of use of the
technologies for both the baseline and the Tier 2 vehicles. Some of the technologies listed, such
as individual cylinder fuel control and retarded spark timing, involve calibration changes only
and have no hardware costs associated with them.
V-10
-------
Chapter V: Economic Impact
Table V-2. Total Estimated Per Vehicle Manufacturer
Incremental Hardware Costs for the Tier 2 Standards
4-cylinder
6-cylinder
8-cylinder
larger 8/10-cylinder*
sales weighted
LDV
($)
24.99
65.16
75.42
N/A
44.69
LDT1
($)
13.16
91.46
N/A
N/A
39.87
LDT2
($)
8.16
90.98
70.97
N/A
84.27
LDT3
($)
N/A
238.86
171.99
N/A
178.74
LDT4/MDPV
($)
N/A
N/A
171.99
291.54
187.53
* Primarily used in MDPVs.
V-ll
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 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
12.20
2.35
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.
0
50
10
100
100
100
100
0
100
0
50
60
70
0
0
100
100
100
0
nc. cost
over Tier 1
(in dollars)
0.00
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
12.20
2.35
2.44
0.00
24.99
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
24.40
3.86
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
:hat will
req. tech.
0
50
10
100
100
100
100
0
100
100
50
0
100
100
0
100
100
100
50
nc. cost
over Tier 1
(in dollars)
0.00
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
24.40
3.86
3.90
25.00
65.16
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
30.50
5.43
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.
0
50
10
100
100
100
100
0
100
100
50
0
60
80
40
100
100
100
50
nc. cost
over Tier 1
(in dollars)
0.00
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
30.50
5.43
5.49
26.00
75.42
(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 8-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-powered 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-4. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT1 Compared to NLEV LDT1
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
Improved 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
2.35
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.
0
50
10
100
100
100
100
0
100
0
50
60
70
0
0
100
100
100
50
Inc. cost
over Tier 1
(in dollars)
0.00
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
0.00
2.35
2.81
0.00
13.16
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
67.10
3.86
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.
0
50
10
100
100
100
100
0
100
100
50
0
100
100
0
100
100
0
75
Inc. cost
over Tier 1
(in dollars)
0.00
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
67.10
3.86
0.00
12.50
91.46
(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, 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-powered air pump with integrated filter and relay, wiring, air-shut-off valve with integral solenoid, check valve, tubing and brackets.
V-13
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 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
2.35
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.
0
50
10
100
100
100
100
0
25
100
0
50
60
70
0
0
0
0
0
100
100
0
Inc. cost
over Tier 1
(in dollars)
0.00
0.00
0.00
0.00
0.00
3.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
2.35
2.81
0.00
8.16
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
67.10
4.32
51.67
3.86
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.
0
50
10
100
100
100
100
0
25
100
100
50
0
100
100
0
100
0
0
100
100
75
Inc. cost
over Tier 1
(in dollars)
0.00
0.00
0.00
0.00
0.00
3.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
67.10
0.00
0.00
3.86
4.52
12.50
90.98
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
42.70
10.13
52.83
5.43
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.
0
50
10
100
100
100
100
0
25
100
100
50
0
60
80
40
100
0
0
100
100
75
Inc. cost
over Tier 1
(in dollars)
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
42.70
0.00
0.00
5.43
6.59
16.25
70.97
(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 8-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-powered 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
Table V-6. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT3 Compared to Current LDTSs
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)
1 ncreased 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
107.54
160.00
67.10
0.44
7.17
4.39
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.
0
50
10
100
100
100
100
0
100
25
100
50
0
100
100
0
100
100
100
100
100
50
Inc. cost
over Tier 1
(in dollars)
0.00
6.00
0.00
0.00
0.00
3.00
1.60
0.00
10.00
0.00
10.00
5.00
0.00
0.00
94.64
0.00
67.10
0.44
7.17
4.39
4.52
25.00
238.86
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
131.44
160.00
42.70
1.03
7.30
6.41
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.
0
50
10
100
100
100
100
0
100
25
100
50
0
60
80
40
100
100
100
100
100
50
Inc. cost
over Tier 1
(in dollars)
0.00
8.00
0.00
0.00
0.00
3.00
1.60
0.00
10.00
0.00
15.00
5.00
0.00
0.00
32.86
0.00
42.70
1.03
7.30
6.41
6.59
32.50
171.99
(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 8-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-powered air pump with integrated filter and relay, wiring, air-shut-off valve with integral solenoid, check valve, tubing and brackets.
V-15
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-7. Estimated Incremental Manufacturer Hardware Cost for Tier 2 LDT4s and MDPVs Compared to Current Vehicles
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 (R)
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 (87%)
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
131.44
160.00
42.70
1.03
7.30
6.41
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.
0
50
10
100
100
100
100
0
100
25
100
50
0
60
80
40
100
100
100
100
100
50
Inc. cost
over Tier 1
(in dollars)
0.00
8.00
0.00
0.00
0.00
3.00
1.60
0.00
10.00
0.00
15.00
5.00
0.00
0.00
32.86
0.00
42.70
1.03
7.30
6.41
6.59
32.50
171.99
Larger 8 & 10-Cylinder (13%)
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
131.44
160.00
42.70
2.06
14.62
12.82
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
0
40
0
0
0
0
0
0
% Tier 2
that will
req. tech.
0
50
10
100
100
100
100
0
100
25
100
50
0
60
80
40
100
100
100
100
100
100
Inc. cost
over Tier 1
(in dollars)
0.00
8.00
0.00
0.00
0.00
3.00
1.60
0.00
10.00
0.00
15.00
5.00
0.00
0.00
105.15
0.00
42.70
2.06
14.62
12.82
6.59
65.00
291 .54
(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 8-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-powered air pump with integrated filter and relay, wiring, air-shut-off valve with integral solenoid, check valve, tubing and brackets.
V-16
-------
Chapter V: Economic Impact
c. Hardware Costs for Evaporative Emissions Control
The standards 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 new 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 finalizing. 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 percent
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
V-17
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
connections and welds are made. Materials and manufacturing techniques exist to reduce these
losses.
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 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 90
Improved
Vehicle
(b)
0.01
0.01
0.01
0.02
0.04
0.02
0.06
0.10
0 90
Chang
e
(a-b)
0.09
0.09
0.22
0.04
0.08
0.04
0.12
0.10
o
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
V-18
-------
Chapter V: Economic Impact
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 $2,400 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
for larger vehicles, but our evaporative standard for HLDTs (1.2 gpt) and MDPVs (1.4 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
proportionally 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 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
V-19
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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
include costs for research 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, LDT4s and
MDPVs. 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.6
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
6 This estimate is based on staff cost of $60 per hour and development vehicle cost of $100,000 per vehicle.
V-20
-------
Chapter V: Economic Impact
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. This estimate does not account for the ability of
manufacturers in some cases to carry-over certification data from California, which would lower
certification costs.
We expect there to be a certification testing cost savings for HLDTs due to the change in
test procedures for these vehicles. For Tier 2, HLDTs will be emissions tested at the same test
weight as is required for the CAFE fuel economy test (i.e., loaded vehicle weight). Currently,
HLDTs are emissions tested at a higher weight (adjusted loaded vehicle weight). This change in
emissions test procedure will allow manufacturers to measure fuel economy and emissions
during the same test, eliminating one of the FTP tests currently required. To be conservative,
however, we have not reduced the certification cost estimate to reflect this likely cost savings.
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
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 $22, as shown
in Table V-9 (aggregate costs are described in the following section). We derived the per
vehicle fixed cost by dividing the total fixed cost per vehicle class over the five year recovery
period by the estimated total sales per vehicle class over the same period. Differences in fixed
costs among vehicle classes occur because we have projected a phase-in of Tier 2 LDVs and
LDTs/MDPVs and changes in sales volumes over time for the vehicle classes. The aggregate
fixed costs, vehicle phase-ins, and sales projections are described in section 3., below.
V-21
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-9. Per Vehicle Fixed Costs
R&D
Tooling
Certification
Total
LDV
($)
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
LDT3
($)
14.34
5.01
0.26
19.61
LDT4/MDPV
($)
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
($)
47.94
22.03
69.97
LDT1
($)
43.12
19.47
62.59
LDT2
($)
87.52
19.26
106.78
LDT3
($)
181.99
19.61
201.60
LDT4/MDPV
($)
190.78
21.18
211.96
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
the first five years of production, after which they would expire. For variable costs, research in
the costs of manufacturing has shown that as manufacturers gain experience in production, they
are able to lower the per-unit cost of production. These effects are often described as the
manufacturing learning curve.6
The learning curve is a well documented 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 of 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
V-22
-------
Chapter V: Economic Impact
by improvements in several areas. Areas involving direct labor and material are usually the
source of the greatest savings. Examples 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.
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.
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 services7. The distribution of these progress ratios is
shown in Figure V-l. 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. The average progress ratio for the whole data set falls between 81 and 82
percent. Other studies (Alchian 1963, Argote and Epple 1990, Benkard 1999) appear to support
the commonly used p value of 80 percent, i.e., each doubling of cumulative production reduces
the former cost level by 20 percent.
The learning curve is not the same in all industries. For example, the effect of the
learning curve seems to be less in the chemical industry and the nuclear power industry where a
V-23
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
15
10
0)
3
D"
CD
0
Distribution of Progress Ratios
i i i
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
(Button and Thomas, 1984)
V-24
-------
Chapter V: Economic Impact
doubling of cumulative output is associated with 11% decrease in cost (Lieberman 1984,
Zimmerman 1982). The effect of learning is more difficult to decipher in the computer chip
industry (Gruber 1992).
We applied a p value of 80 percent in this analysis. Using one year as the base unit of
production, the first doubling would occur at the start of the third model year of production.
Beyond that time, 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, LDT4s, and MDPVs 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
($)
69.97
64.23
42.20
LDT1
($)
62.59
58.38
38.91
LDT2
($)
106.78
99.12
79.86
LDT3
($)
201.60
180.69
161.08
LDT4/MDPV
($)
211.96
189.96
168.78
V-25
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 vehicle standards. Manufacturers will most likely meet the standards through
improvements to existing technologies. The costs of fuel quality improvements are provided in
section B, below.
We do not anticipate that the improvements to technologies will affect fuel economy or
in-use maintenance. We expect the standards to be met through improvements in current
technologies rather than through the use of new technologies. We do not believe these
improvements would adversely affect fuel economy or maintenance costs. Also, we have not
observed fuel economy losses in our testing programs described in Chapter IV.
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 (incremental to ORVR),
as well as the improved exhaust emissions control system.7 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
($)
82.43
75.22
53.19
LDT1
($)
73.80
68.50
49.03
LDT2
($)
129.54
119.90
100.64
LDTS
($)
248.92
222.60
202.99
LDT4/MDPV
($)
261.57
233.52
212.34
7 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-26
-------
Chapter V: Economic Impact
The above analysis presents estimated vehicle costs for Tier 2 exhaust and evaporative
emissions standards. In addition, we are finalizing On-board Diagnostics (OBD n) and On-board
Refueling Vapor Recovery (ORVR) for MDPVs. Light-duty vehicles and light-duty trucks
already must comply with these requirements. The OBD II and ORVR requirements were
proposed as part of a Heavy-duty Engines and Vehicles Regulation (64 FR 58472) and the
detailed cost analyses are presented in the RIA for that rulemaking (Docket A-98-32, Item II-B-
01)
In summary, for OBD II, the vehicles will likely be equipped with additional and
improved hardware such as additional oxygen sensors, solenoids for the evaporative system
purge and leak check, and improved electronic control modules. We estimate the total cost to
consumers for the system to be about $80 per vehicle. For the ORVR system, we estimate the
cost to consumers to be about $10 per vehicle. Also, the ORVR system provides a fuel economy
savings of about $6 over the lifetime of the vehicle. This savings occurs because refueling
vapors are captured, and burned in the engine, rather than escaping to the atmosphere.
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, LDTs, and MDPVs. 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
$269 million in 2004 and peaking at $1,579 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 $1,351 million
in 2014, after which the growth in vehicle population leads to increasing costs that reach $1,392
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.
V-27
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-13. Estimated Annual Nationwide Costs
(thousands of dollars)
Category
LDV
LDT1
LDT2
LDT3
LDT4/MDPV*
Total
2004
253,327
0
0
9,544
5,907
268,778
2009
358,521
98,943
579,898
339,109
201,991
1,578,462
2014
301,938
73,026
499,791
306,125
169,841
1,350,721
2020
311,110
75,245
514,973
315,425
174,000
1,391,753
"Includes costs for OBD II and ORVR requirements for MDPVs
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 LOT fraction of total sales by 1.6 percent per year
from 1998 through 2008.n 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.12 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 period of the
analysis.13 Table V-14 provides EPA's estimates for vehicle sales for 1998 and projections for
select future years.
V-28
-------
Chapter V: Economic Impact
Table V-14. Estimated Annual 49-State Vehicle Sales
(thousands of vehicles)
Category
LDV
LDT1
LDT2
LDT3
LDT4/MDPV8
Total
1998
7,352
1,012
3,374
1,025
541
13,304
2004
6,266
1,268
4,228
1,284
663
13,709
2008
5,502
1,447
4,824
1,465
747
13,985
2072
5,620
1,475
4,917
1,493
762
14,267
2020
5,849
1,535
5,117
1,554
793
14,848
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 29 percent of vehicles overall would fall into the highest bin (0.60 g/mile NOx), 28 percent
in the next highest bin (0.3 g/mile NOx) and the remaining 43 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.
8 To account for sales of MDPVs, we estimated 1998 sales of 70,000 vehicles and projected growth at a
rate of one-half percent per year. The MDPV sales projections were added to the yearly sales estimates for LDT4s.
V-29
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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
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*
(%)
3
9
26
68
100
100
LDT4/MDPV*
(%)
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, LDT4s,and MDPVs starting with LDT3s in 2008. OBD H is required
for MDPVs starting in 2004. The phase-in for ORVR for MDPVs is 40/80/100 percent in 2004-
2006.
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 average 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
V-30
-------
Chapter V: Economic Impact
reduced after the second year of production due to the learning curve factor. Tables V-16
through V-20 present total annualized nationwide costs by vehicle class for years 2004 through
2020. Table V-21(A) presents these cost figures summed for all vehicle categories. In addition,
Table V-20 and V-21(A) include aggregate costs for MDPV OBDII and ORVR requirements.9
Table 21(B) provides the non-annualized costs for the Tier 2 program.
9 The net present value of the fuel savings over the life of the vehicle due to ORVR, estimated to be $5.50,
has been subtracted from the system cost of $10.25 for purposes of estimating the aggregate costs.
V-31
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-16. Annual Nationwide Costs For Tier 2 LDVs
Calendar
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Fixed Cost
($)
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
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Variable Cost
($)
189,306,353
367,257,308
334,503,997
303,014,320
292,689,560
294,501,019
295,973,524
297,453,392
298,940,659
300,435,362
301,937,539
303,447,227
304,964,463
306,489,285
308,021,731
309,561,840
311,109,649
312,665,198
314,228,524
315,799,666
317,378,665
318,965,558
320,560,386
322,163,188
323,774,003
325,392,874
327,019,838
328,654,937
330,298,212
331,949,703
333,609,451
Total Cost
($)
253,326,525
495,297,653
462,544,342
431,054,665
420,729,905
358,521,192
295,973,524
297,453,392
298,940,659
300,435,362
301,937,539
303,447,227
304,964,463
306,489,285
308,021,731
309,561,840
311,109,649
312,665,198
314,228,524
315,799,666
317,378,665
318,965,558
320,560,386
322,163,188
323,774,003
325,392,874
327,019,838
328,654,937
330,298,212
331,949,703
333,609,451
V-32
-------
Chapter V: Economic Impact
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
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Fixed Cost
($)
0
0
27,715,184
27,715,184
27,715,184
27,715,184
27,715,184
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Variable Cost
($)
0
0
73,706,372
76,145,235
70,928,248
71,227,705
71,583,844
71,941,763
72,301,472
72,662,979
73,026,294
73,391,426
73,758,383
74,127,175
74,497,811
74,870,300
75,244,651
75,620,874
75,998,979
76,378,974
76,760,869
77,144,673
77,530,396
77,918,048
78,307,638
78,699,177
79,092,673
79,488,136
79,885,577
80,285,004
80,686,429
Total Cost
($)
0
0
101,421,556
103,860,420
98,643,433
98,942,890
99,299,028
71,941,763
72,301,472
72,662,979
73,026,294
73,391,426
73,758,383
74,127,175
74,497,811
74,870,300
75,244,651
75,620,874
75,998,979
76,378,974
76,760,869
77,144,673
77,530,396
77,918,048
78,307,638
78,699,177
79,092,673
79,488,136
79,885,577
80,285,004
80,686,429
V-33
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-18. Annual Nationwide Costs For Tier 2 LDT2s
Calendar
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Fixed Cost
($)
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
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Variable Cost
($)
0
0
149,650,348
515,340,381
518,028,642
487,480,951
489,918,356
492,367,948
494,829,787
497,303,936
499,790,456
502,289,408
504,800,855
507,324,860
509,861,484
512,410,791
514,972,845
517,547,710
520,135,448
522,736,125
525,349,806
527,976,555
530,616,438
533,269,520
535,935,868
538,615,547
541,308,625
544,015,168
546,735,244
549,468,920
552,216,264
Total Cost
($)
0
0
177,375,502
607,757,561
610,445,822
579,898,131
582,335,536
557,059,974
494,829,787
497,303,936
499,790,456
502,289,408
504,800,855
507,324,860
509,861,484
512,410,791
514,972,845
517,547,710
520,135,448
522,736,125
525,349,806
527,976,555
530,616,438
533,269,520
535,935,868
538,615,547
541,308,625
544,015,168
546,735,244
549,468,920
552,216,264
V-34
-------
Chapter V: Economic Impact
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
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Fixed Cost
($)
869,782
2,609,345
7,538,109
19,715,055
28,992,728
28,122,946
26,383,382
21,454,618
9,277,673
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Variable Cost
($)
8,674,674
26,928,407
79,344,616
213,951,882
324,084,012
310,986,090
300,078,649
301,579,042
303,086,937
304,602,372
306,125,384
307,656,011
309,194,291
310,740,262
312,293,964
313,855,433
315,424,711
317,001,834
318,586,843
320,179,778
321,780,676
323,389,580
325,006,528
326,631,560
328,264,718
329,906,042
331,555,572
333,213,350
334,879,417
336,553,814
338,236,583
Total Cost
($)
9,544,455
29,537,753
86,882,725
233,666,936
353,076,740
339,109,036
326,462,031
323,033,661
312,364,610
304,602,372
306,125,384
307,656,011
309,194,291
310,740,262
312,293,964
313,855,433
315,424,711
317,001,834
318,586,843
320,179,778
321,780,676
323,389,580
325,006,528
326,631,560
328,264,718
329,906,042
331,555,572
333,213,350
334,879,417
336,553,814
338,236,583
V-35
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-20. Annual Nationwide Costs For Tier 2 LDT4s and MDPVs
Calendar
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Tier 2 Tier 2 Tier 2
Fixed Costs Variable Total Costs
($) Costs ($)
($)
000
000
000
000
5,346,756 61,813,742 67,160,497
15,276,445 180,447,506 195,723,952
15,276,445 173,942,913 189,219,358
15,276,445 160,988,307 176,264,752
15,276,445 161,793,249 177,069,694
9,929,689 162,602,215 172,531,904
0 163,415,226 163,415,226
0 164,232,302 164,232,302
0 165,053,464 165,053,464
0 165,878,731 165,878,731
0 166,708,125 166,708,125
0 167,541,665 167,541,665
0 168,379,373 168,379,373
0 169,221,270 169,221,270
0 170,067,377 170,067,377
0 170,917,714 170,917,714
0 171,772,302 171,772,302
0 172,631,164 172,631,164
0 173,494,319 173,494,319
0 174,361,791 174,361,791
0 175,233,600 175,233,600
0 176,109,768 176,109,768
0 176,990,317 176,990,317
0 177,875,268 177,875,268
0 178,764,645 178,764,645
0 179,658,468 179,658,468
0 180,556,760 180,556,760
OBDII&
ORVR
for MDPVs
($)
5,907,154
6,074,415
6,173,995
6,204,865
6,235,889
6,267,068
6,298,404
6,329,896
6,361,545
6,393,353
6,425,320
6,457,446
6,489,733
6,522,182
6,554,793
6,587,567
6,620,505
6,653,607
6,686,875
6,720,310
6,753,911
6,787,681
6,821,619
6,855,727
6,890,006
6,924,456
6,959,078
6,993,874
7,028,843
7,063,987
7,099,307
Total With
OBDII&ORVR
for MDPVs
($)
5,907,154
6,074,415
6,173,995
6,204,865
73,396,386
201,991,020
195,517,762
182,594,648
183,431,239
178,925,257
169,840,545
170,689,748
171,543,197
172,400,913
173,262,918
174,129,232
174,999,878
175,874,878
176,754,252
177,638,023
178,526,213
179,418,844
180,315,939
181,217,518
182,123,606
183,034,224
183,949,395
184,869,142
185,793,488
186,722,455
187,656,068
V-36
-------
Chapter V: Economic Impact
Table V-21 (A). Annual Nationwide Costs For Tier 2 LDVs, LDTs and MDPVs
Calendar
Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Tier 2
Fixed Costs
($)
64,889,954
130,649,690
191,018,792
267,887,764
282,512,192
227,551,928
161,792,192
101,423,090
24,554,118
9,929,689
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Tier 2
Variable Costs
($)
197,981,026
394,185,715
637,205,332
1,108,451,819
1,267,544,205
1,344,643,272
1,331,497,286
1,324,330,452
1,330,952,104
1,337,606,865
1,344,294,899
1,351,016,374
1,357,771,455
1,364,560,313
1,371,383,114
1,378,240,030
1,385,131,230
1,392,056,886
1,399,017,171
1,406,012,256
1,413,042,318
1,420,107,529
1,427,208,067
1,434,344,107
1,441,515,828
1,448,723,407
1,455,967,024
1,463,246,859
1,470,563,093
1,477,915,909
1,485,305,488
Tier 2
Total Costs
($)
262,870,980
524,835,406
828,224,125
1,376,339,582
1,550,056,397
1,572,195,200
1,493,289,478
1,425,753,541
1,355,506,222
1,347,536,554
1,344,294,899
1,351,016,374
1,357,771,455
1,364,560,313
1,371,383,114
1,378,240,030
1,385,131,230
1,392,056,886
1,399,017,171
1,406,012,256
1,413,042,318
1,420,107,529
1,427,208,067
1,434,344,107
1,441,515,828
1,448,723,407
1,455,967,024
1,463,246,859
1,470,563,093
1,477,915,909
1,485,305,488
Including MDPV
OBDIIandORVR
Costs ($)
268,778,135
530,909,821
834,398,119
1,382,544,447
1,556,292,286
1,578,462,268
1,499,587,881
1,432,083,437
1,361,867,767
1,353,929,907
1,350,720,219
1,357,473,820
1,364,261,189
1,371,082,495
1,377,937,907
1,384,827,597
1,391,751,735
1,398,710,493
1,405,704,046
1,412,732,566
1,419,796,229
1,426,895,210
1,434,029,686
1,441,199,835
1,448,405,834
1,455,647,863
1,462,926,102
1,470,240,733
1,477,591,936
1,484,979,896
1,492,404,796
V-37
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-21 (B). Non-Annualized Nationwide Vehicle Costs For Tier 2 LDVs, LDTs and
MDPVs
Calendar
Year
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
Tier 2
Fixed Costs
($)
0
158,787,057
160,915,431
214,584,860
255,856,575
97,985,003
103,494,735
15,074,888
10,243,046
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Tier 2
Variable Costs
($)
0
0
0
0
197,981,026
394,185,715
637,205,332
1,108,451,819
1,267,544,205
1,344,643,272
1,331,497,286
1,324,330,452
1,330,952,104
1,337,606,865
1,344,294,899
1,351,016,374
1,357,771,455
1,364,560,313
1,371,383,114
1,378,240,030
1,385,131,230
1,392,056,886
1,399,017,171
1,406,012,256
1,413,042,318
1,420,107,529
1,427,208,067
1,434,344,107
1,441,515,828
1,448,723,407
1,455,967,024
1,463,246,859
1,470,563,093
1,477,915,909
1,485,305,488
Tier 2
Total Costs
($)
0
158,787,057
160,915,431
214,584,860
453,837,601
492,170,718
740,700,067
1,123,526,707
1,277,787,251
1,344,643,272
1,331,497,286
1,324,330,452
1,330,952,104
1,337,606,865
1,344,294,899
1,351,016,374
1,357,771,455
1,364,560,313
1,371,383,114
1,378,240,030
1,385,131,230
1,392,056,886
1,399,017,171
1,406,012,256
1,413,042,318
1,420,107,529
1,427,208,067
1,434,344,107
1,441,515,828
1,448,723,407
1,455,967,024
1,463,246,859
1,470,563,093
1,477,915,909
1,485,305,488
V-38
-------
Chapter V: Economic Impact
B. Gasoline Desulfurization Costs
1. Overview of Changes Since the NPRM
In the NPRM, we indicated that we expected to work with the Department of Energy
(DOE) in using the Oak Ridge National Laboratory (ORNL) refinery model to estimate gasoline
desulfurization costs. However, we discovered that the ORNL refinery model did not contain
representations of certain technologies which we believe are important in the context of
desulfurizing gasoline, and this was revealed in several of the early modeling case runs which
were conducted by DOE. Thus, we continue to use our refinery model, with a number of
adjustments discussed below, to estimate the gasoline desulfurization costs. We compare our
refinery modeling results to those by DOE, and other cost studies which we received during this
last year, after presenting our cost analysis and results. In general, these other cost studies
support our cost estimates.
One of the principal comments to the NPRM which we wanted to address in our FRM
cost study is that for the NPRM commenters stated that we inappropriately based our cost
estimates on CDTech and and Mobil Oil's Octgain 220 desulfurization technologies which have
not yet been "commercially proven."10 Some refiners feel that these technologies will not have
been operating long enough prior to when they have to decide on what technology they will want
to use. Thus, these refiners may choose among the several commercially proven desulfurization
technologies available today. We incorporated this point of view in our cost analysis for the final
rule by assuming that some refiners in 2004 will use today's proven technologies.
Similarly, we became aware that technologies which desulfurize gasoline through
adsorption, instead of hydrotreating, are commercially available starting this year. Since these
technologies appear to desulfurize gasoline much more efficiently than other processes available
today, we believe that a number of refiners will use these technologies, but to only a very limited
extent starting in 2005, and increasing after that. We are assuming that these technologies will
be used later on because these technologies are so new, and very different from other
desulfurization technologies. A more elaborate discussion on all these desulfurization
technologies can be found in Section IVB, which is the section containing our discussion of the
feasibility of meeting the gasoline sulfur requirements. Many of the small refineries, which must
meet a much less stringent set of interim phase-in requirements which will likely allow them to
push off their capital investments until 2007 and 2008, are expected to take advantage of this
revolutionary technology. The mix of technologies projected to be used in each year is
a. Our understanding of what refiners generally mean when they say a process is
commercially proven is that a process has operated successfully for at least two years in a
refinery producing a refinery product.
V-39
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
summarized below in the section on technology cost.
We also received comments on our determination that available desulfurization capacity
is available and will be used to desulfurize gasoline first before additional investments are made.
Our analysis of the 1996 API/NPRA survey of refining operations and gasoline quality for the
NPRM showed that fluidized catalytic cracker feed hydrotreaters are not operating at capacity
which we used for an initial reduction in gasoline sulfur. The commenter claimed that these
units are already operating at capacity, contrary to what is shown in API/NPRA survey. While
we do not have information from refiners to verify or dispute such claims, we were able to
address this issue through our analysis of gasoline sulfur levels. Refiners must report their
gasoline sulfur levels to EPA to satisfy the RFG and Antidumping program reporting
requirements. We analyzed the 1998 gasoline sulfur levels and found that the average sulfur
level of domestically produced gasoline dropped from 314 ppm to about 270 ppm. This
significant drop in sulfur level may have occurred with the use of excess capacity available from
FCC feed hydrotreaters, and probably to meet the requirements of the 1998 requirements of the
Reformulated Gasoline Program. Consistent with this new data on gasoline sulfur levels and our
assumption that these sulfur reductions resulted from increased FCC feed hydrotreater use, we
adjusted the gasoline pool sulfur levels using the 1998 gasoline sulfur data and dropped any
assumptions that current gasoline sulfur levels could be reduced with existing FCC feed
hydrotreaters. These adjustments made for each PADD are presented below in the section on
blendstocks.
We applied two changes to the Octgain cost estimate methodology used in the NPRM
which improved our cost estimates for this analysis for the FRM, and this improvement also
applied to other fixed bed hydrotreaters as well. In the NPRM, we assumed that the Octgain unit
would be used exclusively to treat the entire FCC naphtha stream. However, Mobil Oil, and the
other vendors of these fixed bed desulfurization technologies, recommend that their processes be
used with a type of distillation column called a splitting column and a catalytic extractive
desulfurization unit for treating the light FCC naphtha. For fixed bed hydrotreaters, this
combination seems to provide a high level of desulfurization at the lowest cost, so we used it for
this analysis. We also based our NPRM cost estimate on the use of a FCC naphtha splitter which
was inappropriate for the task. The naphtha splitter we used is for breaking out individual
streams for additional processing, such as for separating out olefins for petrochemicals, or
producing MTBE. However, for the simple job of creating two substreams for hydrotreating
purposes, it is not necessary to boil away the heavier stream, thus the capital and operating costs
are much lower. We obtained the cost for using such a splitting column from Mobil Oil. This
cost agrees well with the cost of CDTech's CDHydro column which functions in this manner, so
we believe the cost estimate from Mobil Oil is reasosnable and used it in this analysis.
We received a number of comments concerning the cost to refiners of meeting the 80
ppm cap standard. Refiners reported that if the FCC naphtha hydrotreater goes down, then high
sulfur FCC naphtha would likely have to be either stored up or sold off until the hydrotreater can
V-40
-------
Chapter V: Economic Impact
be brought on line. Then the untreated, high sulfur blendstock must be dealt with. In most other
cost studies, the contractor provided a cost estimate to cover this situation. We added the cost of
a storage tank to our cost analysis which would allow for such storage of high sulfur blendstock.
Furthermore, we provided excess desulfurization capacity for treating short term stores of high
sulfur naphtha. These adjustments to our cost estimation methodology show up in our estimated
cost for complying with the low sulfur program.
We maintained many of the aspects of the NPRM analysis. We performed our cost
analysis on a PADD-by-PADD basis, based on gasoline production in each PADD (not gasoline
consumption). Each PADD is represented by a single refinery which consists of refining units
having the average capacity of all refineries of that PADD and which produces gasoline having
the average sulfur level of that PADD. This allows us to compare the cost of desulfurizing
gasoline between different parts of the country which allowed us to address some of the
comments which we received. Like the NPRM, we are assuming that the cost for California
refiners to produce non-California low sulfur gasoline is the same as the cost of producing low
sulfur gasoline in the rest of the country. Since California refiners are already treating all their
gasoline blendstocks, this assumption is probably very conservative. For calculating capital,
fixed and variable operating costs, our methodology for the final rule is very similar to what we
did for the NPRM, with some modifications, which are outlined below in their respective
sections. Our cost analysis is not incremental, studying the cost of a progression of gasoline
desulfurization levels, like the analysis for the NPRM, instead we only evaluated the cost of
achieving 30 ppm, and we are providing an analysis for meeting a 5 ppm standard, and reviewed
the Alliance's cost study for achieving 5 ppm gasoline.
2. Cost Estimation Methodology
a. Technology and Cost Inputs
As we stated above, we are basing our cost analysis for the final rule on a larger group of
desulfurization technologies. To facilitate cost calculations with all these desulfurization
technologies, we are assigning these technologies into three different groups. The first group
comprises those technologies which have already had at least two years of commercial
experience. The second group is comprised of CDTech and Octgain 220, which are the
improved desulfurization technologies upon which we based the NPRM gasoline desulfurization
costs. As stated in the NPRM these technologies are either being demonstrated now, or will start
to be demonstrated in the next few months, as described in Chapter IV. The third group
comprises desulfurization technologies which work through adsorption. Even though these
adsorption technologies are commercially available now, they are newer and different enough
from the other technologies that we felt they should be placed into their own group. Because
they are newer and significantly different, we believe that most refiners would likely be hesitant
in signing a licensing agreement without prior commercial experience, even those refiners which
V-41
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
would be willing to risk using a technology such as CDTech which has partial current
commercial experience. These technologies are discussed in more detail in the feasibility
section. The technologies considered which fall into these various categories are summarized in
Table V-22.
Table V-22. List of Desulfurization Technologies by Technology Group
Technology Group
Proven Technologies
Improved Technologies
Adsorption
Technologies
Desulfurization Technology
Exxon Scanfining, IFF Prime G, Mobil Oil Octgain 125
CDTech CDHydro/HDS, Mobil Oil Octgain 220
Black and Veatch IRVAD, Phillips S-zorb
It is important to point out that there are other desulfurization technologies available
which refiners may use. For example, UOP has developed an improved desulfurization
technology, and Mobil Oil licenses another desulfurization process named Octgain 100,
distinguished by the different catalyst used in the process. However, we decided, as a matter of
practicality, to not try to represent all technologies in our cost analysis. It is also important to
point out that although Octgain 125 will likely be installed in sour crude oil refineries after 2004,
as the program is phasing in, we model the cost of desulfurizing gasoline based on "typical"
refineries with average gasoline sulfur levels. For these moderate applications, Mobil Oil
recommends that these typical refineries use the Octgain 220 process. It is for this reason that we
do not include the Octgain 125 process in the second group of technologies.
These technologies, by virtue of their respective groups, are assumed to be installed for
startup in certain years, consistent with what the perceived status is of the technology when a
refiner must make the decision on a desulfurization technology (approximately 3- 4 years before).
We believe that of the refiners which must meet one of sulfur requirements in 2004, half of them
will install a proven technology, while the other half of the refiners would be more willing to rely
on a technology which has not been proven. A refiner may use the unproven technology for a
variety of reasons. For example, a refiner with poor refining profit margins may assume the risk
of using an unproven technology in the hope of desulfurizing its gasoline at a lower cost which
will help the refiner to improve its refining margins. Another reason why a refiner may choose
an unproven technology is that a refiner may have had a very positive experience with a licensor
that could convince the refiner to use that licensor's technology despite whether the technology
has been proven or not. Our assumptions of the mix of technologies to be installed for use
starting in any one year of the phase-in is summarized in Table V-23 below. Since there are
multiple desulfurization technologies in each group, for our cost analysis, we presume that
refiners would use these technologies equally, rather than attempt to determine if refiners would
V-42
-------
Chapter V: Economic Impact
tend to use one more than another, and then project what the percentage of each desulfurization
technology used.
Table V-23. Projected Use of Desulfurization Technology Types by Refiners During the
Phase-in Period (from Section IV-B)
Initial Year of Full Operation
2004
2005
2006
2007 & 2008
Mix of Technology Groups used
1/2 Proven,
3/4 Improved
1/2 Improved
1/4 Improved
1/2 Improved
, 1/4 Adsorbent
, 1/2 Adsorbent
, 3/4 Adsorbent
As discussed in Chapter IV, a number of desulfurization units are projected to begin
operating prior to 2004. Five of these units will be demonstration units for the improved and
adsorbent technologies. Two to five more units are expected to be operated by refiners desiring
to generate early credits or allotments and to use low sulfur gasoline as a marketing tool. These
latter units are likely to be a mixture of proven and improved technologies, much like that
projected for 2004, possibly with a greater fraction of proven technology. Overall, we
represented the cost of these pre-2004 units using the 2004 technology mix.
We acquired process operations information on each of these technologies through our
participation with the National Petroleum Council (NPC). During 1999 the NPC was conducting
a study of how potential fuel quality control programs will affect the cost and producibility of
domestically produced motor vehicle fuels. During this study, the Technology Workgroup of the
NPC requested input cost data from many different licensors of gasoline FCC naphtha
desulfurization processes to study the cost of desulfurizing gasoline. We obtained that
information and we used it in our cost study.14 This cost input data is summarized in Table V-24.
V-43
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-24. Process Operations Information for FCC Naphtha Desulfurization Processes
(All Technologies are 95% Efficient for Desulfurizing Gasoline)
Capacity
(MMbbl/day)
Capital Cost
(MM$)
Hydrogen
Consumption
(SCF/bbl)
Electricity
(KwH/bbl)
HP Steam
(Lb/bbl)
Fuel Gas
(BTU/bbl)
Catalyst Cost
($/bbl)
Cooling Water
(Gal/bbl)
Yield Loss (%)
Octane Loss
(R+M)/2
Octgain 125
15,000
14.9
370
2.0
-
84,400
0.43
250
5
0
Exxon
Scanfming
25,000
16.8
77
0.61
44.8
14,500
0.22
135
0
1.0
IFF Prime G
24,000
21.7
126
1.3
63
9300
0.01
130
0.8
1.3
Octgain 220
31,000
23.8
130
1.5
75
35,600
0.22
225
0.7
0.1
CDTech
30,000
18.5
102
0.44
24.4
33,000
0.25
53.3
0
1.0
Black&Veatch
IRVAD
30,000
17.9
Negligible
1.82
0
18,300
0.27
16.7
4.5
(2.0)
Phillips S-Zorb
25,000
13.8
70
-
4.5
39,000
0.27
130
0
0.75
V-44
-------
Chapter V: Economic Impact
Besides these desulfurization technologies, we used additional technologies in our
refinery modelling analysis. Depending on the desulfurization case which we were modeling, we
used extractive desulfurization units for desulfurizing both light FCC naphtha and light straight
run. We also needed to include distillation or splitting columns for fixed bed hydrotreaters for
separating the FCC naphtha into two different streams so the light FCC naphtha could be treated
by extractive desulfurization and the medium and heavy FCC naphtha could be treated by the
hydrotreater. Most of the vendors which license fixed bed desulfurization processes already
include the operating and capital costs of both the extractive desulfurization unit and splitting
columns in their information submissions, however, we needed to add these costs to the Octgain
costs. The process operation information for these other technologies are summarized in Table
V-25. The splitting column inputs are from Mobil Oil which provided the information to the
NPC Technology Workgroup along with information on their Octgain units. As we stated above,
for the NPRM we used the splitter input information from the Oak Ridge National Laboratory
refinery model. However, that splitter is really for creating multiple separate substreams out of
the FCC naphtha while the Mobil Oil splitter is for creating a simple cut, which is all that is
needed for this application. The capital and operating costs for the Mobil Oil splitter are much
lower as a result. We provide the ORNL splitter information as a comparison.
In this analysis we also included costs for half of refiners adding an FCC naphtha storage
tank.15 The purpose of the storage tank would be for refiners to store up nonhydrotreated FCC
naphtha, for up to 10 days, during a shutdown of the FCC naphtha hydrotreater. During the
shutdown, the high sulfur blendstock cannot be blended into gasoline because it would cause the
gasoline pool to exceed the 80 ppm cap. Then, after the hydrotreater is brought back on line, the
high sulfur FCC naphtha in the storage tank would either be sent to the hydrotreater, in quantities
which would not exceed the hydrotreater capacity, or it would be slowly blended into finished
gasoline in a manner which allows the refiner to meet the 80 ppm cap. We sized the FCC
naphtha hydrotreater large enough to handle the stored naphtha. The capital costs for the storage
tank are summarized in Table V-25.
V-45
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-25. Process Operations Information for Additional Units
used for Desulfurization Cost Analysis
Capacity
(MMbbl/day)
Capital Cost
(MM$)
Electricity
(KwH/bbl)
HP Steam
(Lb/bbl)
Fuel Gas
(BTU/bbl)
Cooling Water
(Gal/bbl)
Operating Cost
($/bbl)
Extractive
Desulfurization
10000
3.5
—
—
—
—
0.06
Splitter (Mobil
Oil)
(used in this
analysis)
50000
4.1
0.17
36
—
13
—
Splitter
(ORNL)
(used in
NPRM)
20000
10.7
2.5
10
90000
—
—
FCC Naphtha
Storage Tank
50,000 bbls
0.75
—
—
—
—
none*
* No operating costs are estimated directly, however both the ISBL to OSBL factor and the capital contingency
factor used for desulfurization processes is used for the tankage as well, which we believe to be excessive for
storage tanks so it is presumed to cover the operating cost.
b. Capital Costs
Capital costs are the one-time costs incurred by purchasing and installing new hardware
in refineries. The capital costs are calculated similar to how they were calculated for the NPRM,
with some differences. Capital costs for a particular processing unit are supplied by the vendors
for a particular volumetric capacity and desulfurization efficiency, and these costs are adjusted to
match the volume of the particular case being analyzed using the sixth tenths rule.11 The
calendar day volume is increased by 7 percent to size the hydrotreating unit for stream days, the
11 The capital cost is estimated at this other throughput using an exponential equation termed the "six-
tenths rule." 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
splitters and 0.60 for desulfurization units.
V-46
-------
Chapter V: Economic Impact
days which the unit is operating. Unlike the NPRM, the hydrotreating unit volume is not
increased by 15 percent as a safety factor. Instead, a 15 percent factor is applied to the capital
costs after the outside battery limit costs and added capital installation costs (for higher labor
rates) were calculated, and a 10 percent factor is applied to the operating costs. These two
contingency factors are meant to account for costs not accounted for in the principal calculation,
such as running the amine and sulfur plants harder for addressing the additional sulfur removed.
The 5 percent capital adjustment factor applied to noncommercially demonstrated units for the
NPRM is maintained in the final rule. An additional 5 percent factor is applied to size the units
larger so that the unit can process untreated blendstock stored up during a shutdown or
turnaround.
The capital costs are adjusted further to account for the off site costs and differences in
labor costs relative to the Gulf Coast. The same method for calculating the offsite costs and
accounting for differences in labor costs used in the NPRM, which is from Gary and Handewerk,
is used here.16 The offsite and labor factors used for each PADD are summarized here.
Table V-26. Offsite and Location Factors Used for Estimating Capital Costs
Offsite
Factor
Location
Factor
PADD1
1.25
1.5
PADD 2
1.25
1.3
PADD 3
1.2
1.0
PADD 4
1.5
1.4
PADD 5
1.25
1.2
The same economic assumptions used in the NPRM for amortizing the capital costs over
the volume of gasoline produced are used for this analysis. These assumptions and the resulting
capital amortization cost factors are summarized below in Table V-27. These capital
amortization cost factors are used in the following section on the cost of desulfurizing gasoline to
represent the capital cost as a cents per gallon cost.
V-47
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-27. Economic Cost Factors Used in Calculating the 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%
10%
Resulting
Capital
Amortization
Factor
0.11
0.12
0.16
c. Fixed Operating Cost
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.
We are using the same cost factors to estimate fixed operating costs in this analysis as what we
used for the analysis for the NPRM.
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 for PADDs other than PADD 3. This factor is based on the maintenance factor used
in the ORNL refinery model.
Other fixed operating costs are accounted for as well, and these generic cost factors are
also 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 (after adjusting for offsite costs and location factor) 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.
d. 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. The operating cost demands (utilities, hydrogen, octane
V-48
-------
Chapter V: Economic Impact
and yield loss) are from the licensors which license the gasoline desulfurization technologies and
the basis for the values is 95 percent FCC naphtha desulfurization, since that level of
desulfurization adequately exceeded the need by each average refinery modeled for reaching the
sulfur target (30 ppm pool sulfur). We used the same variable operating cost factors, for such
costs as utilities, hydrogen and octane costs, in this analysis as we used in the NPRM. We
summarized these costs in the following table. We are no longer showing the costs for residual
oil and diesel fuel, since we are no longer projecting the use of excess FCC feed hydrotreater
capacity in achieving the 30 ppm standard. We did make one change in our operating cost
calculation methodology. In the NPRM, we estimated the cost of producing steam based on the
premise that heat demand for the steam is met by burning fuel gas, and we used the estimated
price of fuel gas as our cost basis. For this analysis we are using the same methodology, except
our costs are increased upward by a factor of two to be consistent with published cost estimation
methodology which estimates the cost of supplying steam as two times the cost of the fuel gas
consumed.17 Our octane cost estimation methodology used for the analysis in the NPRM was
corroborated by the cost estimating work by API, which estimated an octane cost just less than
ours based on refinery modeling, thus we maintained this cost estimation methodology in our
cost analysis.18 These costs are summarized in Table V-28.
V-49
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-28. Summary of Costs Taken From EIA and NPC Data Tables *
Electricity
(c/KwH)
LPG (c/Gal)
Gasoline
(c/Gal)
Octane Cost
(cents)
Octane
Spread
(R+M)/2
Fuel Gas
($/MMbtu)
Hydrogen
Cost
($/MSCF)
PADD 1
5.9
19.7
27.0
4.3
5.7
3.75
2.5
PADD 2
3.9
18.4
25.9
2.8
5.2
3.75
2.5
PADD 3
4.2
16.5
24.9
3.5
5.4
3
2.0
PADD 4
3.4
17.8
28.9
11.4
5.2
4.5
3.0
PADD 5OC
5.4
19.7
30
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.
e.
Determination of Blendstock Sulfur Levels
We maintained the alkylate, coker, and light straight run sulfur levels estimates which we
summarized in the NPRM; however, we made an adjustment in the FCC gasoline sulfur levels
based on the lower average gasoline sulfur levels in 1998. For the NPRM, we provided a sulfur
balance for an average refinery in each PADD to establish the volumes and sulfur levels of
blendstocks which contribute significantly to the pool sulfur level (FCC 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
V-50
-------
Chapter V: Economic Impact
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 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. The averaged value
was determined to be 22 ppm, however, one refinery had a sulfur level of over 130 ppm. Since
the promulgation of the NPRM, we contacted that refiner with the high alkylate sulfur level and
found out that the operations of their alkylate unit has improved since 1990, and their alkylate is
now averaging about 20 ppm sulfur. When we averaged that sulfur level with the alkylate sulfur
levels of other refineries, the average alkylate sulfur level dropped to 7 ppm for those refineries.
For the NPRM, we also contacted several refining industry consultants to find out what
alkylate sulfur levels they used in their refinery models. The alkylate sulfur levels in those
refinery models averaged about 10 ppm (the values ranged from 0 to 25 ppm). For the final rule,
we are maintaining the 10 ppm average sulfur level for alkylate we used for the NPRM, since
both the RFG data base and refining industry consultants generally support this level.
Other blendstocks, such as isomerate, reformate, raffmate, 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. We believe that for
an analysis of the cost of achieving a 30 ppm gasoline pool sulfur level, that this assumption is
appropriate. Even if their sulfur contribution is somewhat higher, both the 15 andlO percent
capital and operating cost contingency factors and the excess 5 percent treating capacity of the
FCC naphtha hydrotreater are conservative estimates, which could offset the additional
desulfurization treatment cost of these other streams (or for further desulfurizing FCC naphtha to
compensate for the small amount of sulfur in these other streams).
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. If there was disagreement, we adjusted one or the other, as
summarized below.
For the NPRM we assumed that projected unused FCC feed hydrotreating capacity would
V-51
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
be used first by the average refinery to reduce their FCC naphtha sulfur level, and additional
hydrotreating would be estimated from the revised FCC naphtha sulfur level. As stated above,
comments which we received on our proposed rule stating that such capacity does not exist
raised uncertainty about how much excess capacity there might be both now and projecting
availability in the 2004 timeframe.
New analysis since the promulgation of the NPRM of the gasoline sulfur levels in 1998
shows that gasoline sulfur levels dropped significantly since 1997, possibly due to refiners
having to meet the federal RFG and Antidumping requirements using the Complex Model. One
possible explanation of how this reduction came about was that refiners used their existing spare
FCC feed hydrotreater capacity to reduce their gasoline sulfur levels. Assuming that this is the
case, we will use the new gasoline sulfur levels for each PADD to recalculate the FCC naphtha
sulfur levels. All sulfur levels calculated are volume-weighted, not refinery-weighted. These
adjustments are summarized below in the section on each PADD.
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,
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. The gasoline production
volume for the average refinery in PADD 1 is about 77 thousand barrels per day.
We analyzed whether these figures need to be adjusted to account for the implementation
of Phase II RFG in 2000. Phase n 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 Phase II
RFG to be about 150 ppm, no changes in sulfur level are expected to occur to produce Phase II
RFG.
In 1998, PADD 1 gasoline sulfur levels averaged 189 ppm, which is 27 ppm lower than
the previous value. FCC sulfur levels are recalculated to be 381 ppm based on the lower pool
sulfur level.
V-52
-------
Chapter V: Economic Impact
The PADD 1 blendstock sulfur levels and relative volumes are summarized in Table V-29.
Table V-29. PADD 1 Gasoline Blendstock and Pool Sulfur Levels and Pool Fractions
Sulfur (ppm) NPRM
FRM
Percentage of
gasoline pool
Contribution to pool
(ppm) NPRM
FRM
FCC
442
381
42
185
160
A Iky I ate
10
10
1
Straight
Run
343
4
14
Coker
3289
0.44
14
Gasoline Pool
Sulfur Level
214
189
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 effect on the average sulfur level of PADD 2. The gasoline production
volume for the average refinery in PADD 2 is about 66 thousand barrels per day.
In 1998, PADD 2 gasoline sulfur levels averaged 276 ppm, which is 62 ppm lower than
the previous value. FCC sulfur levels are recalculated to be 745 ppm based on the lower pool
sulfur level.
The PADD 2 blendstock sulfur levels and relative volumes are summarized in Table V-30.
V-53
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-30. PADD 2 Gasoline Blendstock and Pool Sulfur Levels and Pool Fractions
Sulfur (ppm) NPRM
FRM
Percentage of
gasoline pool
Contribution to pool
(ppm) NPRM
FRM
FCC
924
745
35
323
261
A Iky I ate
10
13
1
Straight
Run
397
3.4
14
Coker
0
0
0
Gasoline Pool
Sulfur Level
338
276
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 gasoline production volume for the average refinery in PADD 3 is
about 75 thousand barrels per day.
In 1998, PADD 3 gasoline sulfur levels averaged 288 ppm, which is 19 ppm lower than
the previous value. FCC sulfur levels are recalculated to be 673 ppm based on the lower pool
sulfur level.
The PADD 3 blendstock sulfur levels and relative volumes are summarized in Table V-31.
V-54
-------
Chapter V: Economic Impact
Table V-31. PADD 3 Gasoline Blendstock and Pool Sulfur Levels and Pool Fractions
Sulfur (ppm) NPRM
FRM
Percentage of
gasoline pool
Contribution to pool
(ppm) NPRM
FRM
FCC
722
673
40
288
269
A Iky I ate
10
14
1
Straight
Run
139
2.8
4
Coker
3255
0.42
14
Gasoline Pool
Sulfur Level
307
288
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
gasoline production volume for the average refinery in PADD 4 is about 19 thousand barrels per
day.
In 1998, PADD 4 gasoline sulfur levels averaged 282 ppm, which is 17 ppm higher than
the previous value. FCC sulfur levels are recalculated to be 823 ppm based on the higher pool
sulfur level.
The PADD 4 blendstock sulfur levels and relative volumes are summarized in Table V-32.
V-55
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-32. PADD 4 Gasoline Blendstock and Pool Sulfur Levels and Pool Fractions
Sulfur (ppm) NPRM
FRM
Percentage of
gasoline pool
Contribution to pool
(ppm) NPRM
FRM
FCC
762
823
31
236
255
A Iky I ate
10
12
1
Straight
Run
122
21
26
Coker
0
0
0
Gasoline Pool
Sulfur Level
263
282
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 gasoline production volume for the average refinery in PADD 5,
not including California refineries, is about 27 thousand barrels per day.
In 1998, PADD 5 gasoline sulfur levels averaged 301 ppm, which is 205 ppm lower than
the previous value. FCC sulfur levels are recalculated to be 710 ppm based on the lower pool
sulfur level.
The PADD 5 outside of California blendstock sulfur levels and relative volumes are summarized
inTableV-33.
V-56
-------
Chapter V: Economic Impact
Table V-33. PADD 5 Outside of California Gasoline Blendstock
and Pool Sulfur Levels and Pool Fractions
Sulfur (ppm) NPRM
FRM
Percentage of
gasoline pool
Contribution to pool
(ppm) NPRM
FRM
FCC
1197
710
42
503
298
A Iky late
10
10
1
Straight
Run
41
5.9
2
Coker
0
0
0
Gasoline Pool
Sulfur Level
506
301
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. The future volume of gasoline produced is based on the increase
in consumption summarized later on in this Section. These values are the same as those used in
the NPRM.
These values are summarized below in Table V-34.
V-57
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-34. Projected Volume of Gasoline Produced by an Average Refinery in each
PADD and Projected Gasoline Consumption 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.
f.
Phase-In Desulfurization
To estimate the capital and per-gallon cost of the gasoline desulfurization program based
on our projected use of gasoline desulfurization technologies, we needed to estimate the volume
of gasoline each year which would have to be desulfurized to enable refiners to meet the standard
which applies in that particular year. To make this estimation, we needed to project for what year
refineries will need to have new capital investments installed to meet the requirements of this
gasoline sulfur program. We made such an assessement, accounting for the small refiner and
ABT programs contained in the final rule, as well as the geographic phase-in, and it is
summarized in Section IV.
Based on this analysis we tallied the production volume of gasoline desulfurized for each
year and by PADD. This allowed us to calculate our estimated capital and per-gallon costs each
year. Our estimate incorporates the temporary exemption for the geographical phase-in as well
as the small refiners. These volumes are summarized in Table V-35.
V-58
-------
Chapter V: Economic Impact
Table V-35. Cumulative Fraction of the Gasoline Pool Desulfurized by PADD and by Year
2004*
2005
2006
2007
2008+**
PADD1
0.25
0.36
0.99
0.99
1
PADD 2
0.63
0.68
0.93
0.93
1
PADD 3
0.65
0.74
0.96
0.97
1
PADD 4
0.15
0.15
0.15
0.88
1
PADD
50C
0.66
0.66
1
1
1
* Includes early desulfurization units prior to 2004.
** Includes gasoline already meeting the 30 ppm standard which we included in our baseline
gasoline sulfur level for estimating cost, thus it is appropriate to assign some cost to these gallons
of gasoline.
g. Decreasing Costs in Future Years
Like the analysis completed for the NPRM, we are presuming that desulfurization costs
decrease in future years, however, our methodology is somewhat different. For the NPRM, we
presumed that operating costs decrease due to an improvement in catalyst technology. Based on
this presumption, we projected that operating costs, including catalyst cost, hydrogen cost, octane
cost, and yield loss, would decrease by 20 percent after two years. We also assumed that with
debottlenecking, fixed operating costs would stay the same in total and decrease on a per-barrel
basis.
Our analysis for the Final Rule incorporates operational cost reductions, but not the
debottlencking cost reduction. The presumption here is that refiners will either operate the
proven technologies more efficiently, or they would simply change out the catalyst to use the
lowest cost fixed bed desulfurization catalyst, which would result in a 20 percent reduction in
hydrogen consumption cost, octane recovery cost, yield loss, and catalyst cost starting in the third
year. For example, if refiners initially installed a Mobil Oil Octgain 125 process and then later
on decided to install the Octgain 220 process (which could be changed out after operating the
unit for two years when the catalyst desulfurization efficiency begins to degrade), we estimate,
based on the vendors information and our cost factors, that the Octgain 220 process would lower
the aggregate cost of the desulfurization unit by 20 percent. But based on the operating cost
alone, we estimate the cost savings to be almost 30 percent. Since this case is only one of several
proven technologies and there may not be as dramatic as a reduction for the others, we only used
V-59
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
20 percent as our operational cost improvement. We did not assume that the same operational
benefits applied to improved and adsorption technologies. We do presume, though, that after the
proven and improved desulfurization units reach the end of their economic life, which is after 15
years, they would be replaced by the lower cost adsorption units.
3 The Cost of Desulfurizing Gasoline
a. EPA Costs
The refinery blendstock sulfur levels, the vendor desulfurization technology information,
the various cost inputs, and the various desulfurization assumptions described above were
combined together in our refinery model to estimate the cost of desulfurizing gasoline from the
base sulfur level, down to 30 ppm. As stated above, we presume that refiners would choose a
mix of proven and improved desulfurization technologies to meet the requirements of the first
year of the program. Then for meeting the program requirements after 2004, some refiners
would choose to use lower cost adsorption technologies for 2005, with more and more of them
doing so toward the later years. For each technology group, we presume that equal use of each
technology would be used. To estimate costs for each year based on this methodology, we used
the projected volume of gasoline desulfurized for each PADD during each year of the phase-in
period. To estimate national average costs, we volume weighted the PADD-specific cost
estimates.
Based on this methodology we estimated the aggregate operating and capital cost, and the
per-gallon cost, for the U.S. refining industry as a whole, each year starting in 2004. As expected
the program's per-gallon cost decreases over time as lower cost desulfurization technology is
implemented until 2008 when the last desulfurization units are installed. In 2006, a portion of
the proven technologies' operational costs decrease. After 2008, the costs are constant until 2019
when the initial desulfurization units installed in 2004 reach the end of their useful life, and are
replaced by adsorption units, the lowest cost desulfurization technologies. The aggregate
operating costs increase due to the constant increase in growth in gasoline demand. These costs
are summarized in Table V-36.
V-60
-------
Chapter V: Economic Impact
Table V-36. Estimated U.S. Aggregate Operating and Capital Cost, and Per-Gallon Cost of
Desulfurizing Gasoline to 30 ppm (7% ROI, Before Taxes, $1997)
Year
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
Estimated
Aggregate
Operating Cost
($Billion)*
-
1.21
1.36
1.84
1.95
2.02
2.04
2.05
2.07
2.08
2.10
2.12
2.13
2.14
2.16
2.17
1.65
1.63
1.61
1.63
1.66
1.68
1.71
1.73
1.76
1.78
Estimated
Aggregate Capital
Cost
(SBillion)
2.61**
0.29
1.16
0.34
0.14
-
-
-
-
-
-
-
-
-
-
2.20
0.29
1.24
0.40
0.17
-
-
-
-
-
-
Estimated Total
Aggregate Cost
(SBillion)
2.61
1.50
2.52
2.18
2.09
2.02
2.04
2.05
2.07
2.08
2.10
2.12
2.13
2.14
2.16
4.37
1.94
2.87
2.01
1.80
1.66
1.68
1.71
1.73
1.76
1.78
Estimated Per-
Gallon Cost
(c/gal)
1.95
1.90
1.70
1.71
1.70
1.70
1.70
1.70
1.70
1.70
1.70
1.70
1.70
1.70
1.70
1.32
1.30
1.26
1.26
1.26
1.26
1.26
1.26
1.26
1.26
V-61
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Year
2029
2030
Estimated
Aggregate
Operating Cost
($Billion)*
1.80
1.83
Estimated
Aggregate Capital
Cost
($Billion)
-
-
Estimated Total
Aggregate Cost
($Billion)
1.80
1.83
Estimated Per-
Gallon Cost
(c/gal)
1.26
1.26
* Based on fuel consumption data summarized further below in Section V.
** Includes investments made to produce low sulfur gasoline before 2004 to accumulate credits.
Table V-36 shows that the aggregate capital cost to the U.S. refining industry for meeting
the proposed 30 ppm sulfur standard during the initial phase-in is expected to total about 4.5
billion dollars. The program's phase-in causes the capital investments to be spread out over
several years, with a little more than half of the capital investments being spent either during, or
prior to the year 2004. This level of capital expenditure is less than previous capital expenditures
made by the refining industry for environmental programs. As we discussed in the NPRM,
during the early nineties the U.S. refining industry invested 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 made in the early '90s were incurred by less than three quarters of the refining industry,
we believe that a program requiring the entire industry to spend, on average, about one billion
dollars of capital costs per year over several years is not overly burdensome from an economic
perspective. The relative value of the costs and benefits of this program are discussed in Chapter
VII.
As stated above we estimated per-gallon cost by PADD based on an average refinery for
each PADD using different amortization premises. In Table V-37 below, costs are shown for
amortizing capital at a 7 percent rate of return on investment (ROI) before taxes which is to
represent the cost to society. Then we provide a range of costs which is meant to represent the
cost based on a rate of return on capital consistent with how refiners may recover their capital
costs. This range is 6 to 10 percent ROI after taxes. To simplify this comparison, we are
presenting these per-gallon costs for 2008, the year when the costs stabilize.
V-62
-------
Chapter V: Economic Impact
Table V-37. Post Phase-in Cost (year 2008) of Desulfurizing Gasoline to 30 ppm
Based on Different Capital Amortization Rates
Societal Cost
7% ROI before
Taxes
Capital Payback
(6% ROI, after
Taxes)
Capital Payback
(10% ROI, after
Taxes)
PADD 1
2.00
2.04
2.22
PADD 2
1.65
1.69
1.85
PADD 3
1.52
1.54
1.65
PADD 4
2.32
2.41
2.76
PADD 5OC
2.63
2.67
2.87
National
Average
1.70
1.73
1.87
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 refinery operating cost, producing low sulfur gasoline in PADD 4 is
expected to be the most expensive, and, in the analysis for the NPRM, was about twice as costly
to desulfurize gasoline as PADDs 2 and 3. However, because the PADD 4 refineries are subject
to less stringent interim standards until 2007 and 2008 under the small refiner and geographical
phase-in provisions, the costs are much lower and only 50 percent higher than those of 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 1.7 cents per gallon to desulfurize gasoline
down to 30 ppm in 2008 after the program is fully phased-in.
To help the reader better understand the cost of the program for a typical refinery, the per-
refmery capital and operating costs, and the estimated yearly aggregate capital and operating cost
for each PADD and for the country as a whole of meeting a 30 ppm sulfur standards is
summarized in Table V-38 below.
V-63
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-38. Estimated Average Per-Refinery and Aggregate Capital and Operating Cost
of Desulfurizing Gasoline to 30 ppm
Avg. per-
refmery
capital cost
($MM)
Avg. per-
refmery
operating
cost
($MM)
Aggregate
capital cost
Aggregate
operating
cost
PADD 1
64
20
691
193
PADD 2
50
15
1342
341
PADD 3
38
17
1674
663
PADD 4
26
5
327
53
PADD
5OC
26
11
259
92
National
44
16
4294
1343
Table V-38 shows that, on average, refiners would have to pay out $44 million in capital
costs for each refinery to lower gasoline sulfur to 30 ppm. In addition, each refinery would incur
about 16 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. Since
these figures are averages, larger refineries with high gasoline sulfur levels will experience
higher total costs, while smaller refineries with lower sulfur levels will experience lower total
costs. The aggregate operating cost to the U.S. refining industry is expected to be about 1.3
billion dollars per year.
b.
Other Low Sulfur Cost Studies
i. American Petroleum Institute (API) Study
API funded a study by Mathpro to estimate the cost of desulfurizing gasoline in PADDs
1, 2 and 3 down to 40 ppm.19 Their study was based on CDTech and Mobil Oil Octgain 220 used
in a notional refinery which is designed to represent all the refineries in those three PADDs. That
V-64
-------
Chapter V: Economic Impact
study estimated the cost of desulfurizing gasoline down to 40 ppm to be 2.6 c/gal for Octgain,
and 2.25 c/gal for CDTech. The study amortized capital investment at a 10 percent rate of return,
which is higher than the ROI which we use to evaluate and compare cost-effectiveness. In
addition the Mathpro study allocated 0.5 c/gal for ancillary costs, such as reblending of offspec
batches and accounting for overoptimization. These are costs which Mathpro feels is applicable,
however, Mathpro has not justified these costs.
To compare our two studies, it is important to place their cost analysis on the same basis
as ours. We did that by adjusting their capital cost to reflect a capital amortization rate consistent
with a 7 percent ROI before taxes. We summarized the initial costs and the subsequent
adjustments in the following table. The API costs increase by 0.25 c/gal for meeting a 30 ppm
specification.20 We next adjusted the 30 ppm cost to base the capital costs on a 7 percent ROI,
which decreased the cost to 2.2 c/gal. The costs are even more in line with our costs if some of
the ancillary costs are not justified. These costs are summarized in Table V-39.
Table V-39. API Gasoline Desulfurization Estimate, Adjusted and Compared to EPA's
(API cost adjustments are sequential which leads to the comparison with our costs)
Description
API study initial cost for meeting 40 ppm standard
API Study: Average CDTech & Octgain cost for 40 ppm std.
Adjusted API cost estimate to include incremental cost to meet
30 ppm std. by Mathpro cost study for the Alliance
EPA adjusted API Estimated cost based on 7% ROI before taxes
EPA cost based on CDTech and Octgain 220 7% ROI before
taxes
Cost (c/gal)
CDTech Octgain
2.25 2.6
2.4
2.65
2.2
1.7
ii. National Petrochemical and Refiners Association (NPRA) Study
NPRA funded a study by Mathpro to estimate cost to PADD 4 refineries of meeting a 40
ppm gasoline sulfur standard.21 The study yielded a cost of 5.7 c/gal, however, we reviewed the
bases for the study and a number of assumptions used in the study led to the much higher
gasoline desulfurization cost than our analysis. First, the study assumed that only Octgain 125,
which is a proven desulfurization technology, would be used. Then, their cost inputs for Octgain
125 are the older, conservative ones which were abandoned by Mathpro in the later study funded
by API. Mathpro's analysis of the difference in cost between the two versions of Octgain 125
processes is about 1 c/gal for PADDs 1-3. Furthermore, like our NPRM analysis, the splitting
V-65
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
column used in the NPRA analysis to separate the FCC naphtha into two distinct streams for
selectively treating only the heavier part of the FCC naphtha was for an overly conservative
column for boiling the entire stream, not intended for making a crude cut for gasoline
hydrotreating.
We estimated the cost to PADD 4 refiners to desulfurize their gasoline, based on our
finalized program which exempts PADD 4 refineries for the first three years, and thus we assume
that most will install absorption desulfurization technology. Based on this methodology, we
derived a cost of 2.5 c/gal.
In the process of evaluating that cost, we looked at what the cost would be if PADD 4
refiners had to put in Octgain 125 desulfurization technology, and we can even estimate what it
would cost these refineries if they were to install the full boiling range FCC naphtha splitter
which, of course, is unnecessary for the simple cuts needed for hydrotreating. We used these cost
estimates to adjust the NPRA costs downward to see what NPRA costs might be if they used the
more efficient desulfurization and processing equipment, and revised capital amortization
factors. We estimate that the 5.7 c/gal NPRA cost would decrease to 5.2 c/gal if their capital cost
were amortized by a 7 percent ROI before taxes. Then if their cost estimate would have been
based on the revised Octgain 125 cost, we estimate that their cost would decrease to 4.2 c/gal.
Next, if their estimated cost were based on a more efficient FCC naphtha splitting column, we
estimate that their gasoline desulfurization cost would decrease to 3.5 c/gal. Finally, if their
estimated gasoline desulfurization cost were based mostly on adsorption desulfurization
technology, we estimate that their estimated cost would decrease to about 2 c/gal, which would
be a little higher if their cost estimate would have been for meeting a 30 ppm standard. These
costs are summarized in Table V-40.
V-66
-------
Chapter V: Economic Impact
Table V-40. NPRA PADD 4 Gasoline Desulfurization Estimate, Adjusted and Compared
to EPA's
(NPRA cost adjustments are sequential which leads to the comparison with our costs)
Description
NPRA estimated cost for PADD 4 refineries meeting 40 ppm standard
based on 10% ROI after taxes
Incremental Adjustments by EPA
To 7% ROI before taxes
To reflect new Octgain 125 cost
To reflect optimized splitting column
To reflect more efficient adsorption desulfurization technology
EPA cost for PADD 4 refineries meeting a 30 ppm standard based
primarily on adsorption technology and based on a 7% ROI before taxes
Cost (c/gal)
5.7
5.2
4.2
3.5
1.7
2.5
iii. Association of International Automobile Manufacturers (AIAM) Study
AIAM funded a study by Mathpro to analyze the cost of meeting a 30 ppm standard in
PADD 4 using improved desulfurization technology.22 Mathpro used a spreadsheet to estimate
the cost in a refinery-by-refinery analysis of meeting the low sulfur specification. The study
assumed that CDTech would be the desulfurization technology used. The analysis estimated that
it would cost 3.14 c/gal for PADD 4 refiners to meet the 30 ppm sulfur standard. However, the
cost estimate is based on a 15% ROI, and adjusting the cost estimate to be based on a 7% ROI
before taxes, reduces the cost estimate to 2.41 cents per gallon.
If we only base our cost to PADD 4 refiners of desulfurizing their gasoline on CDTech,
our refinery model estimates that it would cost PADD 4 refiners 3.2 c/gal. Thus, our cost is
much more conservative than that by Mathpro. This most likely reflects the higher labor costs
for the installation of capital for PADD 4 which we use.
V-67
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-41. AIAM Gasoline Desulfurization Estimate for PADD 4, Adjusted and
Compared to EPA's
Description
Mathpro's cost for desulfurizing gasoline to 30 ppm in PADD
4 based on 15%ROI
Mathpro's desulfurization cost based on 7% ROI, before taxes
EPA's cost for desulfurizing gasoline in PADD 4 using
CDTech and based on 7% ROI before taxes
Cost (c/gal)
3.14
2.14
3.2
iv Department of Energy (DOE) Study
The Department of Energy used their refinery modeling resources at the Oak Ridge
National Laboratory to estimate the cost of desulfurizing gasoline for the average refinery.23 24
DOE also sought to determine if desulfurization costs varied significantly between average
refineries and those for whom gasoline desulfurization might be more challenging. To answer
these two questions, they evaluated desulfurization costs for two classes of refineries: mid-
capability and challenged. In their analysis, the mid-capability refineries processed crude oil with
a sulfur content of 1.6 weight percent, partially hydrotreated FCC feed and produced gasoline
with an average sulfur content of 200 - 240 ppm. The challenged group processed crude oil with
a sulfur content of 1.94 weight percent, did not hydrotreat FCC feed and produced gasoline with
an average sulfur content of 500 ppm. The study was parametric, evaluating the cost of
desulfurizing gasoline to 50, 30 and 10 ppm for the mid-capability refinery, and 30 ppm for the
challenged refinery. The estimated costs for 10, 30 and 50 ppm sulfur are summarized in Table
V-42.
V-68
-------
Chapter V: Economic Impact
Table V-42. DOE Gasoline Desulfurization Estimate, Adjusted and Compared to EPA's
Description
Mid-capability refinery
Challenged refinery
Based 10% ROI after taxes
Mid-capability refinery
Challenged refinery
Adjusted to 7% ROI before taxes
EPA national average cost to produce
30 ppm gasoline; 7% ROI before
taxes
Cost (c/gal)
50 ppm
2.1
1.9
30 ppm
2.9
3.4
2.6
2.4
1.7
10 ppm
9.0
6.7
For case where mid-capability refineries produced 30 ppm gasoline, the refinery model
chose CDTech as the FCC naphtha hydrotreater. However, the model also chose to install a FCC
naphtha splitter, and treat some of the light FCC naphtha with a catalytic extractive desulfurizing
unit, and send some of the FCC naphtha to the naphtha hydrotreater/reformer train for
hydrotreating and octane recovery. Splitting the FCC naphtha is an integral part of the CDTech
unit, so it is unclear why the refinery model chose to install an additional splitter in front of the
CDTech unit. Also, the FCC naphtha splitting column simulated by the refinery model is a full
boiling range column. This type of column is more costly than a simpler two cut splitter which
should be sufficient for this application, as we discussed above. Finally, it is also not clear why
the refinery model chose to route some of the FCC naphtha to the reformer hydrotreater. We
identified this technique above as a way to reduce sulfur operationally in the period of time prior
to installation of a FCC gasoline desulfurization unit. However, this technique is generally not
considered to be beneficial in the long run, as running FCC naphtha through the reformer affects
the yield and octane of the reformate. It is not clear how this may have affected the costs
projected by the DOE model, as the effect of running FCC naphtha through the reformer on
reformate yield and octane was not presented. However, along with the inclusion of the full-
range splitter, this could have increased costs beyond that necessary to achieve the sulfur
standard.
The study's estimated high cost of producing 10 ppm gasoline also appears to be
explainable. The refinery model did not include the severe hydrotreating representations of the
improved and low cost, proven technologies. The Mathpro model and our model include these
severe desulfurization representations. We present cost information in in Section c. below for
V-69
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
more greater than 95 percent efficient desulfurization. A comparison of these costs to those
presented above in Tables V-24 and V-25 for 95 percent efficient desulfurization shows that
increasing desulfurization efficiency increases costs, but not to the degree indicated by the results
of DOE's refinery model. Thus, the absence of these more efficient units appears to have had a
major impact on the refinery model's ability to achieve the 10 ppm standard. For example, the
DOE refinery model estimated that the achieving the 10 ppm standard would cost $660 million
in capital costs per refinery. This is more than an order of magnitude higher than the cost of a
FCC gasoline hydrotreater and many times higher than the cost of an FCC feed hydrotreater
coupled with a FCC gasoline hydrotreater. Thus, it appears that the model simply did not include
cost effective means with which to achieve such low sulfur levels.
Regarding challenged refineries, the DOE study shows that it is only slightly more costly
for the challenged refineries to meet the 30 ppm standard than for the mid-capability refineries.
This difference disappears altogether using EPA's lower capital cost amortization factor based on
a 7% ROI. This suggests that DOE's projected higher desulfurization cost for challenged
refineries is due primarily to higher capital costs and operating costs may actually be lower. This
suggests that for ROI's below 10%, the difference in costs for average and challenged refineries
is small. However, since it appears that the cost for average refineries included some
unnecessary costs, the actual cost difference for average and challenged refineries may be larger
than indicated in Table V-42.
c. Cost of Meeting a 5 ppm Averaging Standard
We received comments from the automobile industry that we should finalize our gasoline
sulfur program with a 5 ppm average sulfur standard. We analyzed the cost of meeting that
standard. We contacted CDTech and Mobil Oil and obtained input and process information on
how their processes could be used by refiners to desulfurize their FCC naphtha to 5 ppm. The
CDTech unit which was costed out above to desulfurize the FCC naphtha to below 100 ppm for a
pool average of 30 ppm, can be modified to desulfurize FCC naphtha to 5 ppm. The CDTech
unit normally is comprised of two columns, one is the CDHydro column, and the second is
named CDHDS. To attain very low sulfur FCC naphtha, CDTech informed us that they could
use two of their CDHDS columns to attain FCC desulfurization beyond 99 percent. Similar to
the use of CDTech process for treating gasoline down to 30 ppm, the CDHydro unit is
commercially demonstrated, but the CDHDS unit is not.
Mobil Oil has commercial desulfurization experience with their Octgain 125 process
desulfurizing the FCC naphtha by over 99 percent. However, because of the amount of olefm
desulfurization and octane loss by the Mobil process, if it were used to desulfurize light FCC
naphtha, Mobil Oil recommends that their process be coupled with an less aggressive
hydrotreating process for treating the light FCC naphtha to reduce octane loss. We considered
using an extractive desulfurization unit, or CDTech's CDHydro process. The CDHydro process
V-70
-------
Chapter V: Economic Impact
has two advantages over the extractive desulfurization unit. First, it removes more sulfur out of
the light FCC naphtha pool. Second it is a desulfurization unit coupled with a distillation
column, saving the need for a separate splitting column. Therefore, we coupled CDTech's
CDHydro process with Octgain's 125 process to most cost-effectively desulfurize the FCC
naphtha, both of which are commercially demonstrated.
Finally, other desulfurization technologies which can be used to desulfurize gasoline to 5
ppm is the combination of an FCC feed hydrotreater with a CDTech unit, or any other FCC
gasoline hydrotreater. In this case, the FCC feed hydrotreater is commercially demonstrated, but
the CDTech unit has not yet been demonstrated. This strategy is particularly likely for refineries
which already have an FCC feed hydrotreater.
The process operation information for these processes is summarized in Table V-43. The
processing costs for the CDTech unit presented here are greater than those presented in Table V-
24 above, due to the need to achieve a greater degree of desulfurization.
V-71
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-43. Process Operation Information for Deep Desulfurization of FCC Naphtha
Technology
(sulfur removal efficiency)
Capacity
(MMbbl/day)
Capital Cost
(MM$)
Hydrogen Consumption
(SCF/bbl)
Electricity
(KwH/bbl)
HP Steam
(Lb/bbl)
Fuel Gas
(BTU/bbl)
Catalyst Cost
($/bbl)
Cooling Water
(Gal/bbl)
Yield Loss (%)
Octane Loss (R+M)/2
CDTech
(99.4%)
20,000
25.7
165
0.75
-
81,240
0.23
83
0
2.1
CDHydro
(98%)
8800
4.6
30
0.5
-
55,000
0.02
60
0
0
Octgain 125
(99.9%)
8000
14.5
420
2.3
-
51,000
0.50
45
8.5
0
FCC Feed
Hydrotreater
(93%)
34,500
60
290
1.5
14
56,000
0.04
-
0.9
-
Meeting a 5 ppm specification day-in and day-out would require refiners to ensure that
each and every stream is low in sulfur. Thus, gasoline blendstocks which are sufficiently low in
sulfur for meeting a 30 ppm specification may have to be monitored more closely and the sulfur
level would, perhaps, have to be controlled tighter than what they are now. These streams
include reformate, isomerate, alkylate, hydrocrackate, and even MTBE. Since these streams are
already low in sulfur (10 ppm or lower except for MTBE which can be two to three times that)
not much monitoring or treating is necessary to ensure that these streams remain low in sulfur,
and the cost is expected to be low. We did not provide our own estimates of these costs; instead
we used the costs from the Alliance of Automobile Manufacturer's study by Mathpro on the cost
of meeting a 5 ppm gasoline sulfur specification. These monitoring or sulfur controlling
strategies and their respective costs are summarized below in Table V-44. In sum, accounting for
these refinery processing changes add an additional 0.2 cents per gallon to the cost of producing
gasoline.
V-72
-------
Chapter V: Economic Impact
Table V-44. Other Refinery Process Changes Potentially Needed to Meet a 5 ppm Sulfur
Standard ($1997)
Description
Install extractive desulfurization treating
for captive MTBE
Install extractive desulfurization treating
for light straight run and natural gasoline
Provide additional hydrotreating of
hydrocrackate for recombinant
mercaptans
Add three stage washing facilities for
alkylate production
Apply good refinery practice to control
reformate sulfur to <=1 ppm
Unit Cost
see extractive
desulfurization costs
see extractive
desulfurization costs
$400/bbl/day + extractive
desulfurization oper
costs
$200/bbl/day + extractive
desulfurization oper cost
$500/day
Cost Impact on
Gasoline Pool
0.008
0.03
0.04
0.09
0.02
In addition to the information summarized above, we make additional assumptions with
respect to estimating the cost of producing 5 ppm gasoline. To simplify the analysis, we created
a national average refinery based on the individual PADD-average refineries, by volume
weighting those average refineries. We volume-weighted the utility and other operational costs
for the national average refinery. Like the analysis for the 30 ppm standard, we applied a 15
percent contingency factor to the final estimated cost of meeting the 5 ppm standard. We
adjusted the capital capacity upward by 10 percent to account for the uncertainty of meeting the 5
ppm standard, and this is additional to the 7 percent factor to adjust calendar day throughput to
stream day throughput. We added the tankage allowance like the 30 ppm analysis.
Based on the information summarized above, we estimated the cost of desulfurizing
gasoline to 5 ppm. We included the 0.2 c/gal treatment costs for the other gasoline blendstocks
in our cost calculation. Our cost estimate, however, does not include any additional distribution
costs which may be incurred by distributing a much cleaner product. Our costs of achieving 5
ppm are summarized in Table V-45.
V-73
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-45. Estimated Cost of Meeting a 5 ppm Sulfur Standard
($1997)
Technology
CDTech
CDHydro/Octgain 125
FCC Feed HT/CDTech
Cost (c/gal)
3.1
3.4
3.8
Incremental Cost to 30 ppm
Standard (c/gal)
1.4
1.7
2.1
The American Automobile Alliance funded a study by Mathpro to estimate the cost of
producing 5 ppm sulfur gasoline.25 The study is based on the same two of three desulfurization
technologies which we used in our cost study, which is CDTech by itself or Octgain 125 coupled
with CDHydro. The study estimated a cost of 2.0 and 2.5 c/gal incremental to 30 ppm gasoline,
which Mathpro estimated to be 2.5 c/gal. Thus, the study's total estimated cost of meeting a 5
ppm sulfur standard is 4.5 to 5.0 c/gal.
The Alliance's cost study estimated a higher desulfurization cost than our study which is
explainable by two primary differences. One, Mathpro, applied a very large 1.8 inside battery
limit (ISBL) to outside battery limit (OSBL) capital cost adjustment factor, which is two times
larger than typical. Second, the study amortized the capital costs on a 10 percent ROI.
Amortizing the capital costs at a 7 percent ROI before taxes and using our ISBL to OSBL cost
adjustment factor yields a cost which is essentially the same as ours.
4. Other Effects of This Program
a. Effect of the Cap Standard
In addition to the 30 ppm averaging standard, we are finalizing an 80 ppm per-gallon
standard. This additional standard will help to avoid high sulfur batches of gasoline from
causing reversibility problems with the emission control hardware. The per-gallon standard or
cap on sulfur level provides an additional challenge to refiners by preventing them from
producing moderate or high sulfur batches of gasoline, which could be possible while meeting
the 30 ppm average standard.
There are a number of situations when refineries tend to produce batches of gasoline with
high sulfur levels. The most obvious situation is 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
V-74
-------
Chapter V: Economic Impact
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.
In the Draft RIA for the NPRM, we laid out our premise that the cost of meeting the cap
standard could be estimated by estimating the cost of reducing gasoline sulfur to meet the
average sulfur level which refiners would be producing their gasoline at under the cap. This is
based on a past communication with API on how to estimate the cost of the cap standard.26 Since
the averaging standard is at or below the average sulfur level which we expect refiners to operate
at if only a 80 ppm cap standard applied, we assumed that there would be no new cost accrued by
the cap standard. Upon investigating this further, we believe that situations could occur when a
refiner could produce gasoline above an 80 ppm cap while still meeting a 30 ppm average
standard. For example, if a refiner typically produced 25 ppm sulfur gasoline to meet this
program's sulfur requirement, he could produce gasoline with 400-500 ppm sulfur for 3-4 days
or 200-300 ppm gasoline for 4-6 days and still average 30 ppm for the calendar year. For
example, these periods of producing high sulfur gasoline could occur if a refiner had to perform a
turnaround of his FCC naphtha hydrotreater.
We received a couple of comments from refiners on our approach on not estimating a
separate cost for the cap standard. These refiners said that they would accrue additional costs for
the cap standard, especially during turnarounds, and that EPA should include these costs. The
comments point out that refiners will overbuild on hydrotreating capacity to treat the high sulfur
FCC naphtha which will need to be treated due to turnarounds of the desulfurization equipment.
Based on these comments, we modified the costs described above for producing low
sulfur gasoline to account for those situations when refiners would otherwise produce high sulfur
gasoline. There were four aspects to these modifications. First, we believe that refiners could
store FCC naphtha during a shutdown of the FCC naphtha hydrotreater. Gasoline production
would decrease in the short term, but gasoline meeting applicable commercial and regulatory
specifications could still be produced and the rest of the refinery could remain operative. To
facilitate this, we provided for the installation of a tank that would store 10 days of FCC naphtha
production. This amount of storage should be adequate for most unanticipated turnarounds. We
presumed that half of refiners would need to add such storage, the other half of refineries either
already having such storage available, or have the capability to send the untreated blendstock to a
nearby refinery which had spare capacity for treating this high sulfur blendstock. Second, we
assumed that refineries would design and operate their desulfurization units to normally produce
gasoline with 25 ppm sulfur. This would allow them to blend in some higher sulfur blendstock
directly into their gasoline pool. Third, we assumed that refiners would install 5 percent more
desulfurization capacity than necessary, in order to treat the blendstock which had been stored
during a turnaround. We did this for all refiners, though it is possible that only one refiner out of
a number of geographically grouped refiners might actually need to invest in extra capacity.
V-75
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Finally, we include a 15 percent capital cost contingency factor, and a 10 percent operational cost
contingency factor to account for costs related to the FCC naphtha hydrotreater and other units in
the refinery which may not have been accounted for in the licensor estimates.
We believe that refinery managers will have to place a greater emphasis on the proper
operation of all of their desulfurization units, not just the new FCC gasoline desulfurization unit,
in order to consistently deliver low sulfur gasoline. This improved operations management could
involve enhancements to the computer systems which control the refinery operations, as well as
improved maintenance practices.27 Refiners may be able to recoup some or all of these costs
through improved throughput. However, even if they cannot do so, these costs are expected to be
less than 1 percent of those estimated above for FCC gasoline desulfurization.28 29
Refiners will also likely invest in a gasoline sulfur analyzer.30 The availability of a sulfur
analyzer at the refinery would provide essentially real-time information regarding the sulfur
levels of important streams in the refinery and facilitate operational modifications to prevent
excursions above the sulfur cap. Based on information from a manufacturer of such an analyzer,
the cost for a gasoline sulfur analyzer would be about $50,000, and the installation cost would be
another $5000.31 Compared to the capital and operating cost of desulfurizing gasoline, the cost
for this instrumentation is far below 1 percent of the total cost of this program.
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 the quantity of gasoline demanded would be expected due
to the price increase in gasoline. This would mitigate the increase 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.32 This situation, coupled with ever increasing demand for gasoline,
V-76
-------
Chapter V: Economic Impact
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.
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 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.33 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.34 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
V-77
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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.35 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, the final rule contains a number of provisions which are intended to prevent
refinery closures due to financial hardship. The small refiner provisions are projected to give 16
small refineries which are owned by small businesses until 2008 in order to meet the 30 and 80
ppm standards. Between 2004 and 2008, these refiners have to meet interim standards which are
related to their current sulfur levels. The geographic phase in delays the 30 and 80 ppm
standards until 2007 for 14 refiners located in PADD 4, but will also benefit those refiners
located outside of PADD 4 but who sell a significant amount of gasoline in PADD 4. Finally,
this final rule also includes a hardship provision applicable to up to about 1 percent of U. S.
gasoline production. This provision is intended to benefit refiners who are not able to produce
complying gasoline because of extreme and unusual circumstances outside the refiner's control
that could not have been avoided through the exercise of due diligence. In all three of these
cases, the additional time provided to meet the 30 and 80 ppm standards would allow these
refiners 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 or emerging desulfurization technology. Other
refiners not covered by these provisions may also be able to delay compliance with the 30 and 80
ppm standards until 2006 through the Averaging, Banking and Trading program (ABT). The
ABT program allows a refiner to phase-in the gasoline sulfur program across its refineries to its
best financial advantage, or gain even more leeway through the generation and purchase of sulfur
credits. For the Final Rule, we are providing more flexibility to refiners by opening up the
provisions governing the trading of allotments to allow trading among all refineries to meet the
corporate sulfur standard.
We received several comments that we should do a refinery closure analysis. However,
we feel that these provisions, which are all designed to minimize the impact of the sulfur
standards on refiners, will address the concerns related to the issue of refinery closures. We can
also point to Mathpro's refinery-by-refinery analysis for the Alliance which provides us with
additonal assurance that refineries will not close.36 Mathpro first analyzed the cost of the
gasoline sulfur program on each refinery in PADD 4. Then it compared the cost to the cash
operating margins of these refineries, and concluded that the relative cost is insufficient to cause
refinery closures in PADD 4. After our own review of the work completed by Mathpro we
reached the same conclusion as Mathpro.
V-78
-------
Chapter V: Economic Impact
c. Other Fuel Issues Which May Affect the Cost to Desulfurize Gasoline
We received several comments on our proposed rule that we should consider the impact
of the expected phase-down of MTBE use in gasoline, and the potential reduction of diesel
sulfur, in our cost analysis of desulfurizing gasoline. With respect to an expected phase-down of
MTBE, we expect that the MTBE content in gasoline will be limited, but not phased out, which
will still allow for the blending of a small volume of MTBE into gasoline. Thus some refiners
which may not be using MTBE now may actually have more access to MTBE for blending into
their gasoline, while other refiners which make a lot of RFG or oxyfuels, may have to reduce
their MTBE use. For desulfurizing their gasoline, refiners can choose among a number of
different desulfurization technologies which have varying impacts on gasoline octane levels.
Since refiners can expect MTBE levels to be phased down, we believe refiners' technology
choice for desulfurizing gasoline will include how an MTBE phase-down will affect their
particular situation, and they will choose the gasoline desulfurization technology that will reduce
their costs while meeting both requirements. Thus, despite not knowing what the final
requirements will be of an MTBE phase-down program, we believe that the phase-down of
MTBE will not have a significant impact on the costs of desulfurizing gasoline.
With respect to diesel desulfurization, we heard from a number of refiners that they can
address both gasoline and diesel desulfurization most cost-effectively with separate hydrotreating
units. The alternative to separate "end of the pipe" hydrotreaters, is to put in a FCC feed
hydrotreater which would still require the two additional desulfurization units (although an
existing diesel hydrotreater could suffice as the second unit). However, FCC feed hydrotreaters
incurs high capital costs which is a significant disincentive to their use. Because refiners aim to
minimize their costs, with a bias away from capital costs, we are convinced that treating the
diesel and gasoline blendstocks separately will be the method of choice for the majority of
refiners, which is corroborated by the comments we received from the oil industry, and this
strategy ensures that the costs of the two programs will be separate. Since there are still
overlapping elements to both programs, such as hydrogen supply, the costs of which can be
reduced if refiners can plan to implement both programs together, refiners want to know what the
eventual diesel program will be before building their gasoline desulfurization units. We are
working to accommodate them with a proposed rule on desulfurizing diesel fuel soon after this
final rule. In the diesel desulfurization rule, we will evaluate the impact of both programs on the
refining industry.
V-79
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
5. 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 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 gallon37, 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 timeframe12. Using the projected
long-term distribution of 40 percent LDV and 60 percent LDT in the fleet38, 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
miles13. Applying the average fuel economy factor of 19.0 miles per gallon and the initial cost
for low sulfur fuel of 1.93 eVgal leads us to a per-vehicle estimate of $11.68. This is the
additional cost that the average vehicle owner will incur in the first year of the program due to
the sole use of low sulfur gasoline.
12 In the NPRM, the value of 15.5 mpg was used for all light-duty trucks. For the final rulemaking, we
have instead applied different mpg values to the different weight classes of trucks: LDT1, 18.7 mpg; LDT2, 15.7
mpg; LDT3, 13.2 mpg; LDT4, 12.2 mpg. Using the weighting factors in Table VI-4, the weighted average of these
values remains 15.5 mpg.
13 Calculated from the annual miles traveled per vehicle for each year of a vehicle's life, multiplied by a
distribution of vehicle registrations 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-80
-------
Chapter V: Economic Impact
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 Safety 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:
LFC = [(AVMT); • (SURVIVE); • (C) - (FE)]
Where:
LFC = Lifetime fuel costs in $/vehicle
(AVMT); = Annual vehicle miles travelled in year i of a vehicle's operational life39
(SURVIVE); = Fraction of vehicles still operating after i years of service40
C = Cost of low sulfur gasoline in $/gal
FE = Fuel economy in miles per gallon. 24.2 for LDV, 15.5 average for LDT
i = Vehicle years of operation, counting from 1 to 25
The cost of low sulfur gasoline is a function of the year of refinery production as described in
Section V.B.; the initial cost of 1.93 eVgal applies only in the first year of low sulfur gasoline
production. In subsequent years, costs will decrease as refiners make use of more advanced
technology. As a result of these declining fuel costs over time, we determined that it is
appropriate to calculate total lifetime costs for two separate cases:
1) Near-term, representing a vehicle whose operational life begins at the same time
that low sulfur gasoline standards take effect (i.e., 2004)
2) Long-term, representing a vehicle whose operational life begins six years after
low sulfur gasoline standards take effect (i.e., 2010)
The sixth year for calculating long-term costs of low sulfur gasoline was chosen to be consistent
with the sixth year of vehicle manufacture, when the capital cost amortization period ends.
Details of the calculation of long-term vehicle costs are given in Section V.A.
We used the above equation to calculate lifetime fuel costs separately for LDV, LDT1,
LDT2, LDT3, and LDT4. The results are shown in Table V-46.
V-81
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-46. Undiscounted Per-vehicle Costs of
Low Sulfur Gasoline
LDV
LDT1
LDT2
LDT3
LDT4
Near-term ($)
95.03
168.15
200.27
255.95
276.93
Long-term ($)
89.45
157.78
187.93
240.10
259.78
We then weighted the per-vehicle costs for the individual vehicle categories in Table V-46 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 $164.83 on a near-term basis and
$154.77 on a long-term basis.
An alternative approach to calculating lifetime per-vehicle costs of low sulfur gasoline is
to discount future year costs. This approach leads to "net present value" lifetime fuel costs, and
is a useful means for showing what the average vehicle owner would have to spend in the first
year in order to pay for all future year fuel costs. It also provides a means for comparing the
program's costs to its emission reductions in a cost-effectiveness analysis, as described in Section
VI.
Discounted lifetime fuel costs are calculated in an analogous manner to the undiscounted
values, except that each year of the summation is discounted at the average rate of 7%. The
equation given above can be modified to include this annual discount factor:
LFC = £ [{(AVMT); • (SURVIVE); • (C) - (FE)}/(1.07)M]
Once again, we calculated lifetime fuel costs separately for LDV, LDT1, LDT2, LDT3, and
LDT4. These values are shown in Table V-47.
V-82
-------
Chapter V: Economic Impact
Table V-47. Discounted Per-vehicle Costs of
Low Sulfur Gasoline
LDV
LDT1
LDT2
LDT3
LDT4
Near-term ($)
69.38
119.60
142.45
181.21
196.06
Long-term ($)
65.51
112.65
134.17
170.55
184.53
Once again, we then weighted the per-vehicle costs for the individual vehicle categories in Table
V-47 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 $117.82 on a
near-term basis and $111.01 on a long-term basis.
A summary of all per-vehicle fuel costs described in this section is given in Table V-48
below.
Table V-48. Fleet Average Per-vehicle Costs
Of Low Sulfur Gasoline
First year
Lifetime, undiscounted, near-term
Lifetime, undiscounted, long-term
Lifetime, discounted, near-term
Lifetime, discounted, long-term
Cost per vehicle
($)
11.68
164.83
154.77
117.82
111.01
6. Aggregate Annual Fuel Costs
Aggregate fuel costs are those costs associated with the increased price 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
V-83
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
analyses within this document, along with projected fuel economy estimates (mpg) developed by
Standard & Poor's Data Research International (DRI).41 The resultant aggregate annual fuel
costs are summarized in Table V-49. It is important to note that the capital costs associated with
the proposed sulfur controls have been amortized for this analysis at a seven percent rate of
return before taxes. The actual capital investment would occur up-front, prior to and during the
initial years of the program, as described previously in this chapter.
Table V-49. Summary of the Increased Annualized Social Cost of Gasoline
as a Result of the Tier 2 Gasoline Sulfur Controls
(SMillion)
Calendar
Year
2000
2004
2010
2015
2020
Including Non-Road and
Excluding California14
0
1,618
2,553
2,648
2,153
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 for 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 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
14The aggregate fuel costs used in the economic impact analysis include gasoline consumed by non-road
sources and exclude gasoline consumed in the State of California.
V-84
-------
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 (roughly 1 percent) relative to the
gasoline consumed by vehicles and trucks under 10,000 pounds.
The motorcycle (MC) fuel economy value used is a very rough estimate (45 mpg), but the
value chosen has little impact on this analysis given the relatively low VMT of motorcycles
relative to LDVs and LDTs (<1 percent).
The stratification of EPA VMT projections between the 8500 to 10,000 pound trucks and
the over 10,000 pound trucks was done by using draft MOBILE6 fleet characterization data
which showed that approximately 83 percent of heavy-duty gasoline trucks are in the 8500 to
10,000 pound range with the remaining 17 percent in the > 10,000 pound range.
The projected VMT values within each category (MC, LDV, LOT, HDCX 10,000 pounds,
and HDG> 10,000 pounds) were then divided by the corresponding DRI projected fuel economy
estimates (or the MC fuel economy estimate) 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.
b. Explanation of Results
The aggregate fuel costs used in the economic impact analysis 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.15 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.42 The highway gasoline consumption,
including the non-road contribution and excluding the California contribution, was then
15Based on EPA VMT estimates that California accounts for approximately 11 percent of nationwide
VMT.
V-85
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 include non-road sources
because gasoline used to power these sources will incur the increased per gallon cost, but exclude
California because this rule 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 an increased cost due to this rule.
The aggregate annual fuel costs change as projected per gallon costs and annual fuel
consumption change over time. For information on how the per gallon costs change over time,
see the discussion earlier in this Chapter. As a result of these changes, the aggregate annual fuel
costs 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-86
-------
Chapter V: Economic Impact
Table V-50. Calculation of Gasoline Consumption by Highway Sources
CY
1996
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
Motorcycle
AMD
Gasoline
MCVMT Gasoline MC
ex MCVMT Esti- Gasoline
CA.AL.HI nation mated Consump
Bmiles Bmiles MC nation
(1) (2) mpg Bgal
9
10 12 45 0.26
10 12 45 0.26
11 12 45 0.27
11 13 45 0.28
12 13 45 0.29
12 13 45 0.30
12 14 45 0.31
13 14 45 0.32
13 15 45 0.33
13 15 45 0.34
14 16 45 0.35
14 16 45 0.36
14 16 45 0.37
15 17 45 0.37
15 17 45 0.38
16 18 45 0.39
16 18 45 0.40
16 19 45 0.41
17 19 45 0.42
17 19 45 0.43
17 20 45 0.44
18 20 45 0.45
18 21 45 0.46
19 21 45 0.47
19 22 45 0.48
19 22 45 0.49
20 22 45 0.50
20 23 45 0.51
21 23 45 0.52
21 24 45 0.53
21 24 45 0.54
PassCar
AMD
Gasoline Gasoline PassCar
PassCar PassCar Gasoline
VMT ex VMT S&P DRI Consump
CA.AK.HI nation PassCar nation
Bmiles (1) Bmiles (2) mpg (3) Bgal
1272
1204 1368 21.2 64.51
1186 1348 21.3 63.30
1169 1329 21.4 61.96
1152 1309 21.6 60.63
1135 1290 21.7 59.33
1118 1271 21.9 58.02
1101 1251 22.0 56.75
1084 1232 22.2 55.49
1086 1234 22.3 55.25
1089 1237 22.5 55.00
1091 1240 23.2 53.44
1093 1242 23.4 53.02
1096 1245 23.7 52.61
1098 1248 23.9 52.21
1100 1250 24.1 51.82
1103 1253 24.6 50.94
1105 1256 24.8 50.57
1107 1258 25.1 50.20
1110 1261 25.3 49.85
1112 1264 25.5 49.50
1114 1266 25.5 49.66
1117 1269 25.5 49.76
1119 1272 25.5 49.87
1121 1274 25.5 49.97
1124 1277 25.5 50.08
1126 1280 25.5 50.18
1128 1282 25.5 50.28
1131 1285 25.5 50.39
1133 1288 25.5 50.49
1135 1290 25.5 50.60
1138 1293 25.5 50.70
LDK8500
LOT
Gasoline
AMD LOT VMT S&P DRI LOT <8500
VMT ex nation <10k Gasoline
CA.AL.HI Bmiles Truck Consump
Bmiles (1) (2) mpg (3) nation Bgal
711
961 1092 15.9 68.66
1023 1163 16.0 72.67
1086 1234 16.1 76.44
1148 1304 16.3 80.14
1210 1375 16.4 83.78
1273 1446 16.5 87.65
1335 1517 16.6 91.18
1398 1588 16.8 94.66
1439 1635 16.9 96.64
1480 1681 17.1 98.58
1521 1728 17.3 99.88
1562 1774 17.5 101.66
1603 1821 17.6 103.39
1644 1868 17.8 105.09
1685 1914 17.9 106.75
1726 1961 18.2 107.74
1767 2007 18.4 109.31
1808 2054 18.5 110.85
1849 2101 18.7 112.36
1890 2147 18.9 113.83
1931 2194 19.0 115.46
1972 2240 19.0 117.92
2013 2287 19.0 120.37
2054 2334 19.0 122.82
2095 2380 19.0 125.27
2136 2427 19.0 127.73
2177 2473 19.0 130.18
2218 2520 19.0 132.63
2259 2567 19.0 135.08
2300 2613 19.0 137.53
2341 2660 19.0 139.99
HDG 8500-1 Ok
8500-1 Ok
HDG HDG 8500-10k
AMD HDG VMT VMT S&P DRI Gasoline
VMT ex nation nation <10k Consump
CA.AL.HI Bmiles Bmiles Truck nation
Bmiles (1) (2) (4) mpg (3) Bgal
46
54 61 51 15.9 3.18
55 63 52 16.0 3.27
57 65 54 16.1 3.35
59 67 56 16.3 3.43
61 69 58 16.4 3.50
63 71 59 16.5 3.59
65 74 61 16.6 3.67
67 76 63 16.8 3.74
69 78 65 16.9 3.82
71 80 67 17.1 3.90
73 82 68 17.3 3.95
75 85 70 17.5 4.03
77 87 72 17.6 4.10
79 89 74 17.8 4.17
81 91 76 17.9 4.23
83 94 78 18.2 4.28
85 96 80 18.4 4.34
87 98 82 18.5 4.40
89 101 83 18.7 4.46
91 103 85 18.9 4.52
93 105 87 19.0 4.59
95 107 89 19.0 4.69
97 110 91 19.0 4.79
99 112 93 19.0 4.89
101 114 95 19.0 4.99
103 116 97 19.0 5.09
105 119 99 19.0 5.19
106 121 100 19.0 5.29
108 123 102 19.0 5.38
110 126 104 19.0 5.48
112 128 106 19.0 5.58
HDG>10k
>10k
HDG >10k
VMT S&P Gasoline
nation DRI Consump
Bmiles >10k nation
(4) mpg (3) Bgal
10 7.1 1.46
11 7.1 1.50
11 7.2 1.55
11 7.2 1.59
12 7.2 1.63
12 7.4 1.64
13 7.4 1.68
13 7.5 1.72
13 7.5 1.77
14 7.5 1.81
14 7.5 1.87
14 7.5 1.92
15 7.5 1.96
15 7.6 2.01
16 7.6 2.05
16 7.6 2.10
16 7.6 2.14
17 7.6 2.19
17 7.7 2.23
18 7.7 2.28
18 7.8 2.29
18 7.8 2.34
19 7.8 2.39
19 7.8 2.44
19 7.8 2.49
20 7.8 2.54
20 7.8 2.59
21 7.8 2.64
21 7.8 2.69
21 7.8 2.74
22 7.8 2.79
Totals
EPA Total S&PDRI Hwy
Hwy Gasoline
Gasoline Consump
Consump nation Bgal
nation Bgal (5)
120.94
138.07 132.72
141.00 134.90
143.56 137.07
146.07 139.25
148.54 141.43
151.20 142.44
153.59 144.62
155.93 146.79
157.80 148.97
159.63 151.15
159.48 151.56
160.97 152.47
162.43 153.38
163.85 154.29
165.25 155.20
165.44 157.48
166.77 158.39
168.06 159.30
169.33 160.21
170.56 161.12
172.45 161.66
175.16 162.63
177.88 163.60
180.59 164.57
183.31 165.54
186.02 166.51
188.74 167.48
191.45 168.45
194.17 169.42
196.88 170.39
199.60 171.36
(1) See Chapter III of this Tier 2 Final Rule RIA for a discussion of these VMT projections.
(2) CA =11% of nation; CA,AK,HI= 12% of nation
(3) From S&P DRI World Energy Service U.S. Outlook, April 1998, Table 17 (mpg values include diesel), S&P does not provide mpg estimates for 2021-2030 so the 2020 estimate is assumed for those years
(4) Uses Draft MOBILE6 Fleet Characterization Data for MOBILE6; OMS/T.Jackson, March 1999; uses fleet mix projections where -83% of HDG are 8500-10K and -17% of HDG are >10K
(5) Presented for comparison only. Discrepancy in later years due mainly to OMS's larger LOT VMT share (67% of LD VMTWs S&fy~53% of <10k VMT)
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table V-51. Aggregate
CY
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
EPA Total
Hwy
Gasoline
Consump
nation
Bgal
138.07
141.00
143.56
146.07
148.54
151.20
153.59
155.93
157.80
159.63
159.48
160.97
162.43
163.85
165.25
165.44
166.77
168.06
169.33
170.56
172.45
175.16
177.88
180.59
183.31
186.02
188.74
191.45
194.17
196.88
199.60
Annualized
Increased Social
Total Hwy
Gasoline
Consumption
excluding CA
Bgal (2)
122.89
125.49
127.77
130.00
132.20
134.57
136.69
138.78
140.44
142.07
141.94
143.27
144.56
145.83
147.07
147.24
148.42
149.58
150.70
151.80
153.48
155.90
158.31
160.73
163.14
165.56
167.98
170.39
172.81
175.23
177.64
Non-road
Gasoline
Consumption
excluding CA
Bgal (3)
7.86
8.03
8.18
8.32
8.46
8.61
8.75
8.88
8.99
9.09
9.08
9.17
9.25
9.33
9.41
9.42
9.50
9.57
9.64
9.71
9.82
9.98
10.13
10.29
10.44
10.60
10.75
10.91
11.06
11.21
11.37
Fuel Costs
per Year from 2004 to 2030
Costs for Gasoline
% of Total that
is 30/80 ppm
Sulfur Gasoline
Bgal (4)
0
0
0
0
0.5947
0.6744
0.9253
0.9253
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
1 .0000
Annual Tier2
Cost excluding
CA & including
NonRoad
$B
0
0
0
0
1.618
1.819
2.268
2.302
2.526
2.555
2.553
2.577
2.600
2.623
2.645
2.648
2.670
2.690
2.711
2.161
2.153
2.133
2.166
2.200
2.233
2.266
2.299
2.332
2.365
2.398
2.431
(1) See Chapter V, section B of this Tier 2 Final Rule RIA for a discussion of these estimates.
(2) CA = 11% of total nation; CA.AK.HI = 12% of nation
(3) OMS/T.Sherwood; NonRoad fraction = 6.4%; see memo to Docket A-97-10, 2/19/99
(4) Represents the fraction of total consumption, exluding CA, that is 30 ppm average/80 ppm max sulfur gasoline.
V-88
-------
Chapter V: Economic Impact
C.
Combined Vehicle and Fuel Costs
Sections A. and B. of this section provide detailed cost analyses for Tier 2 vehicles and
low sulfur gasoline, respectively. The following sums the costs to consumers to provide total
incremental costs of the Tier 2 program. The per vehicle costs are provided first, followed by the
total annual nationwide costs.
1.
Combined Costs Per Vehicle
Table V-52 provides a summation of our estimated incremental per vehicle costs,
including increased costs for Tier 2 vehicles and for low sulfur gasoline over the life of the
vehicles. We use the cost estimates for our cost-effectiveness analysis presented in the following
Chapter. As described in the previous sections, we expect these costs to decrease over time as
manufacturers make production improvements and recover fixed costs. Table V-52 provides
estimates of near-term costs, which represent costs in the first years of the program, and long-
term costs which account for the cost decreases.16
Table V-52. Total Incremental Per Vehicle Costs to Consumers
Over the Life of a Tier 2 Vehicle
LDV
($)
LDT1
($)
LDT2
($)
LDT3
($)
LDT4/MDPV
($)
Near-term Costs
Vehicle costs
Fuel costs*
Total
82
69
151
74
120
194
130
143
273
249
181
430
273
196
469
Long-term Costs
Vehicle costs
Fuel costs*
Total
53
66
119
49
113
162
101
134
235
203
171
374
223
185
408
* Discounted lifetime fuel costs
16 Includes estimated costs for OBD II and ORVR requirements for MDPVs.
V-89
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
2.
Combined Total Annual Nationwide Costs
Figure V-2 and Table V-53 summarize EPA's estimates of total annual costs to the nation
both for Tier 2 vehicles and low sulfur gasoline.17 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. 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 $4.1 billion in 2009. Total
annualized costs are projected to remain at about $4 billion through 2018. After 2018,
annualized fuel costs are projected to decrease somewhat due to the use of new technologies
which would enable refiners to produce low sulfur fuel at a lower cost. The gradual rise in costs
long term is due to the effects of projected growth in vehicle sales and fuel consumption.
Total Annualized Costs of Tier 2 Vehicles and Low Sulfur Gasoline
5
i 3
2
jo *•
"o
Q
1
0
2004 2006 2008 2010 2012 2014 2016 2018 2020 2022 2024
Year
Note: Capital costs have been amortized for purposes of this analysis
Figure V-2. Total Annualized Costs of Tier 2 Vehicles and Low Sulfur Gasoline.
17 Excluding vehicles and fuel sold in California.
V-90
-------
Chapter V: Economic Impact
Table V-53. Total Annualized Costs to the Nation for
Tier 2 Vehicles and Low Sulfur Gasoline
(Smillion)
Calendar Year
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
Vehicle Costs ($)
$269
$531
$834
$1,383
$1,556
$1,578
$1,500
$1,432
$1,362
$1,354
$1,351
$1,357
$1,364
$1,371
$1,378
$1,385
$1,392
$1,399
$1,406
$1,413
$1,420
Fuel Costs ($)
$1,618
$1,819
$2,268
$2,302
$2,526
$2,555
$2,553
$2,577
$2,600
$2,623
$2,645
$2,648
$2,670
$2,690
$2,711
$2,161
$2,153
$2,133
$2,166
$2,200
$2,233
Total ($)
$1,887
$2,350
$3,102
$3,685
$4,082
$4,133
$4,053
$4,009
$3,962
$3,977
$3,996
$4,005
$4,034
$4,061
$4,089
$3,546
$3,545
$3,532
$3,572
$3,613
$3,653
V-91
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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. "Development of Light-Duty Emission Inventory Estimates in the Notice of Proposed
Rulemaking for Tier 2 and Sulfur Standards", March 1999, EPA 420-R-99-005.
12. "Light-duty Truck Reference Guide", J.D. Power and Associates, July 1998.
13. "Annual Energy Outlook 1999 with Projections to 2020", Energy Information
Administration, Office of Integrated Analysis and Forecasting, U.S., Department of
V-92
-------
Chapter V: Economic Impact
Energy, DOE/EIA - 0383(99), p. 137, December 1999.
14. Tables of FCC Gasoline Desulfurization Technologies' Utility Needs, Capital Costs and
other Operation Costs compiled by the Technology Workgroup of the National Petroleum
Council's Study of the U.S. Refining Industry Meeting Requirements for Cleaner Fuels
and Refineries, September 1999.
15. Very-Low-Sulfur Diesel Distribution Cost, Engine Manufacturers Association, August
1999.
16. Gary, James H., Handwerk, Glenn E., Petroleum Refining: Technology and Economics,
Marcel Dekker, New York (1994).
17. Perry, Robert H., Chilton, Cecil H., Chemical Engineer's Handbook, McGraw Hill 1973.
18. Presentation by the American Petroleum Institute Economics Committee to EPA and
DOE staff, June 16, 1999.
19. Costs of Meeting 40 ppm Sulfur Content Standard for Gasoline in PADDs 1-3 via Mobil
and CDTech Desulfurization Processes, Study Performed by Mathpro Inc. for The
American Petroleum Institute, February 26, 1999.
20. Refining Economics of 5 ppm Sulfur Standard for Gasoline in PADDs 1-3, Study
Performed by Mathpro Inc. for The Alliance of Automobile Manufacturers by Mathpro
Inc., October 1999.
21. Costs of Alternative Sulfur Content Standards for Gasoline in PADD IV, Study
Performed by Mathpro Inc. for The National Petrochemical and Refiners Association,
December 1998.
22. Likely Effects on Gasoline Supply in PADD 4 of a National Standard for Gasoline Sulfur
Content, Study Prepared by Mathpro Inc. for The Association of International
Automobile Manufacturers, Daimler-Chrysler Corporation, Ford Motor Company,
General Motors Corporation, March 1999.
23. Low Sulfur Impacts on a Mid-Capability Refinery, Work Performed by Oak Ridge
National Laboratory Center for Transportation Analysis for the U.S. Department of
Energy Office of Policy, July 1999.
24. Low Sulfur Gasoline Impacts on Challenged Refinery, Work Performed by Oak Ridge
National Laboratory Center for Transportation Analysis for the U.S. Department of
Energy Office of Policy, August 1999.
V-93
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
25. Refining Economics of 5 ppm Sulfur Standard for Gasoline in PADDs 1-3, Study
Performed by Mathpro Inc. for The Alliance of Automobile Manufacturers by Mathpro
Inc., October 1999.
26. Letter from Tim Hogan of the American Petroleum Institute to Charles Freed, U.S. EPA,
October 10, 1997.
27. Jena, Rabi, Take the PC-Based Approach to Process Control, Fuel Reformulation,
November/December 1995.
28. Sutton, IS., Integrated Management Systems Improve Plant Reliabiliy, Hydrocarbon
Processing, January 1995.
29. King, M. J., Evans, H. N., Assessing your Competitors' Application of CEVI/CIP,
Hydrocarbon Processing, July 1993.
30. Christie, David A., Advanced Controls Improve Reformulated Fuel Yield and Quality,
Fuel Reformulation, July/August 1993.
31. Personal conversation with Debbie Pack, ABB Process Analytics Inc., November 1998.
32. 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.
33. Stetzer, C. Martin, Price Waterhouse L.L.C., Redefining the Context of Refinery
Pacesetter Performance, 1997 NPRA Annual Meeting, March 16-18, 1997.
34. Venner, S. F., 1998.
3 5. Refining Industry Profile Study, 1998.
36. Likely Effects on Gasoline Supply in PADD 4 of a National Standard for Gasoline Sulfur
Content, Study Prepared by Mathpro Inc. for AIAM, March 1999.
37. U.S. DOT/ FHA, Highway Statistics for 1995. Total rural and urban miles traveled
divided by total gallons of gasoline consumed in transportation.
38. 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.
39. "MOBILE6 Fleet Characterization Input Data," Tracie R. Jackson, Report Number
M6.FLT.007.
V-94
-------
Chapter V: Economic Impact
40. 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.
41. Standard & Poor's DRI World Energy Service U.S. Outlook, Table 17, April 1998.
42. Memo from Todd Sherwood to Air Docket A-97-10, Non-Road Gasoline Consumption,
dated February 19,1999.
V-95
-------
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 in to estimate the
dollars per ton of total NOx + NMHC emission reductions after the Tier 2 standards have been
fully implemented. The tons reduced used in this analysis are the same as those used in our air
quality modeling analysis. We have also expanded our cost-effectiveness analysis from that
presented in the NPRM to include another approach, aggregate cost-effectiveness, which
accounts for all costs and emission reductions over a 30 year period beginning in 2004. 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. Our responses to comments submitted to us on the subject of cost-
effectiveness can be found in the Response To Comments document, Issue Number 24.
The emission reductions used to calculate the cost-effectiveness levels reported here are
based on those reductions used for our air quality analysis modeling and benefits analysis. This
was done to maintain consistency in the analyses. As noted in section HI. A, we have updated our
inventory model since the air quality modeling inventories were calculated. Table HI. A.-3
compares the updated Tier 2 model with the air quality analysis modeling and shows that the
emission reductions expected from Tier 2/gasoline sulfur will be substantially greater than the
amounts originally calculated. If the updated numbers were incorporated into our cost-
effectiveness we would expect the results to be improved over those shown in this section.
A. Overview of the Analysis
We have calculated the cost-effectiveness of the exhaust emission/gasoline sulfur
standards and the evaporative emission standards, based on two different approaches. The first
considers the net present value of all costs incurred and emission reductions generated over the
life of an average Tier 2 vehicle. This per-vehicle approach focuses on the cost-effectiveness of
the program from the point of view of the Tier 2 vehicles which will be used to meet the new
requirements, and is the method used in our proposal. However, the per-vehicle approach does
not capture all of the costs or emission reductions from the Tier 2/gasoline sulfur program since
it does not account for the use of low sulfur gasoline in pre-Tier 2 vehicles. Therefore, we have
also calculated an aggregate cost-effectiveness using the net present value of costs and emission
reductions for all in-use vehicles over a 30-year time frame. Both approaches have been used in
previous mobile-source programs, though the per-vehicle approach is more common.
VI-1
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Sections A through C describe how we developed our per-vehicle cost-effectiveness
results. This is followed, in Section D, with the extension of these techniques to the aggregate
cost-effectiveness. All of our results are then presented and discussed in Section E.
The per-vehicle cost-effectiveness analysis conducted for our 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 this approach to cost-
effectiveness follow.
1. Temporal and Geographic Applicability
The per-vehicle approach to our cost-effectiveness calculations 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 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 (CASAC) 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
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.
VI-2
-------
Chapter VI: Cost-Effectiveness
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 rulemaking, therefore, we will present cost-
effectiveness of our 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 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 standards would be calculated with respect to
another set of potential standards which is less stringent than our 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.
An average approach to cost-effectiveness, on the other hand, requires that we compare
the costs and emission reductions associated with our 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 standards.
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 standards, both the emission reductions and the fuel cost as a function of sulfur content are
nearly linear, though the vehicle costs do contain some nonlinearity. In addition, nearly all past
mobile source programs have calculated cost-effectiveness with respect to the previous set of
standards. Therefore, we have chosen to calculate cost-effectiveness on an average rather than an
incremental basis.
Since today's program includes both fuel standards and vehicle standards, it was
VI-3
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 sulfur standard would go into effect. The national average sulfur content of
current conventional gasoline is approximately 300 ppm.. This is a change from the NPRM
value of 330 ppm based on more recent survey data. 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 n reformulated gasoline (RFG), the average sulfur content is projected to be 150 ppm
in the summer and 300 ppm in the winter1. Based on the fact that the high ozone season covers
approximately 4.5 months, we estimate that 38 vol% of the annual pool is summer gasoline, with
the remainder being winter gasoline. Applying these fractions to the Phase II RFG sulfur levels
produces an annual 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 285 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 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 Tier 2 evaporative
emission standards. Instead, the 2.0 gram/test standards under the enhanced evaporative
1 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-4
-------
Chapter VI: Cost-Effectiveness
procedure, initially implemented in 1996, have been used as the baseline.
B. Costs
The costs used in our per-vehicle 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 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 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
amortization has ended and all learning curve cost savings have been accounted for. For the
purposes of this rulemaking, therefore, we will present cost-effectiveness of our 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 Tier 2 program. For instance, the costs associated
with our gasoline sulfur control program will decline in steps due to rotating capital
expenditures. 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 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
VI-5
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
our 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 amortization has 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. 1 and V.B.2.
2.
Vehicle and Fuel Costs
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 standards. For the purposes of calculating cost-effectiveness,
we first subtracted out the costs attributable to compliance with our 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 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 evaporative emission standards
were constant across all vehicle classes at $4.10 per vehicle. For low sulfur gasoline, we used
the discounted lifetime costs presented in Table V-46. The costs used in our cost-effectiveness
calculations are repeated in Table VI-1.
Table VI-1. Fleet-average, Per-vehicle Costs Used in Cost-effectiveness
Near-term
Long-term
Vehicle-exhaust
($)
121.04
89.56
Vehicle-evap
($)
4.10
4.10
Fuel
($)
117.82
111.01
Total costs
($)
242.96
204.67
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 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
VI-6
-------
Chapter VI: Cost-Effectiveness
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 program.
The reduction in gasoline sulfur levels that would result from our 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. We are not setting
Tier 2 standards in order to control emissions of SO2, so we have not calculated the cost-
effectiveness of SO2 control. Likewise for sulfate PM, manufacturers are already meeting the
Tier 2 PM standard, so that there are no additional costs for compliance and PM cost-
effectiveness is not relevant. However, reductions in emissions of SO2 and sulfate PM represent
real benefits of our program, and it is appropriate to account for them in some way in our cost-
effectiveness calculations. To do this, 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 have produced both "credited" and "uncredited"
$/ton NOx+NMHC values; the former takes into account the SO2 and direct PM emission
reductions associated with our standards, while the latter does not.
Cost-effectiveness values for the control of SO2 and direct PM represent conservative
estimates of the cost of measures that will need to be implemented in the future in order for all
areas to reach attainment. Such cost-effectiveness values are therefore an appropriate source for
estimating the amount of the costs to credit to these pollutants. As a result, we credited some
costs to SO2 and direct PM through the application of cost-effectiveness ($/ton) values for these
two pollutants drawn from other sources.
In concept, we would consider the most expensive program needed to reach attainment to
be 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 VIE 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 cost-effectiveness of the eight
SO2 control programs ranged from $1600/ton to $111,500/ton. In this particular rulemaking, 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
VI-7
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
NAAQS revisions rule ranged from $2,400/ton (for PM10) to $12,900/ton (for PM2 5). The
particulate matter that would be reduced as a result of our Tier 2/gasoline sulfur program could
be categorized as fine PM having mean particle diameters of less than 2.5 microns. Despite the
fact that the revised NAAQS for PM was remanded, PM2 5 remains a bigger health hazard than
PM10, and it is therefore still valid to examine cost-effectiveness values for both PM10 and PM2 5.
Furthermore, 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 standards.
The cost crediting was applied after all costs associated with compliance with our
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, $50.61 of the total costs
were apportioned to SO2, while $3.72 was apportioned to direct PM. These amounts are
independent of whether we are considering a near-term or long-term cost-effectiveness
calculation, since the lifetime tons reduced for these two compounds is the same, on a per-
vehicle basis, in any year of the program. A summary of the costs used in our cost-effectiveness
calculations is given below in Table VI-2.
Table VI-2. Fleet Average Per-vehicle Costs Used in Cost-effectiveness
Total uncredited costs
SO2 credit allocation
Direct PM credit allocation
Total credited costs
Near-term costs
($)
242.96
-50.61
-3.72
188.63
Long-term costs
($)
204.67
-50.61
-3.72
150.34
c.
Emission Reductions
In order to determine the overall per-vehicle 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, sulfate PM, and SO2 in a manner
VI-8
-------
Chapter VI: Cost-Effectiveness
analogous to the discounted lifetime fuel costs described in Section V.B.4.
1. NOx and NMHC
Our standards are intended primarily to reduce emissions of NOx and NMHC. As a
result, we have determined that the cost-effectiveness of our standards should be determined for
both NOx and NMHC. It is true the our program does include new standards for PM. However,
these standards are already being met by manufacturers. Thus manufacturers will incur no new
costs to comply with the Tier 2 PM standard and a cost-effectiveness analysis of the PM
standards is therefore unnecessary.
Several past rulemakings which produced reductions in both NOx and NMHC 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 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 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:
1) Annual mileage accumulation levels 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:
VI-9
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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, where LDT4 includes MDPV) 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 and 300 ppm for the winter; conventional gasoline in the remaining areas, modeled
at 300 ppm sulfur year-round). 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-EVI 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 program, it was
necessary for us to combine the values for EVI vs. no-EVI areas and RFG vs. conventional gasoline
areas in an effort to represent the national scope of our program. 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 EVI and RFG areas5. We
also made a distinction between summer and winter RFG, since summer-grade Phase n RFG
having approximately 150 ppm sulfur will be used for only 38 percent of the year, while winter-
grade Phase n RFG having approximately 300 ppm sulfur will be used for the remaining 62
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.
VI-10
-------
Chapter VI: Cost-Effectiveness
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
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 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.
The final step before calculating the cost-effectiveness of our 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.
Table VI-4. Vehicle Class Sales Weighting Factors
LDV
LDT1
LDT2
LDT3
LDT4
0.400
0.102
0.340
0.103
0.055
The values in Table VI-4 differ slightly from those in the draft RIA due to the inclusion of larger
trucks above 8500 Ib GVWR (a class now called medium duty passenger vehicles, or MDPV)
into the LDT4 category. 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
VI-11
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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)
8002
285
NOx
(tons)
0.14331
0.11556
Exhaust
NMHC
(tons)
0.03686
0.03319
Evap
NMHC
(tons)
0.04192
0.04192
Table VI-6. Fleet-average, Per-vehicle Discounted
Lifetime Tons for Tier 2 Standards
Sulfur
(ppm)
80
30
NOx
(tons)
0.03565
0.02744
Exhaust
NMHC
(tons)
0.02369
0.02250
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.
2. Irreversibility
As described in Appendix B, we believe that Tier 1, LEV, and Tier 2 vehicles meeting the
SFTP 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.
2 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-12
-------
Chapter VI: Cost-Effectiveness
The increased emissions that result when an SFTP-compliant 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 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 in the range of 20 to 65 percent, meaning that 20 to 65 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, 20 to 65 percent of the
sensitivity effect is permanent or "irreversible" regardless of the 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.
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.
Under our gasoline sulfur program, the average sulfur level will be 30 ppm and the
maximum allowable level will be 80 ppm after full implementation. Per-vehicle lifetime
emissions at these two sulfur levels from Table VI-6 were used to determine the effect of
irreversibility on Tier 2 vehicles. For simplicity, we have used the midpoint of our estimated
range of irreversibility effects, 42.5 percent. The Tier 2 lifetime tonnage values for NOx and
exhaust NMHC at 30 ppm, which included the effects of irreversibility and which was actually
used in our cost-effectiveness analysis, was calculated from the following equation:
ILE30 = (IE).(LE80-LE30)
VI-13
-"30
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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, 0.425
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
285 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 285 ppm and 800 ppm from Table VI-5 were used to
determine the effect of irreversibility on vehicles meeting NLEV standards. As discussed in
Appendix B, we believe that irreversibility applies to any SFTP-compliant vehicle, including
LDT3 and LDT4 meeting Tier 1 standards under the NLEV program. Thus the calculations
followed the same procedure as that used for Tier 2 vehicles:
ILE285 = (IE) • (LE800 - LE285) + LE285
Where:
ILE285 = Irreversibility-impacted, discounted lifetime emissions of SFTP-complaint
NLEV vehicles at 285 ppm sulfur in tons/vehicle, for each vehicle class
IE = Irreversibility impact, 0.425
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 285 ppm sulfur in
tons/vehicle, for each vehicle class
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.
VI-14
-------
Chapter VI: Cost-Effectiveness
Table VI-7. Fleet-average, Per-vehicle Discounted
Lifetime Tons Used in Cost-effectiveness Analysis
Baseline: NLEV
at 285 ppm
Target: Tier 2 at
30 ppm
NOx (tons)
0.12735
0.03093
Exhaust NMHC
(tons)
0.03475
0.02300
Evap NMHC
(tons)
0.04192
0.04020
Total NOx +
NMHC (tons)
0.20402
0.09413
3. Primary Particulate Matter
Vehicles meeting our 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 program, as described in Section Vn.
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 gasoline sulfur
standard are calculated according to the following equation:
LE = [{(AVMT); • (SURVIVE); - (FE) • (D) • (SUL) • (F) • (MC) • (K)}/(1.07)i-1]
Where:
LE
(AVMT);
(SURVIVE);
FE
D
= Discounted lifetime emissions of primary PM in tons/vehicle
= Annual vehicle miles traveled in year i of a vehicle's operational life
= Fraction of vehicles still operating after i years of service
= Fuel economy by vehicle class (see Section VLB. 4)
= Density of gasoline, 6. 17 Ib/gal
VI-15
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
SUL = Change in gasoline sulfur concentration, 2.55xlO"4 Ib sulfur/lb fuel (255 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 0.00037. 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 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, 2.55xlO"4 Ib sulfur/lb fuel (255 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
VI- 16
-------
Chapter VI: Cost-Effectiveness
fleet-average, per-vehicle discounted lifetime tons of SO2 reduced is 0.01054. 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.
D. Aggregate Cost-Effectiveness
Since the per-vehicle approach to cost-effectiveness considers only Tier 2 vehicles, it
does not reflect the costs and emission reductions from pre-Tier 2 vehicles operating on low
sulfur gasoline. An alternative approach for evaluating the cost-effectiveness of our program
involves calculating the net present value of all nationwide emission reductions and costs for a
30 year period. This timeframe captures both the early period of the program when very few Tier
2 vehicles will be in the fleet, and the later period when essentially all vehicles in the fleet will
meet Tier 2 standards. We have calculated this "aggregate" cost-effectiveness using the net
present value of the annual emission reductions and costs. The calculation of aggregate cost-
effectiveness follows the pattern described above for the per-vehicle analysis:
i-2004
Where:
DNAE = Reduction in discounted, nationwide aggregate emissions in tons
(NE); = Reduction in nationwide emissions in tons for year i of the program
i = Year of the program, counting from 2004 to 2034
and
i-2004
Where:
DNAC = Discounted, nationwide aggregate costs in dollars
(NC); = Nationwide costs in dollars for year i of the program
i = Year of the program, counting from 2004 to 2034
The inputs for annual nationwide emission reductions and costs used to calculate the aggregate
cost-effectiveness are given in Appendix VI-C. Aggregate cost-effectiveness is produced by
dividing DNAC by DNAE. The results are given in Section VIE below.
E. Results
VI-17
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
The results of our cost-effectiveness analysis are given in Tables VI-8 (per-vehicle) and
VI-9 (30 year aggregate). We calculated the per-vehicle cost-effectiveness of our 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 285 ppm, and our Tier 2 standards at
our sulfur standard of 30 ppm. The aggregate values were calculated as described in Section D
above.
Table VI-8. Per-vehicle cost-effectiveness of the Tier 2/gasoline sulfur standards
Near term
Long term
Credited
costs ($)
188.63
150.34
Uncredited
costs ($)
242.96
204.67
Tons
NOx+NMHC
0.10989
0.10989
Credited
$/ton
1717
1368
Uncredited
$/ton
2211
1863
Table VI-9. Aggregate cost-effectiveness of the standards
Discounted
aggregate
vehicle &fuel
costs
$48.1 billion
Discounted aggregate
NMHC + NOx
reduction (tons)
23.5 million
Discounted
aggregate cost-
effectiveness per ton
$2,047
Discounted aggregate
cost-effectiveness per ton
with SO 2 and direct PM
credit"
$1,311
a $13.8 billion credited to SO2 ($4800/ton), $3.5 billion to direct PM ($10,000/ton).
The values in Table VI-8 differ slightly from those in the NPRM for several reasons.
First, the truck category has been expanded to include the larger trucks weighing greater than
8500 Ib GVWR (the medium-duty passenger vehicles), causing a small increase in both the
emission reductions and vehicle costs associated with our Tier 2 standards. Second, the baseline
sulfur level changed from 305 ppm in the NPRM to 285 ppm in this final rule as described in
Section VI.A.2. The reduction in baseline sulfur means that emissions from baseline vehicles
were slightly lower than presented in the NPRM, and thus the emissions benefit of reducing
sulfur to 30 ppm is also slightly lower. Third, there was a change in our approach to
irreversibility, in that we revised our estimate of the irreversibility effect to encompass the range
VI-18
-------
Chapter VI: Cost-Effectiveness
of 20 to 65 percent, as described in Appendix B. Using the midpoint of 42.5 percent resulted in a
small decrease in overall emission reductions resulting from Tier 2 vehicles operating on 30 ppm
fuel. Finally, the costs associated with both fuel desulfurization and vehicle aftertreatment
changed, as described in Sections V.A and V.B.
Because the primary purpose of cost-effectiveness is to compare our program to
alternative programs, we made a comparison between the values in Tables VI-8 and VI-9 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
1,980-2,6902
1,859
1,128-1,778
2,228
By comparing the values from Table VI-8 and VI-9 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 Tier 2/gasoline sulfur standards. Vehicle standards, most similar to
today's proposal, have comparable or higher values than our 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, the values in
2 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-19
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table VI-10 might imply that further reductions in NOx and VOC from heavy-duty engines
could be more cost-effective than the reductions that will be produced from our Tier 2/gasoline
sulfur program. However, we do not believe that to be the case. While we are indeed developing
a proposal for further control from heavy-duty engines, we expect that substantial further
emission reductions will require advanced after-treatment devices. These devices will be more
costly than methods use to meet our past standards, and will have difficulty functioning properly
without changes to diesel fuel. We therefore expect that the cost effectiveness of future heavy-
duty standards is not likely to be significantly less than the cost effectiveness of today's rule.
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 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)3. The Regulatory Impact Analysis (RIA) associated with that rulemaking
contained a listing of potential future emission control programs and their cost-effectiveness.8
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.
We recognize that the cost effectiveness calculated for our 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 Tier
2/gasoline sulfur program would apply nationwide. However, in dealing with the question of
comparing local and national programs, it is also relevant to point out that, because of air
transport, the need for NOx control is a broad regional issue not confined to non-attainment areas
only. To reach attainment, future controls will need to be applied over widespread areas of the
3 This rulemaking was remanded by the D.C. Circuit Court on May 14, 1999. However, the analyses
completed in support of that rulemaking are still relevant, since they were designed to investigate the cost-
effectiveness of a wide variety of potential future emission control strategies.
VI-20
-------
Chapter VI: Cost-Effectiveness
country. In the analyses supporting the recent NOx standards for highway diesel engines,4 we
looked at this question in some detail and concluded that the regions expected to impact ozone
levels in ozone nonattainment areas accounted for over 85% of total NOx emissions from a
national heavy-duty engine control program. Similarly, NOx emissions in attainment areas also
contribute to particulate matter nonattainment problems in downwind areas. Thus, the
distinction between local and national control programs for NOx is less important than it might
appear.
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
source technologies that might be considered.
4 Final Regulatory Impact Analysis: Control of Emissions of Air Pollution from Highway Heavy-Duty
Engines, September 16, 1997
VI-21
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
APPENDIX VI-A : Discounted Lifetime Tonnage Values for Exhaust
Emissions
Stnd Veh clas IM 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
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
NLEV
NLEV
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
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
LDT1
LDT1
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
IM
IM
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
30
80
80
Convent ional
RFC
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
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
.04494
.06646
.06478
.07705
.07503
.09894
.09642
.15696
.15282
.18307
.17836
.23321
.22703
.26188
.25512
.03043
.02963
.03939
.03839
.02183
.02128
.04163
.04060
.02033
.01982
.04101
.04000
.02730
.02661
.05087
.04961
.02970
.02894
.05402
.05268
.01364
.01328
.02237
.02181
.06296
.06132
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
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
.01565
.03540
.03000
.02205
.01878
.03943
.03344
.05429
.04632
.07525
.06396
.06443
.05498
.08646
.07351
.01124
.00957
.01892
.01605
.01839
.01565
.03540
.03000
.01832
.01559
.03535
.02996
.02130
.01813
.04114
.03486
.02152
.01831
.04138
.03506
.01124
.00957
.01892
.01605
.01989
.01694
VI-22
-------
Chapter VI: Cost-Effectiveness
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
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
Tier
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
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
LDT1
LDT1
LDT2
LDT2
LDT3
LDT3
LDT4
LDT4
LDV
LDV
LDT1
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
No
IM
No
IM
No
IM
No
IM
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
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
80
150
150
150
150
150
150
150
150
150
150
150
Convent ional
RFC
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
RFG
RFG
RFG
RFG
RFG
RFG
RFG
RFG
RFG
RFG
RFG
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.08716
.08495
.08783
.08552
.11092
.10809
.15929
.15508
.18545
.18068
.23669
.23042
.26534
.25849
.04183
.04073
.05250
.05116
.02903
.02828
.05338
.05206
.02685
.02617
.05236
.05106
.03626
.03533
.06519
.06358
.03954
.03853
.06935
.06763
.01831
.01783
.02905
.02832
.07523
.10209
.09307
.11650
.15830
.18399
.29048
.32511
.05016
.06201
.03424
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.03669
.03110
.02329
.01984
.04051
.03436
.05585
.04765
.07659
.06510
.06632
.05659
.08807
.07489
.01224
.01043
.01983
.01683
.01989
.01694
.03669
.03110
.01982
.01687
.03663
.03105
.02302
.01960
.04260
.03611
.02326
.01981
.04286
.03633
.01224
.01043
.01983
.01683
.01787
.03192
.02059
.03502
.04960
.06677
.06432
.08390
.01106
.01740
.01787
VI-23
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
Tier
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
NLEV
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
LDT1
LDT2
LDT2
LDT3
LDT3
LDT4
LDT4
LDV
LDV
LDT1
LDT1
LDT2
LDT2
LDT3
LDT3
LDT4
LDT4
LDV
LDV
LDT1
LDT1
LDT2
LDT2
LDT3
LDT3
LDT4
LDT4
LDV
LDV
LDT1
LDT1
LDT2
LDT2
LDT3
LDT3
LDT4
LDT4
LDV
LDV
LDT1
LDT1
LDT2
LDT2
LDT3
LDT3
LDT4
LDT4
LDV
LDV
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
No
IM
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
150
150
150
150
150
150
150
150
150
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
300
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
RFC
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
Convent ional
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.06180
.03157
.06047
.04274
.07544
.04667
.08032
.02170
.03386
.09463
.12597
.10225
.12671
.16546
.19134
.30482
.33951
.06330
.07713
.04253
.07537
.03909
.07357
.05307
.09198
.05802
.09802
.02709
.04157
.09718
.12925
.10502
.13003
.16996
.19639
.30513
.33962
.06501
.07916
.04366
.07729
.04012
.07544
.05446
.09432
.05956
.10052
.02781
.04264
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.03192
.01780
.03187
.02066
.03703
.02089
.03728
.01106
.01740
.01901
.03297
.02147
.03586
.05410
.07062
.06969
.08822
.01182
.01812
.01901
.03297
.01893
.03291
.02197
.03822
.02221
.03848
.01182
.01812
.02232
.03887
.02520
.04227
.06339
.08304
.07907
.10044
.01386
.02134
.02232
.03887
.02222
.03880
.02579
.04507
.02608
.04538
.01386
.02134
VI-24
-------
Chapter VI: Cost-Effectiveness
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
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
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
Convent ional
RFC
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
Convent ional
RFG
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.13343
.12953
.18824
.16619
.11983
.11660
.14896
.14218
.18512
.17755
.22195
.20478
.28329
.27546
.30934
.30272
.08982
.08723
.11664
.10329
.05863
.05683
.11031
.09735
.05357
.05191
.10723
.09464
.07307
.07083
.13467
.11874
.08008
.07763
.14375
.12673
.03766
.03653
.06155
.05435
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.02424
.02061
.04059
.03424
.02668
.02271
.04306
.03688
.07818
.06646
.09850
.08046
.09441
.08026
.11431
.09452
.01517
.01291
.02264
.01903
.02470
.02055
.04041
.03416
.02459
.02045
.04033
.03413
.02848
.02369
.04681
.03979
.02883
.02398
.04714
.04006
.01556
.01293
.02244
.01930
VI-25
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 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
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
Conventional
RFG
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
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-26
-------
Chapter VI: Cost-Effectiveness
APPENDIX VI-C Aggregate Annual Tons and Costs
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
NOx
(tons)
338,231
469,037
748,269
856,471
977,740
1,105,762
1,235,882
1,364,290
1,488,166
1,605,738
1,715,040
1,816,767
1,911,270
1,998,345
2,078,026
2,151,690
2,220,210
2,284,625
2,345,739
2,404,807
2,461,670
2,523,034
2,573,768
2,623,506
2,693,468
2,745,571
2,795,551
2,853,945
2,911,214
2,967,538
3.020.448
VOC
(tons)
85,688
91,310
136,232
143,507
153,281
165,486
178,886
191,563
204,728
217,743
230,828
244,080
256,575
269,066
281,325
293,408
305,470
315,447
325,009
334,331
343,560
352,415
361,364
370,210
384,152
392,438
400,968
411,917
422,174
432,141
441.308
PM10
(tons)
14,127
17,307
22,865
23,427
24,049
24,609
25,131
25,728
26,275
26,836
27,404
27,950
28,504
29,042
29,607
30,144
30,685
31,220
31,762
32,288
32,813
33,339
33,864
34,390
34,944
35,474
36,004
36,540
37,078
37,614
38.146
SOx
(tons)
123,849
147,096
189,462
193,779
198,127
202,374
206,480
210,601
214,688
218,668
222,591
226,458
230,288
234,068
237,813
241,517
245,179
248,825
252,461
256,049
259,638
263,215
266,785
270,347
273,906
277,462
281,016
284,581
288,138
291,695
295.253
Fuel cost
($Million)
1,618
1,819
2,268
2,302
2,526
2,555
2,553
2,577
2,600
2,623
2,645
2,648
2,670
2,690
2,710
2,161
2,153
2,134
2,166
2,200
2,233
2,266
2,299
2,332
2,365
2,398
2,431
2,464
2,497
2,530
2.563
Vehicle
costs
fXMimnn)
269
531
834
1,383
1,556
1,578
1,500
1,432
1,362
1,354
1,351
1,357
1,364
1,371
1,378
1,385
1,392
1,399
1,406
1,413
1,420
1,427
1,434
1,441
1,448
1,456
1,463
1,470
1,478
1,485
1.492
VI-27
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 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
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-28
-------
Chapter VII: Benefit-Cost Analysis
Chapter VII: Benefit-Cost Analysis
This chapter reports EPA's analysis of the economic benefits of the final Tier 2/Gasoline
Sulfur rulemaking reducing air pollution from mobile sources. EPA is required by Executive
Order 12866 to estimate the benefits of major new pollution control regulations. The analysis
presented here attempts to answer three questions: 1) what are the physical effects of changes in
ambient air quality resulting from reductions in NOx and SO2 emissions?; 2) how much are the
changes in air quality worth to U.S. citizens as a whole in monetary terms?; and 3) how do the
benefits compare to the costs? It constitutes one part of EPA's thorough examination of all
aspects of the relative merits of regulatory alternatives.
The BCA that we performed for our final rule 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 ozone
and paniculate matter (PM) that will result from our proposed standards.
3. A benefits analysis to determine the changes in human health and welfare, both in
terms of physical effects and monetary value, that result from the changes in
ambient concentrations of various pollutants.
4. Calculation of the costs of the standards for purposes of comparison to the
monetized benefits.
EPA has used the best available information and tools of analysis to quantify the expected
changes in public health and environment and the economic benefits of the final Tier 2/Gasoline
Sulfur rule, given the constraints on time and resources available for the analysis. We have
attempted to be as clear as possible in presenting our assumptions, sources of data, and sources of
potential uncertainty in the analysis. We urge the reader to particularly pay attention to the fact
that not all the benefits of the rule can be estimated with sufficient reliability to be quantified and
included in monetary terms. The omission of these items from the total of monetary benefits
reflects our inability to measure them. It does not indicate their lack of importance in the
consideration of the benefits of this rulemaking. When it is possible to qualitatively characterize
a benefits category, we provide a discussion, although the benefit is not included in the estimate
of total benefits.
We use the term benefits to refer to any and all positive effects of emissions changes on
vn-i
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
social welfare that we expect to result from the final rule. We use the term environmental costs
(also commonly referred to as "disbenefits") to refer to any and all negative effects of emissions
changes in social welfare that result from the final rule. Where it is possible to quantify benefits
and environmental costs, our measures are those associated with economic surplus in accepted
applications of welfare economics. They measure value by estimating (primarily through
benefits transfer) the willingness of the affected population to pay for changes in environmental
quality and associated health and welfare effects.
This analysis presents estimates of the potential benefits from the Tier 2/Gasoline Sulfur
rule occurring in 2030. The emissions reductions that will result from the Tier 2/Gasoline Sulfur
rule have of course not actually occurred yet. The actual changes in human health and welfare
outcomes to which economic values are ascribed are predictions. These predictions are based on
the best available scientific evidence and judgment, but there is unavoidable uncertainty
associated with each step in the complex process between regulation and specific health and
welfare outcomes. The ways in which we deal with these uncertainties are discussed in Section
C.
Figure VII-1 illustrates the steps necessary to link the Tier 2/Gasoline Sulfur rule with
economic measures of benefits. The first two steps involve the specification and implementation
of the regulation. First, the specific standards for reducing air pollution from mobile sources are
established. Next, the necessary changes in vehicle technology and fuels are determined (see
Chapters IV and V). The changes in pollutant emissions resulting from the hypothesized vehicle
and fuel changes are then calculated for input into an air quality model, along with predictions of
emissions for other industrial sectors in the baseline. Next, the predicted emissions are used as
inputs to air quality models that predict ambient concentrations of pollutants over time and
space. These concentrations depend on climatic conditions and complex chemical interactions.
We have used the best available air quality models to estimate the changes in ambient
concentrations (from baseline levels) that are used as the basis for this benefits analysis.
The predicted changes in ambient air quality then serve as inputs into functions to predict
changes in health and welfare outcomes. We use the term "endpoints" to refer to specific effects
that can be associated with changes in air quality. Table VII-1 lists the human health and welfare
effects identified for ozone, PM, CO, and hazardous air pollutants (HAP). This list includes both
those effects quantified (and/or monetized) in this analysis and those for which we are unable to
provide quantified estimates. All of the effects related to CO and HAPs are not quantified for
this analysis due to a lack of appropriate air quality models for these pollutants. For changes in
risks to human health from ozone and PM, quantified endpoints include changes in mortality and
in a number of pollution-related non-fatal health effects. To estimate these endpoints, EPA
combines changes in ambient air quality levels with clinical and epidemiological evidence about
population health response to pollution exposure. For welfare effects, the endpoints are defined
in terms of levels of physical damage (for materials damage), economic output (agriculture and
forestry),
vn-2
-------
Chapter VII: Benefit-Cost Analysis
Tier 2/Gasoline Sulfur
standards
Evaluate changes in vehicles
and fuels
Estimate Expected
Reductions in Pollutant
Emissions
Model Reductions in
Ambient Concentrations of
Ozone and Other Pollutants
Estimate Expected Changes in
Visibility, Agricultural Yields and |
Other Welfare Effects
Estimate Expected Changes in
Human Health Symptoms and Risk
Estimate Changes in Monetary
Value of Visibility and Other
Welfare Effects
Estimate Monetary Value of
Changes in Human Health
Symptoms and Risk
Figure VII-1
Steps in the Tier 2/Gasoline Sulfur Benefits Analysis
vn-3
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table VII-1. Human Health and Welfare Effects of Pollutants Affected by the Tier 2/Gasoline Sulfur Rule
Pollutant
Primary Quantified and Monetized
Effects
Alternative Quantified and/or Monetized
Effects
Unquantified Effects
Ozone Health
Chronic asthma3
Minor restricted activity days/acute
respiratory symptoms
Hospital admissions - respiratory and
cardiovascular
Emergency room visits for asthma
Premature mortality15
Increased airway responsiveness to stimuli
Inflammation in the lung
Chronic respiratory damage
Premature aging of the lungs
Acute inflammation and respiratory cell damage
Increased susceptibility to respiratory infection
Non-asthma respiratory emergency room visits
Ozone
Welfare
Decreased worker productivity
Decreased yields for commercial
crops
Decreased yields for commercial forests
Decreased yields for fruits and vegetables
Decreased yields for non-commercial crops
Damage to urban ornamental plants
Impacts on recreational demand from damaged
forest aesthetics
Damage to ecosystem functions
PM Health
Premature mortality
Bronchitis - chronic and acute
Hospital admissions - respiratory and
cardiovascular
Emergency room visits for asthma
Lower and upper respiratory illness
Shortness of breath
Minor restricted activity days/acute
respiratory symptoms
Work loss days
Infant mortality
Low birth weight
Changes in pulmonary function
Chronic respiratory diseases other than chronic
bronchitis
Morphological changes
Altered host defense mechanisms
Cancer
Non-asthma respiratory emergency room visits
PM Welfare
Visibility in California, Southwestern,
and Southeastern Class I areas
Visibility in Northeastern, Northwestern,
and Midwestern Class I areas
Nitrogen and
Costs of nitrogen controls to reduce
Impacts of acidic sulfate and nitrate deposition on
-------
Chapter VII: Benefit-Cost Analysis
Pollutant
Primary Quantified and Monetized
Effects
Alternative Quantified and/or Monetized
Effects
Unquantified Effects
CO Health
Premature mortality15
Behavioral effects
Hospital admissions - respiratory, cardiovascular,
and other
Other cardiovascular effects
Developmental effects
Decreased time to onset of angina
Non-asthma respiratory ER visits
HAPS Health
Cancer (benzene, 1,3-butadiene, formaldehyde,
acetaldehyde)
Anemia (benzene)
Disruption of production of blood components (
benzene)
Reduction in the number of blood platelets
(benzene)
Excessive bone marrow formation (benzene)
Depression of lymphocyte counts (benzene)
Reproductive and developmental effects
(1,3-butadiene)
Irritation of eyes and mucus membranes
(formaldehyde)
Respiratory irritation (formaldehyde)
Asthma attacks in asthmatics (formaldehyde)
Asthma-like symptoms in non-asthmatics
(formaldehyde)
Irritation of the eyes, skin, and respiratory tract
(acetaldehyde)
HAPS
Welfare
Direct toxic effects to animals
Bioaccumlation in the food chain
a While no causal mechanism has been identified linking new incidences of chronic asthma to ozone exposure, a recent epidemiological study shows a statistical association between long-term exposure
to ozone and incidences of chronic asthma in some non-smoking men, but not in women.
b Premature mortality associated with ozone is not separately included in this analysis. It is assumed that the Pope, et al. C-R function for premature mortality captures both PM mortality benefits and
any mortality benefits associated with other air pollutants.
vn-s
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
light transmission (for visibility), and increases in terrestrial and estuarine nutrient loading (for
ecological effects).
As with emissions and air quality estimates, EPA's estimates of the effect of ambient
pollution levels on all of these endpoints represent the best science available to the Agency. The
majority of the analytical assumptions used to develop our estimates have been reviewed and
approved by the EPA Science Advisory Board. However, like all estimates, they also contain
unavoidable uncertainty, as does any prediction of the future. In Section C and in the subsections
on health and welfare endpoints, this uncertainty is discussed and characterized.
This chapter proceeds as follows: Sections A and B summarize emissions and air quality
results and discuss the way that emissions and air quality changes are used as inputs to the
benefits analysis. Section C introduces the kinds of benefits that are estimated, presents the
techniques that are used, and provides a discussion of how we incorporate uncertainty into our
analysis. In Section D, we describe individual health effects and report the results of the analysis
for human health effects. In Section E, we describe individual welfare effects and report the
results of the analysis for welfare effects. Section F reports our estimates of total monetized
benefits and alternative calculations. Finally, Section G presents a comparison of monetized
benefits and costs.
A. Emissions
In order to determine the air quality impact of the Tier 2 program, 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 nationwide1 inventories for NOx,
NMHC, SO2, and PM. This Section describes how these inventory impacts were determined.
At proposal, we evaluated the impact of the Tier 2 program using a 1990 emissions
inventory from the CAA Section 812 study (Ref), and considered the effect of full-fleet turnover
that was expected to occur well into the future on populations estimated for 2010. This approach
to the analysis was necessary because at the time of proposal, we had no available baseline data
set beyond the year 2010, since the Section 812 inventory was developed only for this year. The
analysis at proposal, 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 the final rule's analysis, we
have enhanced the analysis significantly. We updated the emissions inventory to reflect new
CAA programs and changes in inventories through the year 1996. We then evaluated the impact
1 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.
vn-6
-------
Chapter VII: Benefit-Cost Analysis
of the program in 2030 and on 2030 populations.
The inventories developed for our air quality assessment and for the benefit-cost analysis
have already been presented and discussed in Chapter HI and in the supporting documents
referenced in that chapter. Interested readers desiring more information about the inventory
methodologies or results should consult that chapter for details.
The Tier 2/gasoline sulfur program 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. Because
of this inconsistency and our desire to more appropriately match the costs and emission
reductions of our proposed program, our analysis uses a future year when the fleet is nearly fully
turned over. For today's rule this stability does not occur until well into the future. For the
purpose of the benefit calculations, we assume that 2030 is a representative year to consider in
comparison with the costs.
The resulting analysis represents a snapshot of benefits and costs in a future year in which
the light-duty fleet consists almost 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 because of growth in
vehicle miles traveled and population.) 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).
B. Air Quality Impacts
In Chapter HI, we described the Tier 2 program's impact on air quality in 2007. The 2007
analysis shows the initial impact of the rule on area that must attain the NAAQS by 2007. Using
this information the Agency provides its justification for the need for the rule. For purposes of
the benefit-cost analysis, EPA prepared a second air quality analysis to evaluate the impact of the
rule after it is fully implemented (i.e., when all on-highway vehicles are expected to be compliant
with the new Tier 2 controls). We chose 2030 as this analytical year reflecting full-fleet turnover
and, thus, all costs are realized as well as most benefits.2
This section summarizes the methods for and results of estimating air quality for the 2030
2 We recognize that program costs and benefits will continue to accrue as new vehicles are purchased.
vn-7
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
base case and Tier 2 control scenario. EPA has focused on the air quality changes that have been
linked to health, welfare, and ecological effects. These air quality changes include the
following:
Ambient ozone-as estimated using a regional-scale version of the Urban Airshed
Model (UAM-V),
• Ambient particulate matter (PM10 and PM25)-as projected from a Source-Receptor
Matrix (S-R Matrix) based on the Climatological Regional Dispersion Model
(CRDM),
• Airborne nitrogen deposition-as predicted using local and regional coefficients of
nitrogen deposition for selected estuaries from the Regional Acid Deposition
Model (RADM) in combination with modeled reduction in NOx emissions, and
• Visibility degradation (i.e., regional haze), as developed using empirical estimates
of light extinction coefficients and efficiencies in combination with modeled
reductions in pollutant concentrations.
The air quality estimates in this section are based on the emission changes discussed in
Section A. Using the methods identified and described in Section C, the air quality impacts
listed above are then associated with human populations and ecosystems to estimate changes in
health and welfare effects.
The air quality analysis used in the benefits estimation at proposal was based on results
from the UAM-V and S-R Matrix models to estimate 1990 baseline and 2010 base case air
quality for ozone and paniculate matter, respectively. We then applied the Tier 2 control
scenario to the 2010 estimates to derive the associated air quality changes. For the final rule's
analysis, we updated all aspects of the analysis by estimating 1996 baseline (rather than 1990)
and 2030 base case air quality (rather than using 2010 as a surrogate for full-implementation of
the program). We applied the same Tier 2 control scenario (i.e., same level of stringency and
control) as was used at proposal and evaluated the impact on 2030 air quality (using 2030 VMT
and population projections). These updates to the analysis have augmented the preliminary
benefit-cost analysis provided at proposal.
Section VII.B.l describes the estimation of ozone air quality using UAM-V, while
Section VII.B.2 covers the estimation of PM air quality using the CRDM S-R Matrix. Section
VII.B.3 discusses the estimation of nitrogen deposition. Lastly, Section VII.B.4 covers the
estimation of visibility degradation.
vn-s
-------
Chapter VII: Benefit-Cost Analysis
1. Ozone Air Quality Estimates
We use the previously described emissions inputs with a regional-scale version of UAM-
V to estimate ozone air quality. UAM-V is an "eulerian" three-dimensional grid photochemical
air quality model designed to calculate the concentrations of both inert and chemically reactive
pollutants by simulating the physical and chemical processes in the atmosphere that affect ozone
formation. 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 impacts of the Tier 2 rule on U.S.
ozone concentrations.3 Our analysis applies the modeling system for a base-year of 1996 and for
two future-year scenarios: a 2030 base case and a 2030 Tier 2 control scenario. As discussed
later, we use the two separate years because ambient air quality observations from 1996 are used
to calibrate the model. These results are used solely in the benefits analysis and are not used as
part of the justification for the rule. A 2007-based analysis described in Chapter HI is used for
that purpose.
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 continental United States into two regions: East and West.
It then segments the area in each region into square blocks called grids (roughly equal in size to
counties), 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 calibrate the results of this process to develop 2030 ozone profiles at
monitor sites by normalizing the observations to the actual 1996 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 to the health and welfare concentration-response (C-R) functions of the
benefits analysis. The following sections provide a more detailed discussion of each of the steps
in this evaluation and a summary of the results.
a. Modeling Domain
The modeling domain representing the eastern U.S. is the same as that used in EPA's
"Regulatory Impact Analysis for the NOX SIP Call, FIP, and Section 126 Petitions" (EPA,
1998b). As shown in Figure VII-2, this domain encompasses most of the eastern U.S. from the
east-coast to mid-Texas and consists of two grids with differing resolutions. The shaded area of
Figure VII-2 uses a relatively fine grid of 12 km consisting of seven vertical layers. The
unshaded area of Figure VII-2 has less resolution, as it uses a 36 km grid consisting of five
3Douglas and Iwamiya (1999) provide further information on the UAM-V modeling used in this analysis.
vn-9
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
vertical layers. The vertical height of the modeling domain is 4,000 meters above ground level,
for both the shaded and unshaded regions.
The modeling domain used to obtain results for the western U.S. comprises the entire
contiguous 48 states. Even though the modeling domain covers the entire United States, the
modeling results are only used for benefit analysis of western U.S. locations (i.e., within the
region not shown in Figure VII-2). 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 approximately 56 km (or 2/3 longitude by !/2 latitude) resulting in
a 90 by 56 grid for each vertical layer, with eight vertical layers in all. The smaller 36 km and 12
km resolution for the eastern U.S. runs better capture the photochemical processes for that
region.
b. Simulation Periods
A simulation period, or episode, consists of meteorological data characterized over a
block of days that are used as inputs to the air quality model. A simulation period is selected to
characterize a variety of ozone conditions including some days with high ozone concentrations in
one or more portions of the U.S. and observed exceedances of the 1-hour NAAQS for ozone
being recorded at monitors. This study used four multi-day simulation periods to prepare the
future-year ozone profiles. For the eastern U.S. ozone analysis, we modeled two simulation
periods: July 12-24 and July 5-15, 1995. For the western U.S. analysis, the simulation periods
were July 5-15 and July 18-31, 1996. These episodes include a 2-3 day "ramp-up" period to
initialize the model, but the results for these days are not used in this analysis.
VII-10
-------
Chapter VII: Benefit-Cost Analysis
Figure VII-2. UAM-V Modeling Domain for Eastern U.S.
VII-11
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
c. Converting UAM-V Outputs to Full-Season Profiles for Benefits Analysis
This study extracted hourly, surface-layer ozone concentrations for each grid-cell from
the standard UAM-V output file containing hourly average ozone values. These model
predictions are used in conjunction with the observed concentrations as obtained from the
Aerometric Information Retrieval System (AIRS) to generate ozone concentrations for the entire
ozone season.4'5 The predicted changes in ozone concentrations from the 2030 basecase to 2030
policy scenario serve as inputs to the health and welfare concentration-response (C-R) functions
of the benefits analysis, i.e., the Criteria Air Pollutant Modeling System (CAPMS). In order to
estimate ozone-related health and welfare effects for the entire United States, full-season ozone
data is required for every CAPMS grid-cell. Given available ozone monitoring data, we
generated full-season ozone profiles for each location in the contiguous 48 states in two steps: (1)
we combine monitored observations and modeled ozone predictions to interpolate hourly ozone
concentrations to a grid of eight km by eight km population grid-cells, and (2) we converted
these full-season hourly ozone profiles to an ozone measure of interest, such as the daily average
6'7 For the analysis of ozone impacts on agriculture, we use a similar approach except air quality
is interpolated to county centroids as opposed to population grid-cells. Each approach is fully
detailed in Abt Associates (1999).
d. Ozone Air Quality Results
Table VII-2 provides a summary of the predicted ambient ozone concentrations from the
UAM-V model for the 2030 base case and changes associated with Tier 2 control scenario. As
shown, the mean seasonal average ozone concentrations across all U.S. population grid-cells
declines by almost 2 percent, or 0.6 ppb. A similar relative decline is predicted for the
population-weighted average, which indicates rather uniform reductions in these concentrations
across urban and rural areas. The impact of Tier 2 on seasonal SUM06 ozone metric are
significantly greater with the average across all U.S. counties declining by almost 26 percent, or
4 The ozone season for this analysis is defined as the 5-month period from May to September; however, to
estimate certain crop yield benefits the modeling results were extended to include months outside the 5-month ozone
season.
5Based on AIRS, there were 949 ozone monitors with sufficient data, i.e., at least 9 hourly observations per
day (8 am to 8 pm) in a given season.
6The 8 km grid squares contain the population data used in the health benefits analysis model, CAPMS.
See Section C of this chapter for a discussion of this model.
7This approach is a generalization of planar interpolation that is technically referred to as enhanced
Voronoi Neighbor Averaging (VNA) spatial interpolation (See Abt Associates (1999) for a more detailed
description).
VII-12
-------
Chapter VII: Benefit-Cost Analysis
2.2 ppb. Alternatively, although the absolute change predicted for the population-weighted is
similar to that for the simple average (i.e., 2.4 ppb versus 2.2 ppb), the relative change is less at
12.4 percent because of higher observed baseline values for this ozone metric.
VII-13
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table VII-2. Summary of UAM-V Derived Hourly Ozone Air Quality for 2030 Base Case
and Change Due to Tier 2 Standards
Statistic "
2030 Base Case
Change b
Percent Change b
Seasonal Average (ppb)
Minimum c
Maximum °
Average
Median
Population- Weighted Average d
11.98
77.23
29.85
29.45
29.80
-0.51
-0.16
-0.59
-0.50
-0.47
-4.26%
-0.20%
1.98%
-1.69%
1.57%
Seasonal SUM06 (ppb)
Minimum °
Maximum °
Average
Median
Population- Weighted Average d
0.00
118.58
8.31
4.97
19.60
0.00
-9.72
-2.20
-2.33
-2.44
0.00%
-8.20%
-26.48%
-46.88%
-12.43%
a The seasonal average and SUM06 are calculated at the CAPMS gridcell level and at the county level, respectively.
There are at two levels since health effects estimates are generated at each CAPMS gridcell, and agricultural benefits
(which require SUM06) are generated at the county level. Both ozone measures are based on the results of enhanced
spatial interpolation. The seasonal average is the average over all hours in May through September. SUM06 is defined
as the cumulative sum of hourly ozone concentrations over 0.06 ppb that occur during daylight hours (from Sam to
8pm) in the months of May through September.
b The change is defined as the control case value minus the base case value. The percent change is the "Change"
divided by the "2030 Base Case."
0 The base case minimum (maximum) is the value for the CAPMS gridcell with the lowest (highest) seasonal average.
d Calculated by summing the product of the projected 2030 CAPMS gridcell population and the estimated 2030
CAPMS gridcell seasonal ozone concentration, and then dividing by the total population. The SUM06 estimates are
calculated analogously at the county level.
VII-14
-------
Chapter VII: Benefit-Cost Analysis
2. PM Air Quality Estimates
EPA used the previously described emissions inputs with a national-scale S-R Matrix
based on CRDM to evaluate the effects of the Tier 2 rule on ambient concentrations of both PM10
and PM25. Ambient concentrations of PM are composed of directly emitted particles and of
secondary aerosols of sulfate, nitrate, ammonium, and organics. Relative to more sophisticated
and resource-intensive three-dimensional modeling approaches, the CRDM and its associated S-
R Matrix do not fully account for all the complex chemical interactions that take place in the
atmosphere in the secondary formation of PM. Instead it relies on more simplistic species
dispersion-transport mechanisms supplemented with chemical conversion at the receptor
location.
The S-R Matrix consists of fixed-coefficients that reflect the relationship between annual
average PM concentration values at a single receptor in each county (i.e., 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 as well as receptors in 10 Canadian provinces and 29 Mexican cities/states. The
methodology used in this RIA for estimating PM air quality concentrations is detailed in Pechan-
Avanti (1999) and is similar to the method used in the July 1997 PM and Ozone NAAQS RIA
(U.S. EPA, 1997e) and the RIA for the final Regional Haze Rule (U.S. EPA, 1999). The
following sections summarize the steps taken to apply the S-R Matrix for this analysis and to
derive the resulting changes in PM air quality.
a. Development of the S-R Matrix
The S-R Matrix was developed using the CRDM, which uses assumptions similar to the
Industrial Source Complex Short Term model (ISCST3), an EPA-recommended short range
Gaussian dispersion model. The 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. Meteorological data for 1990 coupled with
emissions data from version 2.0 of the 1990 National Particulate Inventory (NPI) were used with
CRDM to develop the S-R Matrix.
The NPI was separated into 5,944 sources (i.e., industrial point, utility, area, nonroad, and
motor vehicle) of primary and precursor emissions. Each individual unit in the inventory was
associated with one of four modeled source types (i.e., area, point sources with effective stack
height of 0 to 250m or 250m to 500m, and individual point sources with effective stack height
above 500m) for each county. Emissions that were modeled include SO2, NOX, and ammonia,
which are needed to calculate ammonium sulfate and ammonium nitrate concentrations; VOC,
which are needed to calculate secondary organic aerosols; and directly emitted PM10 and PM2 5.
VII-15
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Both anthropogenic and natural sources of each species were included.
The resulting transfer coefficients are adjusted to account for the chemical balance
between sulfate, nitric acid, and ammonium (Latimer, 1996). The coefficients for SO2, NOX, and
ammonia were multiplied by the ratios of the molecular weights of sulfate/SO2, nitrate/nitrogen
dioxide and ammonium/ammonia to obtain concentrations of sulfate, nitrate and ammonium.8 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. So, ammonium nitrate forms under conditions of excess
ammonium, and only under relatively low temperatures. Accordingly, for each county receptor,
the sulfate-nitrate-ammonium equilibrium is estimated based on the following simplifying
assumptions:
1. All sulfate is neutralized by ammonium;
2. Ammonium nitrate forms only when there is excess ammonium;
3. Average annual particle nitrate concentrations are divided by four assuming that
sufficiently low temperatures are present only one-quarter of the year.
The total particle mass of ammonium sulfate and ammonium nitrate is calculated by multiplying
the anion concentrations of sulfate and nitrate by 1.375 and 1.290 respectively.
b. Fugitive Dust Adjustment Factor
As demonstrated in the RIA for the PM and Ozone NAAQS (U.S. EPA, 1997e), 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; however, monitoring data from the IMPROVE network show that minerals
(i.e., crustal material) comprise only about five percent of PM25 mass in the East and roughly 15
percent of PM25 mass in the West (U.S. EPA, 1996a). 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 NPI significantly overestimates fugitive dust
emissions. A comparison with a more recent National Emissions Trends inventory indicates that
the NPI overestimates fugitive dust PM10 and PM2 5 emissions by 40 percent and 73 percent
respectively9 (U.S. EPA, 1997c).
8 Ratio of molecular weights: Sulfate/SO2= 1.50; nitrate/nitrogen dioxide = 1.35; ammonium/ammonia =
1.06.
9 Natural and man-made fugitive dust emissions account for 86 percent of PM10 emissions and 59 percent
of PM2 5 emissions in the 1997 version of the National Emission Trends Inventory.
VII-16
-------
Chapter VII: Benefit-Cost Analysis
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 PM10 and PM25 and actual ambient data. This adjustment results in a fugitive dust
contribution to modeled ambient PM2 5 concentrations of 10 percent to 17 percent.10 Even after
this adjustment the fugitive dust fraction of total eastern PM2 5 mass is 10.4 percent, which is still
greater than the five percent indicated by IMPROVE monitors. However, given that the
adjustment factor brings the modeled fugitive dust contribution to PM2 5 mass more 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 in some locations.
c. Normalizing S-R Matrix Results to Observed Data
In an attempt to further ensure comparability between S-R Matrix results and measured
annual average PM values, we also calibrated these results to observed monitoring data using
factors developed for the PM and Ozone NAAQS RIA (U.S. EPA, 1997e). For the NAAQS
RIA, a "calibration factor" was developed for each monitored county based on monitoring data
from 1993 to 1995 for PM10 from the AIRS database.11 This calibration procedure was applied to
all S-R Matrix predictions, regardless of overprediction or underprediction relative to monitored
values, and equally across all particle species contributing to the annual average PM value at a
county-level receptor. 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
forPM10.12
10 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 PM25 mass for: Central U.S. = 44.6 percent; Eastern U.S. = 30.9 percent;
Western U.S. = 31.5 percent.
11 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.
Tier 4 covers the remaining 2369 non-monitored counties.
12 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-17
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Because there is little PM2 5 monitoring data available, we developed a general linear
model to predict PM25 concentrations directly from the monitored 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
PM2 5 data to calibrate the S-R Matrix model predictions of annual average PM2 5.
d. PM Air Quality Results
Table VII-3 provides a summary of the predicted ambient PM10 and PM2 5 concentrations
from the S-R Matrix for the 2030 base case and changes associated with Tier 2 control scenario.
As shown, the average annual mean concentrations of PM10across all U.S. counties declines by
almost 1 percent, or 0.22 g/m3. The same relative decline is predicted for the population-
weighted average for mean PM10, which indicates rather uniform reductions in these
concentrations across urban and rural areas. The impact of Tier 2 on PM2 5 concentrations are
slightly greater with average annual mean concentrations of PM25 across all U.S. counties
declining by almost 2 percent, or 0.22 g/m3. Similar to PM10 concentrations, the relative
change predicted for the population-weighted average does not differ much from the spatial
average.
VII-18
-------
Chapter VII: Benefit-Cost Analysis
Table VII-3. Summary of 2030 Base Case PM Air Quality and Changes Due to Tier 2
Standards
Statistic
2030 Base Case
Change"
Percent Change
PM10
Minimum Annual Mean PM10 ( g/m3) b
Maximum Annual Mean PM10 ( g/m3) b
Average Annual Mean PM10 ( g/m3)
Median Annual Mean PM10 ( g/m3)
Population- Weighted Average Annual Mean PM10 ( g/m3) °
6.64
145.11
24.89
23.90
36.21
-0.03
-0.09
-0.22
-0.20
-0.31
-0.5%
-0.1%
-0.9%
-0.6%
-0.9%
PM25
Minimum Annual Mean PM2 5 ( g/m3) b
Maximum Annual Mean PM2 5 ( g/m3) b
Average Annual Mean PM2 5 ( g/m3)
Median Annual Mean PM10 ( g/m3)
Population- Weighted Average Annual Mean PM2 5 ( g/m3) °
0.86
88.47
11.93
11.96
15.52
0.00
-0.08
-0.22
-0.20
-0.31
0.0%
-0.1%
-1.8%
-1.7%
-2.0%
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 is the observed change for the county with the lowest (highest) annual average in the base case.
0 Calculated by summing the product of the projected 2030 county population and the estimated 2030 county PM
concentration, and then dividing by the total population in the 48 contiguous states.
Table VII-4 provides additional insights on the changes in PM air quality resulting from
the motor vehicle Tier 2 and fuel standards. This table focuses on the absolute change (in terms
of g/m3) and relative change (in terms of percent) observed across individual U.S. counties. As
shown, the absolute reduction in annual mean PM10 concentration ranged from a low of 0.01
g/m3 to and high of 1.25 g/m3, while the relative reduction ranged from a low of 0.04 percent
to a high of 3.3 percent. Alternatively, for mean PM25, the absolute reduction ranged from zero
to 1.23 g/m3, while the relative reduction ranged from zero to 5.4 percent.
VII-19
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table VII-4. Summary of Absolute and Relative Changes in PM Air Quality Due to Tier 2
Standards
Statistic
Absolute Change from
2030 Base Case
(g/m3)"
Relative Change from
2030 Base Case
(%)b
PM10
Minimum
Maximum
Average
Median
Population- Weighted Average °
-0.01
-1.25
-0.22
-0.20
-0.31
-0.04%
-3.30%
-0.87%
-0.94%
-0.91%
PM25
Minimum
Maximum
Average
Median
Population- Weighted Average °
0.00
-1.23
-0.22
-0.20
-0.31
0.00%
-5.42%
-1.77%
-1.88%
-1.95%
a The absolute change is defined as the control case value minus the base case value for each county.
b The relative change is defined as the absolute change divided by the base case value, or the percentage change, for
each county. The information reported in this column does not necessarily reflect the same county as is portrayed in the
absolute change column.
0 Calculated by summing the product of the projected 2030 county population and the estimated 2030 county PM
absolute/relative measure of change, and then dividing by the total population in the 48 contiguous states.
3. Visibility Degradation Estimates
Visibility degradation is often directly proportional to decreases in light transmittal in the
atmosphere. Scattering and absorption by both gases and particles decrease light transmittance.
To quantify changes in visibility, our analysis computes a light-extinction coefficient, based on
the work of Sisler (1996), which shows the total fraction of light that is decreased per unit
distance. This coefficient accounts for the scattering and absorption of light by both particles and
gases, and accounts for the higher extinction efficiency of fine particles compared to coarse
particles. Fine particles with significant light-extinction efficiencies include sulfates, nitrates,
organic carbon, elemental carbon (soot), and soil (Sisler, 1996).
Based upon the light-extinction coefficient, we also calculated a unitless visibility index,
VII-20
-------
Chapter VII: Benefit-Cost Analysis
called a "deciview," which is 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.
Because the visibility benefits analysis (see Section VII.C) distinguishes between general
regional visibility degradation and that particular to Federally-designated Class I areas (i.e.,
national parks, forests, recreation areas, wilderness areas, etc.), we separated estimates of
visibility degradation into "residential" and "recreational" categories. The estimates of visibility
degradation for the "recreational" category apply to Federally-designated Class I areas, while
estimates for the "residential" category apply to non-Class I areas. Deciview estimates are
developed from the estimated county-level changes in particulate matter generated from results
of the S-R Matrix for the 2030 base case and Tier 2 control scenarios. These deciview estimates
are then aggregated to one of eight regions in the case of the residential category (as defined by
the underlying study) and one of six regions in the case of the recreational category (as defined
by Class I Visibility Regions described in Section VII.C). More detail on this approach and
results may be found in Pechan-Avanti (1999).
Table VII-5 provides a summary of the visibility degradation estimates in terms of
deciviews by residential category across U.S. regions. As shown, the national improvement in
residential visibility is 1 percent, or 0.23 deciviews. Predicted visibility improvements are the
largest for the North Central and Northwest (both at 1.3 percent), the South Central (1.2 percent),
and the Southeast (1. Ipercent). Smaller visibility improvements are predicted in the Southwest
(0.5 percent) and the Northeast (0.5 percent).
Table VII-6 provides a summary of the visibility degradation estimates in terms of
deciviews for Class I areas (i.e., recreational category) across U.S. visibility regions. As shown,
the national improvement in visibility for these areas is 0.6 percent, or 0.12 deciviews. Predicted
visibility improvements are the largest for the Northwest (1.4 percent), the Southeast (0.9
percent), and the Northeast/Midwest (0.7 percent). Smaller visibility improvements are predicted
in the Southwest (0.3 percent).
VII-21
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table VII-5. Summary of 2030 Visibility Degradation Estimates by Region: Residential
(Annual Average Deciviews)
Study Regions
Southeast
Southwest
California
Northeast
North Central
South Central
Rocky Mountain
Northwest
National Average (unweighted)
2030 Base Case
23.79
17.49
20.91
24.53
22.52
20.28
18.27
21.08
22.00
Change"
-0.27
-0.09
-0.04
-0.12
-0.29
-0.24
-0.16
-0.28
-0.23
Percent Change
-1.1%
-0.5%
-0.2%
-0.5%
-1.3%
-1.2%
-0.9%
-1.3%
-1.0%
1 The change is defined as the control case deciview level minus the base case deciview level.
Table VII-6. Summary of 2030 Visibility Degradation Estimates by Region: Recreational
(Annual Average Deciviews)
Class I Visibility Regions
Southeast
Southwest
California
Northeast/Midwest
Rocky Mountain
Northwest
National Average (unweighted)
2030 Base Case
22.78
17.61
20.54
21.34
17.80
22.09
19.99
Change0
-0.20
-0.05
-0.04
-0.15
-0.10
-0.32
-0.12
Percent Change
-0.9%
-0.3%
-0.2%
-0.7%
-0.6%
-1.4%
-0.6%
The change is defined as the control case deciview level minus the base case deciview level.
VII-22
-------
Chapter VII: Benefit-Cost Analysis
4. Nitrogen Deposition Estimates
This section presents the methods and results of estimating the potential reductions in
airborne nitrogen deposition loadings to estuaries associated with the motor vehicle Tier 2 and
fuel standards. A sampling of 12 estuaries (10 East Coast and 2 Gulf Coast estuaries) were used
for this analysis because of the availability of necessary data and their potential
representativeness. For each estuary, we completed the following steps as part of this analysis:
Baseline loadings of atmospherically supplied nitrogen were obtained from data provided in
Valigura et al (1996) and from local offices of the Chesapeake Bay Program and the National
Estuary Program,
Deposition from atmospheric emissions were divided into local and regional areas that
contribute to airborne nitrogen deposition,
• Deposition coefficients, which relate NOx emission changes from a source region to
nitrogen deposition changes at a receptor region, were derived for local and regional
contributors, and
• Changes in nitrogen deposition loadings were estimated by multiplying NOx emission
changes for the local and regional contributing areas by the appropriate deposition
coefficients.
For five of the 12 estuaries, estimates of both direct deposition to the tidal waters and
indirect deposition to the entire watershed were available from the literature. For the remaining
seven estuaries, only the direct deposition estimates were available. Therefore, to obtain indirect
deposition estimates where missing, we used RADM-derived nitrogen flux for the watershed
(Dennis, 1997). This analysis assumes that 10 percent of nitrogen deposited onto the watershed
is delivered via export (pass-through) to the estuary.13 This calculated indirect deposition value
is then added to the direct deposition value obtained from the literature to arrive at the total load
from atmospheric deposition.
As stated in Step 4 above, the nitrogen deposition results are heavily dependent upon the
deposition coefficients that estimate the impact of NOx emission changes on nitrogen deposition
loadings. For this analysis, two S-R coefficients, an alpha and a beta, were developed for each
estuary. The alpha coefficient relates local emissions to deposition and the beta coefficient
relates regional emissions to deposition. These coefficients are calculated for each estuary using
13 This assumption is consistent with reported case studies such as Valiela et al, 1997. These authors
report that 89% of atmospherically deposited nitrogen was retained by the watershed of Waquoit Bay, suggesting an
11% pass through factor.
VII-23
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
deposition outputs from RADM as employed for the final Regional NOx SIP Call (EPA, 1998b).
More detail on this approach and results may be found in Pechan-Avanti (1999).
Table VII-7 provides a summary of the baseline deposition and change in nitrogen
deposition estimates for the selected estuaries as a result of the Tier 2 rule. As shown,
implementation of the Tier 2 controls results in a 14.1 percent reduction in the average annual
deposition across these estuaries. These predicted reductions range from a low of 10.1 percent
for Delaware Inland Bay to a high of 15.5 percent for Long Island Sound.
Table VII-7. Summary of 2030 Nitrogen Deposition in Selected Estuaries and Changes Due
to Tier 2 Standards (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
2030 Base Case
13.20
4.63
20.71
4.08
0.59
1.44
4.59
6.71
1.58
1.37
0.46
2.96
62.32
Change"
-1.83
-0.65
-2.80
-0.57
-0.06
-0.21
-0.69
-1.04
-0.22
-0.20
-0.06
-0.44
-8.77
Percent Change
-13.8%
-14.1%
-13.5%
-14.0%
-10.1%
-14.9%
-15.1%
-15.5%
-14.2%
-14.3%
-14.1%
-14.9%
-14.1%
1 Change is defined here as the emissions level after implementing the Tier 2 rule minus the base case emissions.
VII-24
-------
Chapter VII: Benefit-Cost Analysis
C. Benefit Analysis
1. Methods for Estimating Benefits from Air Quality Improvements
Environmental and health economists have a number of methods for estimating the
economic value of improvements in (or deterioration of) environmental quality. The method
used in any given situation depends on the nature of the effect and the kinds of data, time and
resources that are available for investigation and analysis. This section provides an overview of
the methods EPA selected to monetize the benefits included in the Tier 2/Gasoline Sulfur RIA.
We note at the outset that EPA rarely has the time or resources to perform extensive new
research to measure economic benefits for individual rulemakings. As a result, our estimates are
based on the best available methods of benefits transfer. Benefits transfer is the science and art
of adapting primary benefits research from similar contexts to obtain the most accurate measure
of benefits for the environmental quality change under analysis. Where appropriate, adjustments
are made for the level of environmental quality change, the sociodemographic and economic
characteristics of the affected population, and other factors in order to improve the accuracy and
robustness of benefits estimates.
In general, economists tend to view an individual's willingness-to-pay for a improvement
in environmental quality as the appropriate measure of the value of a risk reduction. An
individual's willingness-to-accept (WTA) compensation for not receiving the improvement is
also a valid measure. However, WTP is generally considered to be a more readily available and
conservative measure of benefits. Adoption of WTP as the measure of value implies that the
value of environmental quality improvements is dependent on the individual preferences of the
affected population and that the existing distribution of income (ability to pay) is appropriate.
For many goods, WTP can be observed by examining actual market transactions. For
example, if a gallon of bottled drinking water sells for one dollar, it can be observed that at least
some persons are willing to pay one dollar for such water. For goods not exchanged in the
market, such as most environmental "goods," valuation is not as straightforward. Nevertheless, a
value may be inferred from observed behavior, such as sales and prices of products that result in
similar effects or risk reductions, (e.g., non-toxic cleaners or bike helmets). Alternatively,
surveys may be used in an attempt to directly elicit WTP for an environmental improvement.
One distinction in environmental benefits estimation is between use values and non-use
values. Although no general agreement exists among economists on a precise distinction
between the two (see Freeman, 1993), the general nature of the difference is clear. Use values
are those aspects of environmental quality that affect an individual's welfare more or less
directly. These effects include changes in product prices, quality, and availability, changes in the
quality of outdoor recreation and outdoor aesthetics, changes in health or life expectancy, and the
costs of actions taken to avoid negative effects of environmental quality changes.
VII-25
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Non-use values are those for which an individual is willing to pay for reasons that do not
relate to the direct use or enjoyment of any environmental benefit. Non-use values are most
frequently divided into two categories: existence values and bequest values. Existence values
refer to situations where individuals value (are willing to pay for) the knowledge of an improved
environmental state (or avoidance of a deteriorating environmental state). An example is the
willingness to pay (WTP) for the preservation of the blue whale even when an individual has no
plan to take a trip to observe the species nor to derive any direct benefit from its survival.
Existence values commonly rise from philosophical, ethical, or religious attitudes about the
rights of nature and the responsibilities of humans. The other commonly posited category of
non-use benefits is bequest value. People are willing to devote resources to environmental
preservation because of their perceived obligation or desire to leave higher states of
environmental quality to future generations. Bequest values can also be thought of as arising
from the philosophical, ethical, and religious beliefs of individuals.
Non-use values are not traded, directly or indirectly, in markets. For this reason, the
measurement of non-use values has proved to be significantly more difficult than the
measurement of use values. The air quality changes produced by the final Tier 2/Gasoline Sulfur
rule cause changes in both use and non-use values, but the monetary benefit estimates are almost
exclusively for use values.
More frequently than not, the economic benefits from environmental quality changes are
not traded in markets, so direct measurement techniques can not be used. Avoided cost methods
are ways to estimate the costs of pollution by using the expenditures made necessary by pollution
damage. For example, if buildings must be cleaned or painted more frequently as levels of PM
increase, then the appropriately calculated increment of these costs is a reasonable estimate of
true economic benefits when PM levels are reduced. A variation on the avoided cost method is
used to provide an alternative estimate of the benefits of reductions in nitrogen deposition to
estuaries (see Sections C.4 and F). Avoided costs methods are also used to estimate some of the
health-related benefits related to morbidity, such as hospital admissions (see section D).
Indirect market methods can also be used to infer the benefits of pollution reduction. The
most important application of this technique for our analysis is the calculation of the value of a
statistical life for use in the estimate of benefits from mortality reductions. There exists no
market where changes in the probability of death are directly exchanged. However, people make
decisions about occupation, precautionary behavior, and other activities associated with changes
in the risk of death. By examining these risk changes and the other characteristics of people's
choices, it is possible to infer information about the monetary values associated with changes in
mortality risk (see section D). For measurement of health benefits, this analysis captures the
WTP for most use and non-use values, with the exception of the value of avoided hospital
admissions, which only captures the avoided cost of illness.
The most direct way to measure the economic value of air quality changes is in cases
VII-26
-------
Chapter VII: Benefit-Cost Analysis
where the endpoints have market prices. For the final rule, this can only be done for effects on
commercial agriculture and forestry. Well-established economic modeling approaches are used
to predict price changes that result from predicted changes in agricultural and forestry outputs.
Consumer and producer surplus measures can then be developed to give reliable indications of
the benefits of changes in ambient air quality for these categories (see section E).
Estimating benefits for visibility and ecosystem services is a more difficult and less
precise exercise because the endpoints are not directly or indirectly valued in markets. For
example, the loss of a species of animal or plant from a particular habitat does not have a well-
defined price. The contingent valuation method (CVM) has been employed in the economics
literature to value endpoint changes for both visibility and ecosystem functions (Chestnut and
Dennis, 1997). CVM values endpoints by using carefully structured surveys to ask a sample of
people what amount of compensation is equivalent to a given change in environmental quality.
There is an extensive scientific literature and body of practice on both the theory and technique
of CVM. EPA believes that well-designed and well-executed CVM studies are valid for
estimating the benefits of air quality regulation14.
2. Methods for Describing Uncertainty
In any complex analysis using estimated parameters and inputs from numerous models,
there are likely to be many sources of uncertainty15. This analysis is no exception. As outlined
both in this and preceding chapters, there are many inputs used to derive the final estimate of
benefits, including emission inventories, air quality models (with their associated parameters and
inputs), epidemiological estimates of concentration-response (C-R) functions, estimates of values
(both from WTP and cost-of-illness studies), population estimates, income estimates, and
1 "Concerns about the reliability of value estimates that come from CVM 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 WTP responses are consistent
with established economic principles.
15 It should be recognized that in addition to uncertainty, the annual benefit estimates for the final Tier
2/Gasoline Sulfur rule presented in this analysis are also inherently variable, due to the truly random processes that
govern pollutant emissions and ambient air quality in a given year. Factors such as electricity demand and weather
display constant variability regardless of our ability to accurately measure them. As such, the estimates of annual
benefits should be viewed as representative of the types of benefits that will be realized, rather than the actual
benefits that would occur every year.
VII-27
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
estimates of the future state of the world (i.e., regulations, technology, and human behavior).
Each of these inputs may be uncertain, and depending on their location in the benefits analysis,
may have a disproportionately large impact on final estimates of total benefits. For example,
emissions estimates are used in the first stage of the analysis. As such, any uncertainty in
emissions estimates will be propagated through the entire analysis. When compounded with
uncertainty in later stages, small uncertainties in emission levels can lead to much larger impacts
on total benefits. A more thorough discussion of uncertainty can be found in the benefits
technical support document (TSD) for this RIA, Tier IIFinal Rule: Air Quality Estimation,
Selected Health and Welfare Benefits Methods, and Benefit Analysis Results (Abt Associates,
1999).
Some key sources of uncertainty in each stage of the benefits analysis are:
• gaps in scientific data and inquiry
variability in estimated relationships, such as C-R functions, introduced through
differences in study design and statistical modeling
errors in measurement and projection for variables such as population growth
rates
errors due to misspecification of model structures, including the use of surrogate
variables, such as using PM10 when PM2 5 is not available, excluded variables, and
simplification of complex functions
• biases due to omissions or other research limitations.
Some of the key uncertainties in the benefits analysis are presented in Table VII-8. Given
the wide variety of sources for uncertainty and the potentially large degree of uncertainty about
any primary estimate, it is necessary for us to address this issue in several ways. These include
qualitative discussions, probabilistic assessments, and alternative calculations. For some
parameters or inputs it may be possible to provide a statistical representation of the underlying
uncertainty distribution. For other parameters or inputs, the information necessary to estimate an
uncertainty distribution is not available. Even for individual endpoints, there is usually more
than one source of uncertainty. This makes it difficult to provide a quantified uncertainty
estimate. For example, the C-R function used to estimate avoided premature mortality has an
associated standard error which represents the sampling error around the pollution coefficient in
the estimated C-R function. It would be possible to report a confidence interval around the
estimated incidences of avoided premature mortality based on this standard error. However, this
would omit the contribution of air quality changes, baseline population incidences, projected
populations exposed, and transferability of the C-R function to diverse locations to uncertainty
about premature mortality. Thus, a confidence interval based on the standard error would
provide a misleading picture about the overall uncertainty in the estimates. Information on the
uncertainty surrounding particular C-R and valuation functions is provided in the benefits TSD
for this RIA (Abt Associates, 1999). But, this information should be interpreted within the
context of the larger uncertainty surrounding the entire analysis.
VII-28
-------
Chapter VII: Benefit-Cost Analysis
Our approach to characterizing model uncertainty in the estimate of total benefits is to
present a primary estimate, based on the best available scientific literature and methods, and to
then provide alternative calculations to illustrate the effects of uncertainty about key analytical
assumptions. We do not attempt to assign probabilities to these alternative calculations, as we
believe this would only add to the uncertainty of the analysis or present a false picture about the
precision of the results16. Instead, the reader is invited to examine the impact of applying the
16 Some recent benefit-cost analyses in Canada and Europe (Holland, et al., 1999; Lang, et al., 1995) have
estimated ranges of benefits by assigning ad hoc probabilities to ranges of parameter values for different endpoints.
Although this does generate a quantitative estimate of an uncertainty range, the estimated points on these
distributions are themselves highly uncertain and very sensitive to the subjective judgements of the analyst. To
avoid these subjective judgements, we choose to allow the reader to determine the weights they would assign to
alternative estimates.
VII-29
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table VII-8. Primary 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.
-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 literature supporting a direct biological mechanism for observed epidemiological evidence.
-Direct causal agents within the complex mixture of PM responsible for reported health effects have not been
identified.
-The extent to which adverse health effects are associated with low level exposures that occur many times in the year
versus peak exposures.
-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.
'. 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
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
will cause total benefits to be underestimated.
VII-30
-------
Chapter VII: Benefit-Cost Analysis
different assumptions on the estimate of total benefits. While it is possible to combine all of the
alternative calculations with a positive impact on benefits to form a "high" estimate or all of the
alternative calculations with a negative impact on benefits to form a "low" estimate, this would
not be appropriate because the probability of all of these alternative assumptions occurring
simultaneously is extremely low. Instead, the alternative calculations are intended to
demonstrate the sensitivity of our benefits results to key parameters which may be uncertain.
Alternative calculations are presented in Table VII-18.
Many benefits categories, while known to exist, do not have enough information
available to provide a quantified or monetized estimate. The uncertainty regarding these
endpoints is such that we could determine neither a primary estimate nor a plausible range of
values.
Our estimate of total benefits should be viewed as an approximate result because of the
sources of uncertainty discussed above (see Table VII-8). The total benefits estimate may
understate or overstate actual benefits of the rule. The remainder of this section describes in
greater detail two potential sources of uncertainty that can impact multiple aspects of the
analysis: 1) the inability to quantify or monetize many of the benefits and costs associated with
the rule; and 2) adjustments for changes in income in the future.
a. Unquantifiable Environmental Benefits and Costs
In considering the monetized benefits estimates, the reader should remain aware of the
many limitations for conducting these analyses mentioned throughout this RIA. One significant
limitation of both the health and welfare benefits analyses is the inability to quantify many of the
PM and ozone-induced adverse effects listed in Table VII-1. For many health and welfare
effects, such as PM-related materials damage, reliable C-R functions and/or valuation functions
are not currently available. In general, if it were possible to monetize these benefits categories,
the benefits estimates presented in this analysis would increase. Unquantified benefits are
qualitatively discussed in the health and welfare effects sections. In addition to unquantified
benefits, there may also be environmental costs that we are unable to quantify. Several of these
environmental cost categories are related to nitrogen deposition, while one category is related to
the issue of ultraviolet light. These endpoints are qualitatively discussed in the health and
welfare effects sections as well. The net effect of excluding benefit and disbenefit categories
from the estimate of total benefits depends on the relative magnitude of the effects.
VII-31
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
b. Projected Population and Income Growth
As indicated above, our analysis predicts the benefits of the Tier 2/Gasoline Sulfur rule in
2030. As such, we use projections of populations in 2030. The total projected population in the
47 states covered by the Tier 2/Gasoline Sulfur rule in 2030 is 300 million. The total projected
population potentially affected by air quality changes resulting from the Tier 2/Gasoline Sulfur
rule includes California, which adds an additional 45 million people. These projections are
uncertain, although they are based on projection methods used by the U.S. Census Bureau (Abt
Associates, 1999b). To the extent that populations are over- or under-predicted, benefits may be
over- or under-stated.
Our analysis does not attempt to adjust benefits estimates to reflect expected growth in
real income. Economic theory argues, however, that WTP for most goods (such as
environmental protection) will increase if real incomes increase. There is substantial empirical
evidence that the income elasticity17 of WTP for health risk reductions is positive, although there
is uncertainty about its exact value. While many analyses assume that the income elasticity of
WTP is unit elastic (i.e., ten percent higher income level implies a ten percent higher willingness
to pay to reduce risk changes), empirical evidence suggests that income elasticity is substantially
less than one and thus inelastic. The effects of income changes on WTP estimates can influence
benefit estimates in two different ways: (i) as changes that reflect estimates of income change in
the affected population over time; and (ii) as changes based on differences in income between
study populations and the affected populations at a particular time. Empirical evidence of the
effect of income on WTP gathered to date is based on studies examining the latter. Income
elasticity adjustments to better account for changes over time, therefore, will necessarily be based
on potentially inappropriate data. The degree to which WTP may increase for the specific health
and welfare benefits provided by the final Tier 2/Gasoline Sulfur rule is not estimated due to the
high degree of uncertainly in the income elasticity information.
D. Assessment of Human Health Benefits
The most significant monetized benefits of reducing ambient concentrations of PM and
ozone are attributable to reductions in health risks associated with air pollution. EPA's criteria
documents for ozone and PM list numerous health effects known to be linked to ambient
concentrations of these pollutants (EPA, 1996). This section describes individual effects and the
methods used to quantify and monetize changes in the expected number of incidences of various
health effects.
"Income elasticity is a common economic measure equal to the percentage change in WTP for a one
percent change in income.
VII-32
-------
Chapter VII: Benefit-Cost Analysis
In section 1, we discuss how we have determined the baseline incidences for the health
effects impacted by changes in PM and ozone. In section 2, we explain how we deal with the
issue of health effects thresholds. In section 3, we describe how we quantify and value changes
in individual health effects. Finally, in section 4 we present quantified estimates of the
reductions in health effects resulting from the Tier 2 standards and their associated monetary
values.
1. Estimating Baseline Incidences for Health Effects
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 total mortality rates used in
the estimation of PM-related premature 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) (U.S. Centers for Disease
Control, 1999a). An additional adjustment was necessary to provide baseline incidences for
adults 30 and older for use in the Pope, et al. (1995) PM mortality C-R function. We estimated
county-specific baseline mortality incidences for this population 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. We used national
incidence rates whenever possible, because these data are most applicable to a national
assessment of benefits. However, for some studies, the only available incidence information
comes from the studies themselves; in these cases, incidence in the study population is assumed
to represent typical incidence at the national level.
2. Accounting for Potential Health Effect Thresholds
When conducting clinical (chamber) and epidemiological studies, C-R functions may be
estimated with or without explicit thresholds. Air pollution levels below the threshold are
assumed to have no associated adverse health effects. When a threshold is not assumed, as is
often the case in epidemiological studies, any exposure level is assumed to pose a non-zero risk
VII-33
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
of response to at least one segment of the population.
The possible existence of an effect threshold is a very important scientific question and
issue for policy analyses such as the final Tier 2/Gasoline Sulfur RIA. In the benefits analysis for
the Tier 2 Proposal RIA, the low-end estimate of benefits assumed a threshold in PM health
effects at 15 |ig/m3. However, the most recent advice from EPA's Science Advisory Board is
that there is currently no scientific basis for selecting a threshold of 15 |ig/m3 or any other
specific threshold for the PM-related health effects considered in this analysis (EPA-SAB-
Council-ADV-99-012, 1999). Therefore, for our benefits analysis of the final Tier 2/Gasoline
Sulfur rule, we assume there are no thresholds for modeling health effects. It is not appropriate
to adopt a threshold for use in either the primary analysis or any alternative calculations because
there is no adequate scientific evidence to support such a calculation. The potential impact of a
health effects threshold on avoided incidences of PM-related premature mortality is explored as a
key sensitivity analysis presented in Appendix VII-A.
3. Quantifying and Valuing Individual Health Endpoints
Health benefits of the final Tier 2/Gasoline Sulfur rule may be related to ozone only, PM
only, or both pollutants. The ozone only health effects included in our primary benefits estimate
are chronic asthma in adult males and decreased worker productivity. The PM only health
effects include premature mortality, chronic bronchitis, acute bronchitis, upper and lower
respiratory symptoms, shortness of breath, and work loss days18. The health effects related to
both PM and ozone include hospital admissions, and minor restricted activity days.
For this analysis, we rely on C-R functions estimated in published epidemiological
studies relating adverse health effects to ambient air quality. The specific studies from which C-
R functions are drawn are included in Table "VTI-9. A complete discussion of the C-R functions
used for this analysis is contained in the benefits TSD for this RIA (Abt Associates, 1999).
18 In the benefits analysis for the Tier 2 Proposal RIA, we also estimated reductions in the incidence of
premature mortality associated with reduced exposures to ozone. At least some evidence has been found linking
both PM and ozone with premature mortality. The SAB has raised concerns that mortality-related benefits of air
pollution reductions may be overstated if separate pollutant-specific estimates, some of which may have been
obtained from models excluding the other pollutants, are aggregated. In addition, there may be important
interactions between pollutants and their effect on mortality (EPA-SAB-Council-ADV-99-012, 1999)
The Pope et al. (1995) study used to quantify PM-related mortality included only PM, so it is unclear to
what extent it may include the impacts of ozone or other gaseous pollutants. Because of concern about overstating
of benefits and because the evidence associating mortality with exposure to paniculate matter is currently stronger
than for ozone, only the benefits of PM-related premature mortality avoided are included in the total benefits
estimate. The benefits associated with ozone reductions are presented as a sensitivity analysis in Appendix VII-A
but are not included in the estimate of total benefits.
VII-34
-------
Chapter VII: Benefit-Cost Analysis
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-1 and described more fully in
the ozone and PM criteria documents (EPA, 1996a, 1996b), we include only a subset of health
effects in this quantified benefit analysis. Health effects are excluded from this analysis for three
reasons: (i) the possibility of double counting (such as hospital admissions for specific
respiratory diseases); (ii) uncertainties in applying effect relationships based on clinical studies to
the affected population; or (iii) 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 pollutant
and a given health effect, or "endpoint," applying the C-R function is straightforward. This is the
case for most of the health endpoints selected for inclusion in the benefits analysis. A
Table VII-9. Endpoints and Studies Included in the Primary Analysis
Endpoint
Mortality
Ages 30 and Older
Chronic Illness
Chronic Bronchitis
Chronic Asthma
Hospital Admissions
All Respiratory
Total Cardiovascular
Asthma-Related ER Visits
Other Illness
Acute Bronchitis
Upper Respiratory Symptoms
Lower Respiratory Symptoms
Shortness of Breath
Work Loss Days
Minor Restricted Activity Days / Any of 19
respiratory Symptoms
Pollutant
PM
PM
Ozone
PM, Ozone
PM, Ozone
PM, Ozone
PM
PM
PM
PM
PM
PM, Ozone
Study
Pope etal. (1995)
Multiple Studies
McDonnell et al. (1999)
Multiple Studies
Multiple Studies
Multiple Studies
Dockeryetal. (1996)
Pope etal. (1991)
Schwartz et al. (1994)
Ostroetal. (1995)
Ostro(1987)
Multiple Studies
Study Population
Adults, 30 and older
Multiple Studies
Non-asthmatic adults, 27
and older
Multiple Studies
Multiple Studies
Multiple Studies
Children, 8-12
Asthmatic children, 9-11
Children, 7-14
African American
asthmatic children, 7-12
Adults, 18-65
Multiple Studies
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
VII-35
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
more generalized and comprehensive.
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. The benefits TSD provides details of the procedures used to combine multiple C-R
functions (Abt Associates, 1999). Pooled C-R functions are used to estimate incidences of
chronic bronchitis related to PM exposure, hospital admissions from cardiovascular and
respiratory causes related to PM and ozone exposure, and emergency room visits for asthma
related to PM and ozone exposure.
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 at all locations in the U.S. 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 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 the general application of a single C-R function everywhere.
The appropriate economic value of a change in a health effect depends on whether the
health effect is viewed ex ante (before the effect has occurred) or ex post ( after the effect has
occurred). Reductions in ambient concentrations of air pollution generally lower the risk of
future adverse health affects by a fairly small amount for a large population. The appropriate
economic measure is therefore ex-ante WTP for changes in risk. However, epidemiological
studies generally provide estimates of the expected number of incidences of a particular health
effect avoided due to a reduction in air pollution. A convenient way to use this data in a
consistent framework is to convert probabilities to units of avoided statistical incidences. This
measure is calculated by dividing individual WTP for a risk reduction by the related observed
change in risk. For example, suppose a measure is able to reduce the risk of premature mortality
from 2 in 10,000 to 1 in 10,000 (a reduction of 1 in 10,000). If individual WTP for this risk
reduction is $100, then the WTP for an avoided statistical premature mortality amounts to $1
million ($100/0.0001 change in risk). Using this approach, the size of the affected population is
automatically taken into account by the number of incidences predicted by epidemiological
studies applied to the relevant population. The same type of calculation can produce values for
statistical incidences of other health endpoints.
For some health effects, such as hospital admissions, WTP estimates are generally not
available. In these cases, we use the cost of treating or mitigating the effect as an alternative
estimate. For example, for the valuation of hospital admissions we use the avoided medical costs
as an estimate of the value of avoiding the health effects causing the admission. These costs of
VII-36
-------
Chapter VII: Benefit-Cost Analysis
illness (COI) estimates generally understate the true value of avoiding a health effect. They tend
to reflect the direct expenditures related to treatment but not the value of avoided pain and
suffering from the health effect. Table VII-10 summaries the value estimates per health effect
that we use in this analysis. Alternative values used to derive the alternative estimates listed in
Table VII-18 are indicated in parentheses. Note that there is not a specific value listed for
hospital admissions. This reflects the fact that there are a range of symptoms for which
individuals are admitted, each of which has a different associated cost. The estimated benefit of
avoided hospital admissions reflects the distribution of symptoms across the total incidence of
hospital admissions. The study-specific values for hospital admissions can be found in the
benefits TSD for this RIA (Abt Associates, 1999).
In the following sections, we describe individual health endpoints and the C-R functions
we have selected to provide quantified estimates of the avoided health effects associated with the
final Tier 2/Gasoline Sulfur rule. In addition, we discuss how these changes in health effects
should be valued and indicate the value functions selected to provide monetized estimates of the
value of changes in health effects.
VII-37
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table VII-10. Unit Values Used for Economic Valuation of Health Endpoints
Health or Welfare
Endpoint
Mortality
Chronic Bronchitis (CB)
Chronic Asthma
Estimated Value
Per Incidence
(1997$)
Central Estimate
$5.9 million per
statistical life
$319,000
$31,000
Derivation of Estimates
Value is the mean of value-of-statistical-life estimates from 26
studies (5 contingent valuation and 21 labor market studies)
reviewed for the section 812 Prospective analysis.
Value is the mean of a generated distribution of WTP to avoid a
case of pollution-related CB. WTP to avoid a case of pollution-
related CB is derived by adjusting WTP (as described in Viscusi et
al., 1991) to avoid a severe case of CB for the difference in severity
and taking into account the elasticity of WTP with respect to
severity of CB.
Based on results reported in two studies (Blumenschein and
Johannesson, 1998; O'Connor and Blomquist, 1997). Assumes a
5% discount rate and reflects adjustments for age distribution
among adults (ages 27 and older) and projected life years
remaining.
Hospital Admissions
All Respiratory
(ICD codes: 460-519)
All Cardiovascular
(ICD codes: 390-429)
Emergency room visits for
asthma
variable —
function of the
analysis
variable —
function of the
analysis
$280
The COI estimates are based on ICD-9 code level information (e.g.,
average hospital care costs, average length of hospital stay, and
weighted share of total respiratory illnesses) reported in Elixhauser
(1993).
The COI estimates are based on ICD-9 code level information (e.g.,
average hospital care costs, average length of hospital stay, and
weighted share of total cardiovascular illnesses) reported in
Elixhauser (1993).
COI estimate based on data reported by Smith et al. (1997).
Respiratory Ailments Not Requiring Hospitalization
Upper Respiratory Symptoms
(URS)
Lower Respiratory. Symptoms
(LRS)
$23
$15
Combinations of the 3 symptoms for which WTP estimates are
available that closely match those listed by Pope et al. result in 7
different "symptom clusters," each describing a "type" of URS. A
dollar value was derived for each type of URS, using mid-range
estimates of WTP (lEc, 1994) to avoid each symptom in the cluster
and assuming additivity of WTPs. The dollar value for URS is the
average of the dollar values for the 7 different types of URS.
Combinations of the 4 symptoms for which WTP estimates are
available that closely match those listed by Schwartz et al. result in
11 different "symptom clusters," each describing a "type" of LRS.
A dollar value was derived for each type of LRS, using mid-range
estimates of WTP (lEc, 1994) to avoid each symptom in the cluster
and assuming additivity of WTPs. The dollar value for LRS is the
average of the dollar values for the 11 different types of LRS.
VII-38
-------
Chapter VII: Benefit-Cost Analysis
Health or Welfare
Endpoint
Acute Bronchitis
Shortness of Breath
Estimated Value
Per Incidence
(1997$)
Central Estimate
$55
$5.30
Derivation of Estimates
Average of low and high values recommended for use in Section
812 analysis (Neumann, et al. 1994)
From Ostro, et al. 1995. This is the mean of the median estimates
from two studies to avoid a day of shortness of breath, Dickie et al.
1991 (0.00) and Loehman et al., 1979 (13.00).
Restricted Activity and Work Loss Days
Work Loss Days (WLDs)
Minor Restricted Activity Days
(MRADs)
Variable
$47
Regionally adjusted median weekly wage for 1990 divided by 5
(adjusted to 1997$) (U.S. Bureau of the Census, 1992).
Median WTP estimate to avoid 1 MRRAD - minor respiratory
restricted activity day — from Tolley et al.(1986) .
a. Premature Mortality: Quantification
Both acute and chronic exposures to ambient levels of air pollution have been associated
with increased risk of premature mortality. Because of the extreme nature of this endpoint and
the high monetary value associated with risks to life, reductions in the risk of premature mortality
are the most important health endpoint quantified in this analysis, accounting for over 90 percent
of the total monetized benefits. However, considerable uncertainty exists, both among
economists and policymakers, as to the appropriate way to value reductions in mortality risks.
Because of these factors, we include a more detailed discussion for premature mortality than for
other health effects.
Health researchers have consistently linked air pollution, especially PM, with increases in
premature mortality. A substantial body of published scientific literature recognizes a correlation
between elevated PM concentrations and increased mortality rates. Much of this literature is
summarized in the 1996 PM Criteria Document (U.S. EPA, 1996a). There is much about this
relationship that is still uncertain, however, as stated in preamble to the 1997 PM National
Ambient Air Quality Standards (U.S. EPA. 40 CFR 50, 1997), "the consistency of the results of
the epidemiological studies from a large number of different locations and the coherent nature of
the observed effects are suggestive of a likely causal role of ambient PM in contributing to the
reported effects," which include premature mortality. The National Academy of Sciences, in
their report on research priorities for PM (National Academy of Sciences, 1998), indicate that
"there is a great deal of uncertainty about the implications of the findings [of an association
between PM and premature mortality] for risk management, due to the limited scientific
information about the specific types of particles that might cause adverse health effects, the
contributions of particles of outdoor origin to actual human exposures, the toxicological
mechanisms by which the particles might cause adverse health effects, and other important
VII-39
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
questions." EPA acknowledges these uncertainties, however, for this analysis, we assume a
causal relationship between exposure to elevated PM and premature mortality, based on the
consistent evidence of a correlation between PM and mortality reported in the scientific literature
(U.S. EPA, 1996a).
It is currently unknown whether there is a time lag (a delay between changes in PM
exposures and changes in mortality rates) in the chronic PM/premature mortality relationship.
The existence of such a lag is important for the valuation of premature mortality incidences
because economic theory suggests that benefits occurring in the future should be discounted.
Although there is no specific scientific evidence of the existence or structure of a PM effects lag,
current scientific literature on adverse health effects, such as those associated with PM (e.g.,
smoking-related disease) and the difference in the effect size between chronic exposure studies
and daily mortality studies suggest that all incidences of premature mortality reduction associated
with a given incremental change in PM exposure probably would not occur in the same year as
the exposure reduction. This same smoking-related literature implies that lags of up to a few
years are plausible. Adopting the lag structure used in a illustrative calculation for the Tier 2
Proposal RIA and endorsed by the SAB (EPA-SAB-COUNCIL-ADV-00-001, 1999), we assume
a five-year lag structure, with 25 percent of premature deaths occurring in the first year, another
25 percent in the second year, and 16.7 percent in each of the remaining three years. To explore
the uncertainty surrounding this lag structure, Appendix VII-A contains a sensitivity analysis
showing how different lag structures affect the estimated value of reductions in premature
mortality.
Two types of exposure studies (short-term and long-term) have been used to estimate a
PM/premature mortality relationship. Short-term exposure studies attempt to relate short-term
(often day-to-day) changes in PM concentrations and changes in daily mortality rates up to
several days after a period of elevated PM concentrations. Long-term exposure studies examine
the potential relationship between longer-term (e.g., annual) changes in exposure to PM and
annual mortality rates. Researchers have found significant correlations using both types of
studies (U.S. EPA, 1996a); however, for this analysis, following SAB advice (EPA-SAB-
COUNCIL-ADV-99-005, 1999) we rely exclusively on long-term studies to quantify PM
mortality effects.
Following guidance from the SAB (EPA-SAB-COUNCIL-ADV-99-005, 1999), we
prefer studies to use long-term studies that employ a prospective cohort design over those that
use an ecologic or population-level design. Prospective cohort studies follow individuals
forward in time for a specified period, periodically evaluating each individual's exposure and
health status. While the long-term study design is preferred, they are expensive to conduct and
consequently there are relatively few well designed long-term studies. For PM, there have been
VII-40
-------
Chapter VII: Benefit-Cost Analysis
only a few, and the SAB has explicitly recommended use of only one — the Pope, et al. (1995)
prospective cohort study in estimating avoided premature mortality from reductions in ambient
PM concentrations (EPA-SAB-COUNCIL-ADV-99-005, 1999). We follow this
recommendation and are consistent with the modeling of mortality effects of PM in both the
Section 812 Retrospective and Prospective Reports to Congress. The Pope et al. study is
recommended 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.
Although we use the Pope study exclusively to derive our primary estimates of avoided
premature mortality, the C-R function based on Dockery et al. (1993) may provide a reasonable
alternative estimate (EPA-SAB-COUNCIL-ADV-99-012, 1999). While the Dockery et al. study
used a smaller sample of individuals from fewer cities than the study by Pope et al., it features
improved exposure estimates, a slightly broader study population (adults aged 25 and older), and
a follow-up period nearly twice as long as that of Pope et al. The Dockery et al. (1993) study
finds a larger effect of PM on premature mortality. We present an alternative estimate of
premature adult mortality associated with long-term PM exposure based on Dockery et al. (1993)
in Table VII-18. We emphasize, however, that based on SAB advice, the Pope et al. (1995)
derived estimate is our primary estimate of the effect of the final Tier 2/Gasoline Sulfur rule on
this important health effect.
b. Premature Mortality: Valuation
We estimate the monetary benefit of reducing premature mortality risk using the "value of
statistical lives saved" (VSL) approach, even though the actual valuation is of small changes in
mortality risk experienced by a large number of people. The VSL approach applies information
from several value-of-life studies to determine a reasonable benefit of preventing premature
mortality. The mean value of avoiding one statistical death is estimated to be $5.9 million in
1997 dollars. This represents an intermediate value from a variety of estimates that appear in the
economics literature, and is a value EPA has frequently used in RIAs for other rules and in the
Section 812 reports to Congress. This estimate is the mean of a distribution fitted to the
estimates from 26 value-of-life studies identified in the Section 812 reports 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 as 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-11. As
indicated in the previous section on quantification of premature mortality benefits, we assume for
VII-41
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
this analysis that some of the incidences of premature mortality related to PM exposures occur in
a distributed fashion over the five years following exposure. To take this into account in the
valuation of reductions in premature mortality, we apply an annual five percent discount rate to
the value of premature mortality occurring in future years.
The economics literature concerning the appropriate method for valuing reductions in
premature mortality risk is still developing. Some of the alternative approaches that have been
proposed for valuing reductions in mortality risk are discussed in Text Box 1. There is general
agreement that the value to an individual of a reduction in mortality risk can vary based on
several factors, including the age of the individual, the type of risk, the level of control the
individual has over the risk, the level of risk aversion, and the health status of the individual.
While the empirical basis for adjusting the $5.9 million VSL for many of these factors does not
yet exist, a thorough discussion of these factors is contained in the benefits TSD for this RIA
(Abt Associates, 1999). EPA recognizes the need for investigation by the scientific community
to develop additional empirical support for adjustments to VSL for the factors mentioned above.
VII-42
-------
Chapter VII: Benefit-Cost Analysis
Table VII-11. Summary of Mortality Valuation Estimates'1
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 1997 $)
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 S, as detailed in (Abt Associates, 1999).
VII-43
-------
Text Box 1
Alternative Approaches for Assessing the Value of Reduced Mortality Risk
Stated preference studies - These studies use survey responses to estimate WTP to avoid risks. Strengths:
flexible approach allowing for appropriate risk context, good data on WTP for individuals. Weaknesses: Risk
information may not be well-understood by respondents and questions may be unfamiliar.
Consumer market studies - These studies use consumer purchases and risk data (e.g. smoke detectors) to
estimate WTP to avoid risks. Strengths: uses revealed preferences and is a flexible approach. Weaknesses:
very difficult to estimate both risk and purchase variables.
Value of statistical life year - Provides an annual equivalent to value of statistical life estimates. Strengths:
provides financially accurate adjustment for age at death. Weaknesses: adjustment may not reflect how
individuals consider life-years; assumes equal value for all remaining life-years.
Quality adjusted life year - Applies quality of life adjustment to life-extension data, uses cost-effectiveness
data to value. Strengths: widely used in public health literature to assess private medical interventions.
Weaknesses: lack of data on health state indices and life quality adjustments that are applicable to an air
pollution context.
WTP for a change in survival curve - Reflects WTP for change in risk, potentially incorporates age-specific
nature of risk reduction. Strengths: theoretically preferred approach that most accurately reflects risk reductions
from air pollution control. Weaknesses: almost no empirical literature available; difficulty in obtaining reliable
values.
WTP for a change in longevity - Uses stated preference approach to generate WTP for longevity or longer life
expectancy. Strengths: life expectancy is a familiar term to most individuals. Weaknesses: does not incorporate
age-specific risk information; problems in adapting to air pollution context.
Cost-effectiveness - Determines the implicit cost of saving a life or life-year. Strengths: widely used in public
health contexts. Weaknesses: health context is for private goods, dollar values do not necessarily reflect
individual preferences.
One important factor in Text Box 1 for which the impact on total benefits can be
illustrated is the difference in age distribution between the population affected by air pollution
and the population for which most of the VSL estimates were developed. To address this factor
we use the "value of statistical life-years lost" (VSLY) approach, recommended by the SAB as an
appropriate alternative to the VSL approach (EPA-SAB-COUNCIL-ADV-98-003, 1998). To
employ the value of statistical life-year (VSLY) approach, we first estimated the age distribution
of those lives projected to be saved by reducing air pollution. Based on life expectancy tables, we
calculate the life-years saved from each statistical life saved within each age and gender cohort.
To value these statistical life-years, we hypothesized a conceptual model which depicted the
VII-44
-------
Chapter VII: Benefit-Cost Analysis
relationship between the value of life and the value of life-years. The average number of life-
years saved across all age groups for which data were available is 14 for PM-related mortality.
The average for PM, in particular, differs from the 35-year expected remaining lifespan derived
from existing wage-risk studies. Using the same distribution of value of life estimates used
above, we estimated a distribution for the value of a life-year and combined it with the total
number of estimated life-years lost. The details of these calculations are presented in the TSD
for this RIA (Abt Associates, 1999).
c. Chronic Bronchitis: Quantification
Chronic bronchitis is characterized by mucus in the lungs and a persistent wet cough for
at least three months a year for several years in a row. Chronic bronchitis affects an estimated
five percent of the U.S. population (American Lung Association, 1999). There are a limited
number of studies that have estimated the impact of air pollution on chronic bronchitis.
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 Section 812
Prospective Report (U.S. EPA, 1999a), our analysis pools the estimates from these studies 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. Following the §812 Prospective Report, we assume that it is appropriate 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 this percentage change applies to a
baseline incidence rate obtained from another source. For example, 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..
VII-45
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
d. Chronic Bronchitis: Valuation
The best available estimate of WTP to avoid a case of chronic bronchitis (CB) comes
from Viscusi et al. (1991)19. The Viscusi et al. study, however, describes a severe case of CB to
the survey respondents. We therefore employ an estimate of WTP to avoid a pollution-related
case of CB, based on adjusting an adjustment of the Viscusi et al. (1991) estimate of the WTP to
avoid a sever case. This is done to account for the likelihood that an average case of pollution-
related CB is not as severe. The adjustment is made by applying the elasticity of WTP with
respect to severity reported in the Krupnick and Cropper (1992) study. Details of this adjustment
procedure are provided in the benefits TSD for this RIA (Abt Associates, 1999).
We use the mean of a distribution of WTP estimates as the central tendency estimate of
WTP to avoid a pollution-related case of CB in this analysis. The distribution incorporates
uncertainty from three sources: (1) the WTP to avoid a case of severe CB, as described by
Viscusi et al.; (2) the severity level of an average pollution-related case of CB (relative to that of
the case described by Viscusi et al.; and (3) the elasticity of WTP with respect to severity of the
illness. Based on assumptions about the distributions of each of these three uncertain
components, we derive a distribution of WTP to avoid a pollution-related case of CB by
statistical uncertainty analysis techniques. The expected value (i.e., mean or average) of this
distribution, which is about $320,000 (1997$), is taken as the central tendency estimate of WTP
to avoid a PM-related case of CB. We describe the three underlying distributions, and the
generation of the resulting distribution of WTP, in the benefits TSD for this RIA (Abt
Associates, 1999).
e. Chronic Asthma: Quantification
Chronic asthma is characterized by repeated incidences of inflammation of the lungs.
This causes restriction in the airways and results in shortness of breath, wheezing, and coughing.
Asthma is also characterized by airway hyper responsiveness to stimuli. Chronic asthma affects
over seven percent of the U.S. population (U.S. Centers for Disease Control, 1999b). Most
studies have not identified an association between air quality and asthma. However, a recent
study by McDonnell et al. (1999) provides a statistical association between ozone and the
development of asthma in adult white, non-Hispanic males. Following the advice of the EPA
19The Viscusi et al. (1991) study was an experimental study intended to examine new methodologies for
eliciting values for morbidity endpoints. Although these studies were not specifically designed for policy analysis,
the EPA Science Advisory Board (EPA-SAB-COUNCIL-ADV-00-002, 1999) has indicated that the severity
adjusted values from this study 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.
VII-46
-------
Chapter VII: Benefit-Cost Analysis
Science Advisory Board (EPA-SAB-COUNCIL-ADV-00-001, 1999) and the Section 812
Prospective Report, we have added this significant health effect to our benefit analysis since the
proposal RIA. However, it should be noted that it is not clear that the intermittent, short-term,
and relatively small changes in annual average ozone concentrations resulting from this rule are
likely to measurably change long-term risks of asthma. The McDonnell et al. study is a
prospective cohort analysis, measuring the association between long-term exposure to ambient
concentrations of ozone and development of chronic asthma in adults. The study found a
statistically significant effect for adult males, but none for adult females.
Some commentors have raised questions about the statistical validity of the associations
found in this study and the appropriateness of transferring the estimated C-R function from the
study populations (white, non-Hispanic males) to other male populations (i.e. African-American
males). Some of these concerns include 1) no significant association was observed for female
study participants also exposed to ozone; 2) the estimated C-R function is based on a cross-
sectional comparison of ozone levels, rather than incorporating information on ozone levels over
time; 3) information on the accuracy of self-reported incidence of chronic asthma was collected
but not used in estimating the C-R function; 4) the study may not be representative of the general
population because it included only those individuals living 10 years or longer within 5 miles of
their residence at the time of the study; and 5) the study had a significant number of study
participants drop out, either through death, loss of contact, or failure to provide complete or
consistent information.
EPA believes that while these issues may result in increased uncertainty about this effect,
none can be identified with a specific directional bias in the estimates. In addition, the study has
been reviewed by the SAB and has been specifically recommended for inclusion in benefits
analyses of changes in ozone concentrations (EPA-SAB-COUNCIL-ADV-00-001, 1999). EPA
also believes it to be appropriate to apply the C-R function to all adult males over age 27 because
no evidence exists to suggest that non-white adult males have a lower responsiveness to air-
pollution. For other health effects such as shortness of breath, where the study population was
limited to a specific group potentially more sensitive to air pollution than the general population
(Ostro, et al., 1995), EPA has applied the C-R function only to the limited population. EPA
recognizes the need for further investigation by the scientific community to confirm the statistical
association identified in the McDonnell et al. study.
f. Chronic Asthma: Valuation
Similar to the valuation of chronic bronchitis, WTP to avoid chronic asthma is presented
as the net present value of what would potentially be a stream of costs and lower well-being
incurred over a lifetime. Estimates of WTP to avoid asthma are provided in two studies, one by
Blumenschein and Johannesson (1998) and one by O'Conor and Blomquist (1997). Both
studies use the contingent valuation method to solicit annual WTP estimates from individuals
VII-47
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
who have been diagnosed as asthmatics. The central estimate of lifetime WTP to avoid a case of
chronic asthma among adult males, approximately $31,000, is the average of the present
discounted value from the two studies. Details of the derivation of this central estimate from the
two studies is provided in the benefits TSD for this RIA (Abt Associates, 1999).
g. Hospital Admissions: Quantification
There is a wealth of epidemiological information on the relationship between air pollution
and hospital admissions for various respiratory and cardiovascular diseases; in addition, some
studies have examined the relationship between air pollution and emergency room (ER) visits.
Because most emergency room visits do not result in an admission to the hospital (the majority
of people going to the ER are treated and return home) we treat hospital admissions and ER visits
separately, taking account of the fraction of ER visits that do get admitted to the hospital.
Hospital admissions require the patient to be examined by a physician, and on average
may represent more serious incidents than ER visits. The two main groups of hospital admissions
estimated in this analysis are respiratory admissions and cardiovascular admissions. There is not
much evidence linking ozone or PM with other types of hospital admissions. The only type of
ER visits that have been linked to ozone and PM in the U.S. or Canada are asthma-related visits.
To estimate the number of hospital admissions for respiratory illness, cardiovascular illness, and
asthma ER visits, we pool the incidence estimates from a variety of U.S. and Canadian studies,
using a random effects weighting procedure20. Details of the pooling procedure and a complete
listing of the hospital admission studies used in our estimates can be found in the benefits TSD
for this RIA (Abt Associates, 1999).
h. Hospital Admissions: Valuation
An individual's WTP to avoid a hospital admission will include, at a minimum, the
amount of money he or she pays for medical expenses (i.e., payment towards the hospital charge
and the associated physician charge) and the loss in earnings. In addition, an individual is likely
to be willing to pay some amount to avoid the pain and suffering associated with the illness itself.
20Because we are estimating ER visits as well as hospital admissions for asthma, we must avoid counting
twice the ER visits for asthma that are subsequently admitted to the hospital. To avoid double-counting, the
baseline incidence rate for emergency room visits is adjusted by subtracting the percentage of patients that are
admitted into the hospital. The reported incidence rates suggest that ER visits for asthma occur 2.7 times as
frequently as hospital admissions for asthma. The baseline incidence of asthma ER visits is therefore taken to be
2.7 times the baseline incidence of hospital admissions for asthma. To avoid double-counting, however, only 63%
of the resulting change in asthma ER visits associated with a given change in pollutant concentrations is counted in
the ER visit incidence change.
VII-48
-------
Chapter VII: Benefit-Cost Analysis
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.
In the absence of estimates of WTP to avoid hospital admissions for specific illnesses,
estimates of total cost-of-illness (COI) are typically used as conservative estimates. These
estimates are biased downward because they do not include the value of avoiding the illness
itself. 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. Consistent with the guidance offered by the EPA Science
Advisory Board, the COI values used in this benefits analysis will not be adjusted to better reflect
the total WTP (EPA-SAB-COUNCIL-ADV-98-003, 1998).
For the valuation of avoided hospital admissions, the current literature provides well-
developed and detailed cost estimates of hospitalization by health effect or illness. Using illness-
specific estimates of avoided medical costs and avoided costs of lost work-time that Elixhauser
(1993) developed, we construct COI estimates specific to the suite of health effects defined by
each C-R function. For example, we use twelve distinct C-R functions to quantify the expected
change in respiratory admissions. Consequently in this analysis, we develop twelve separate COI
estimates, each reflecting the unique composition of health effects considered in the individual
studies. Details of the derivation of the values of avoided hospital admissions for respiratory and
cardiovascular illnesses and asthma-related ER visits are provided in the benefits TSD for this
RIA (Abt Associates, 1999).
i. Other Health Effects: Quantification
As indicated in Table VII-1, in addition to mortality, chronic illness, and hospital
admissions, there are a number of acute health effects not requiring hospitalization that are
associated with exposure to ambient levels of ozone and PM. The sources for the C-R functions
used to quantify these effects are described below. A more complete description of these
estimates is provided in the benefits TSD for this RIA (Abt Associates, 1999).
Around five percent of U.S. children between ages five and seventeen experience
episodes of acute bronchitis annually (Adams, et al., 1995). Acute bronchitis is characterized by
coughing, chest discomfort, and extreme tiredness. Incidences of acute bronchitis in children
between the ages of five and seventeen are estimated using a C-R function developed from
Dockeryetal. (1996).
Incidences of lower respiratory symptoms (i.e., wheezing, deep cough) in children aged
seven to fourteen are estimated using a C-R function developed from Schwartz et al. (1994).
Because asthmatics have greater sensitivity to stimuli (including air pollution), children with
VII-49
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
asthma can be more susceptible to a variety of upper respiratory symptoms (i.e., runny or stuffy
nose; wet cough; and burning, aching, or red eyes). Incidences of upper respiratory symptoms in
asthmatic children aged nine to eleven are estimated using a C-R function developed from Pope
etal. (1991).
Health effects from air pollution can also result in missed days of work (either from
personal symptoms or from caring for a sick family member). Work loss days are estimated
using a C-R function developed from Ostro (1987).
The endpoint minor restricted activity days (MRAD), which is also represented by the
occurrence of any of 19 acute respiratory symptoms as defined by Krupnick et al. (1990), is a
pooled estimate using estimates of C-R functions derived from Ostro and Rothschild (1989) and
Krupnick et al. (1990).
As noted above, asthma affects over seven percent of the U.S. population. Air pollution
is sometimes linked to development of asthma and occurrences of asthma symptoms
(McDonnell, et al, 1999; Ostro, et al., 1991; Whittemore and Korn, 1980). Incidences of
shortness of breath (in African American asthmatics21) are estimated using a C-R function
derived from Ostro, et al. (1995). Other asthma related symptoms are included in the incidences
of MRAD and any of 19 acute respiratory symptoms. Inclusion of separate estimates for these
endpoints would result in double-counting of these benefits. Supplemental calculations for
separate asthma only endpoints are included in Appendix VII-A.
In addition to the health effects discussed above, human exposure to PM and ozone is
believed to be linked to health effects such as ozone-related premature mortality (Ito and
Thurston, 1996; Samet, et al. 1997), PM-related infant mortality (Woodruff, et al., 1997), cancer
(U.S. EPA, 1996b), increased emergency room visits for non-asthma respiratory causes (U.S.
EPA, 1996a; 1996b), impaired airway responsiveness (U.S. EPA, 1996a), increased susceptibility
to respiratory infection (U.S. EPA, 1996a), acute inflammation and respiratory cell damage (U.S.
EPA, 1996a), premature aging of the lungs and chronic respiratory damage (U.S. EPA, 1996a;
1996b). An improvement in ambient PM and ozone air quality may reduce the number of
incidences within each effect category that the U.S. population would experience. Although
these health effects are believed 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 lends a downward bias to the monetized benefits presented in this analysis.
Another category of potential effects that may change in response to ozone strategies
"Shortness of breath due to PM exposure is not necessarily limited to African-American asthmatics.
However, the Ostro, et al. study was based on a sample of African-American children, who may be more sensitive
to air pollution than the general population so we chose not to extrapolate the findings to the general population.
VII-50
-------
Chapter VII: Benefit-Cost Analysis
results from the shielding provided by ozone against the harmful effects of ultraviolet radiation
(UV-B) derived from the sun. The great majority of this shielding results from naturally
occurring ozone in the stratosphere, but the 10% of total "column"ozone present in the
troposphere also contributes (NAS, 1991). A variable portion of this tropospheric fraction of
UV-B shielding is derived from ground level or "smog" ozone related to anthropogenic air
pollution. Therefore, strategies that reduce ground level ozone will, in some small measure,
increase exposure to UV-B from the sun.
While it is possible to provide quantitative estimates of benefits associated with globally
based strategies to restore the far larger and more spatially uniform stratospheric ozone layer, the
changes in UV-B exposures associated with ground level ozone reduction strategies are much
more complicated and uncertain. Smog ozone strategies, such as mobile source controls, are
focused on decreasing peak ground level ozone concentrations, and it is reasonable to conclude
that they produce a far more complex and heterogeneous spatial and temporal pattern of ozone
concentration and UV-B exposure changes than do stratospheric ozone protection programs. In
addition, the changes in long-term total column ozone concentrations are far smaller from
ground-level programs. To properly estimate the change in exposure and impacts, it would be
necessary to match the spatial and temporal distribution of the changes in ground-level ozone to
the spatial and temporal distribution of exposure to ground level ozone and sunlight. More
importantly, it is long-term exposure to UV-B that is associated with effects. Intermittent, short-
term, and relatively small changes in ground-level ozone and UV-B are not likely to measurably
change long-term risks of these adverse effects.
For all of these reasons, we were unable to provide reliable estimates of the changes in
UV-B shielding associated with ground-level ozone changes. This inability lends an upward bias
to the net monetized benefits presented in this analysis. It is likely that the adverse health effects
associated with increases in UV-b exposure from decreased tropospheric ozone will, however, be
relatively small because 1) the expected long-term ozone change resulting from this rule is small
relative to total anthropogenic tropospheric ozone, which in turn is small in comparison to total
column natural stratospheric and tropospheric ozone; 2) air quality management strategies are
focused on decreasing peak ozone concentrations and thus may change exposures over limited
areas for limited times, 3) people often receive peak exposures to UV-B in coastal areas where
sea or lake breezes reduce ground level pollution concentrations regardless of strategy, and 4)
ozone concentration changes are greatest in urban areas and areas immediately downwind of
urban areas. In these areas, people are more likely to spend most of their time indoors or in the
shade of buildings, trees or vehicles.
VII-51
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
j. Other Health Effects: Valuation
The valuation of a specific short-term morbidity endpoint is generally estimated by
representing the illness as a cluster of acute symptoms. For each symptom, the WTP is
calculated. These values, in turn, are aggregated to arrive at the WTP to avoid a specific short
term condition. For example, the endpoint lower respiratory symptoms (LRS) is represented by
two or more of the following symptoms: runny or stuffy nose; coughing; and eye irritation. The
WTP to avoid one day of LRS is the sum of values associated with these symptoms. The
primary advantage of this approach is that is provides some flexibility in constructing estimates
to represent a variety of health effects.
Valuation estimates for individual minor health effects are listed in Table 11-10.
Derivation of the individual valuation estimates is provided in the benefits TSD for this RIA.
Mean estimates range from $5.30 for an avoided incidence of shortness of breath to $45 for an
avoided incidence of acute bronchitis. The value of work loss days varies depending on the
location of an affected population. Using the median daily wage, the value of a work loss day is
$83.
k. Lost Worker Productivity: Quantification and Valuation
While not technically a health effect, lost worker productivity related to pollution
exposure is presumably linked to reductions in the physical capabilities of workers in outdoor
jobs. The value of lost worker productivity due to ozone exposure is directly estimated based on
a study of California citrus workers (Crocker and Horst, 1981 and U.S. EPA, 1994). 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 (or the percentage
change in income for a one percent change in ambient ozone concentration). The reported
elasticity translates a ten percent reduction in ozone to a 1.4 percent increase in income.
1. Estimated Reductions in Incidences of Health Endpoints and Associated
Monetary Values
Applying the C-R and valuation functions described above to the estimated changes in
ozone and PM yields estimates of the number of avoided incidences (i.e. premature mortalities,
cases, admissions, etc.) and the associated monetary values for those avoided incidences. These
estimates are presented in Table VII-12. All of the monetary benefits are in constant 1997
dollars.
Not all known PM and ozone related health effects could be quantified or monetized.
These unmonetized benefits are indicated by place holders, labeled Bx and B2. Unquantified
VII-52
-------
Chapter VII: Benefit-Cost Analysis
physical effects are indicated by U1 and U2. The estimate of total monetized health benefits is
thus equal to the subset of monetized PM and ozone related health benefits plus BH, the sum of
the unmonetized health benefits.
The largest monetized health benefit is associated with reductions in the risk of premature
mortality. The next largest benefit is for chronic bronchitis reductions, although this value is
more than an order of magnitude lower than for premature mortality. Minor restricted activity
days, work loss days, and worker productivity account for the majority of the remaining benefits.
The remaining categories account for less than $10 million each, however, they represent a large
number of avoided incidences affecting many individuals.
Alternative calculations for premature mortality incidences and valuation are presented in
Table VII-18. An alternative calculation is also provided in that table for chronic bronchitis
incidences.
VII-53
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table VII-12. Estimated Annual Health Benefits Associated With Air Quality Changes
Resulting from the Tier 2/Gasoline Sulfur Rule in 2030
Endpoint
PM-related Endpoints*
Premature mortality1" (adults, 30 and over)
Chronic bronchitis
Hospital Admissions from Respiratory Causes
Hospital Admissions from Cardiovascular Causes
Emergency Room Visits for Asthma
Acute bronchitis (children, 8-12)
Lower respiratory symptoms (LRS) (children, 7-14)
Upper respiratory symptoms (URS) (asthmatic children, 9-1 1)
Shortness of breath (African American asthmatics, 7-12)
Work loss days (WLD) (adults, 18-65)
Minor restricted activity days (MRAD)/Acute respiratory symptoms
Other PM-related health effects6
Ozone-related Endpoints
Chronic asthma (adult males, 27 and over)
Hospital Admissions from Respiratory Causes
Hospital Admissions from Cardiovascular Causes
Emergency Room Visits for Asthma
Minor restricted activity days (MRAD)/Acute respiratory symptoms
Decreased worker productivity (adult working population)
Other ozone-related health effects6
CO-related health effects6
HAPS-related health effects6
Monetized Total Health-related Benefits1
Avoided Incidence0
(cases/year)
4,300
2,300
1,200
500
900
7,900
87,100
86,500
17,400
682,900
3,628,500
u,
400
1,000
300
400
2,226,500
—
U2
U3
U4
—
Monetary Benefits'1
(millions 1997$)
$23,380
$730
$10
$10
$<1
$<1
$<5
$<5
$<1
$70
$170
B!
$10
$10
$<5
$<1
$100
$140
B2
B3
B4
$24,630+BH
"PM reductions are due to reductions in NOx and SO2 resulting from the Tier 2/Gasoline Sulfur rale.
b Premature mortality associated with ozone is not separately included in this analysis. It is assumed that the Pope, et al.
premature mortality captures both PM mortality benefits and any mortality benefits associated with other air pollutants.
estimated value assumes the 5 year distributed lag structure described earlier.
0 Incidences are rounded to the nearest 100. d Dollar values are rounded to the nearest 10 million.
' A detailed listing of unquantified PM, ozone, CO, and HAPS related heatlh effects is provided in Table VII-1.
f Bg is equal to the sum of all unmonetized categories, i.e. B[+B2
C-R function for
Also note that the
VII-54
-------
Chapter VII: Benefit-Cost Analysis
E. Assessment of Human Welfare Benefits
Particulate matter and ozone have numerous documented effects on environmental
quality that affect human welfare. These welfare effects include direct damages to property,
either through impacts on material structures or by soiling of surfaces, direct economic damages
in the form of lost productivity of crops and trees, indirect damages through alteration of
ecosystem functions, and indirect economic damages through the loss in value of recreational
experiences or the existence value of important resources. EPA's criteria documents for ozone
and PM list numerous physical and ecological effects known to be linked to ambient
concentrations of these pollutants (U.S. EPA, 1996a, 1996b). This section describes individual
effects and how we quantify and monetize them. These effects include changes in crop yields,
visibility, and nitrogen deposition to estuaries.
In section 1, we describe how we quantify and value changes in visibility, both in federal
Class I areas (national parks and wilderness areas) and in the areas where people live and work.
In section 2, we describe how we value the benefits of increased agricultural and commercial
forest yields resulting from decreased levels of ambient ozone. In section 3, we describe the
damage to materials caused by particulate matter. In section 4, we discuss the effects of nitrogen
deposition on ecosystems (especially estuarine ecosystems) and describe how we quantify
changes in nitrogen loadings. Finally, in section 6, we summarize the monetized estimates for
welfare effects.
1. Visibility Benefits
Changes in the level of ambient parti culate matter caused by the final Tier 2/Gasoline
Sulfur rule will change the level of visibility in much of the U.S. Visibility directly affects
people's enjoyment of a variety of daily activities. Individuals value visibility both in the places
they live and work, in the places they travel to for recreational purposes, and at sites of unique
public value, such as the Grand Canyon. This section discusses the measurement of the
economic benefits of visibility.
It is difficult to quantitatively define a visibility endpoint that can be used for valuation.
Increases in PM concentrations cause increases in light extinction. Light extinction is a measure
of how much the components of the atmosphere absorb light. More light absorption means that
the clarity of visual images and visual range is reduced, ceteris paribus. Light absorption is a
variable that can be accurately measured. Sisler (1996) created a unitless measure of visibility
based directly on the degree of measured light absorption called the deciview. Deciviews are
standardized for a reference distance in such a way that one deciview corresponds to a change of
about 10 percent in available light. Sisler characterized a change in light extinction of one
deciview as "a small but perceptible scenic change under many circumstances." Air quality
models were used to predict the change in visibility, measured in deciviews, of the areas affected
VII-55
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
by the final Tier 2/Gasoline Sulfur rule22.
EPA considers benefits from two categories of visibility changes: residential visibility
and recreational visibility. In both cases economic benefits are believed to consist of both use
values and non-use values. The use values include the aesthetic benefits of better visibility,
improved road and air safety, and enhanced recreation in activities like hunting and
birdwatching. The non-use values are based on people's beliefs that the environment ought to
exist free of human-induced haze. Non-use values may be a more important component of value
for recreational areas, particularly national parks and monuments.
Residential visibility benefits are those that occur from visibility changes in urban,
suburban, and rural areas, and also in recreational areas not listed as federal Class I areas23.
Recreational visibility improvements are those that occur specifically in federal Class I areas. A
key distinction is that only those people living in residential areas are assumed to receive benefits
from residential visibility, while all households in the U.S. are assumed to derive some benefit
from improvements in Class I areas. Values are assumed to be higher if the Class I area is
located close to their home.24
The results of air quality modeling of the Tier 2/Gasoline Sulfur rule show consistent
improvements in visibility in all areas of the country. The mean improvement across all U.S.
counties was 0.24 deciviews. The biggest improvements in visibility were most often found in
heavily populated urban areas. Of the central counties of metropolitan areas with more than one
million in population in 1993, 21 percent show an improvement of 0.5 deciviews or more. For
suburban counties of these same regions, 11 percent are predicted to have visibility
improvements of 0.5 deciview or more. For the 21 percent of metropolitan areas showing an
improvement of 0.5 deciviews or more, the baseline visibility is 25.3. For the 11 percent of
suburban counties in these regions, the baseline visibility is 23.3. And, baseline visibility in the
10 percent of counties with the largest improvements in visibility (baseline = 24.4 deciviews) is
much worse than baseline visibility in those counties with no change in visibility (baseline =18.1
deciviews). This suggests that the Tier 2/Gasoline Sulfur rule has the potential to provide large
improvements in visibility in those areas with the worst baseline visibility conditions.
22A 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. Some of the average regional changes in visibility
are 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.
23 The Clean Air Act designates 156 national parks and wilderness areas as Class I areas for visibility
protection.
24 For details of the visibility estimates discussed in this chapter, please refer to the benefits technical
support document for this RIA (Abt Associates).
VII-56
-------
Chapter VII: Benefit-Cost Analysis
Only two existing studies provide defensible monetary estimates of the value of visibility
changes. One is a study on residential visibility conducted in 1990 (McClelland, et. al., 1993) and
the other is a 1988 survey on recreational visibility value (Chestnut and Rowe, 1990a, 1990b).
Both utilize the contingent valuation method. There has been a great deal of controversy and
significant development of both theoretical and empirical knowledge about how to conduct CVM
surveys in the past decade. In EPA's judgment, the Chestnut and Rowe study contains many of
the elements of a valid CVM study and is sufficiently reliable to serve as the basis for monetary
estimates of the benefits of visibility changes in recreational areas25. This study serves as an
essential input to our estimates of the benefits of recreational visibility improvements in the
primary benefits estimates. Based on SAB advice (EPA-SAB-COUNCIL-ADV-00-002, 1999),
EPA has designated the McClelland et al. study as significantly less reliable for regulatory
benefit-cost analysis, but it does provide useful estimates on the order of magnitude of residential
visibility benefits. Residential visibility benefits are therefore only included as an alternative
calculation in Table VII-18. The methodology for this alternative calculation, explained below,
is similar to the procedure for recreational benefits.
The Chestnut and Rowe study measured the demand for visibility in Class I areas
managed by the National Park Service (NFS) in three broad regions of the country: California,
the Southwest, and the Southeast. Respondents in five states were asked about their willingness
to pay to protect national parks or NPS-managed wilderness areas within a particular region.
The survey used photographs reflecting different visibility levels in the specified recreational
areas. The visibility levels in these photographs were later converted to deciviews for the current
analysis. The survey data collected were used to estimate a willingness-to-pay equation for
improved visibility. In addition to the visibility change variable, the estimating equation also
included household income as an explanatory variable.
The Chestnut and Rowe study did not measure values for visibility improvement in Class
I areas outside the three regions. Their study covered 86 of the 156 Class I areas in the U.S. We
can infer the value of visibility changes in the other Class I areas by transferring values of
visibility changes at Class I areas in the study regions. However, these values are not as
defensible and are thus presented only as an alternative calculation in Table VII-18. A complete
description of the benefits transfer method used to infer values for visibility changes in Class I
areas outside the study regions is provided in the benefits TSD for this RIA (Abt Associates,
1999).
The estimated relationship from the Chestnut and Rowe study is only directly applicable
25An SAB advisory letter (EPA-SAB-COUNCIL-ADV-00-002, 1999) indicates that "many members of the
Council believe that the Chestnut and Rowe study is the best available," however, the council did not formally
approve use of these estimates because of concerns about the peer-reviewed status of the study. EPA believes the
study has received adequate review and has been cited in numerous peer-reviewed publications (Chestnut and
Dennis, 1997).
VII-57
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
to the populations represented by survey respondents. EPA used benefits transfer methodology
to extrapolate these results to the population affected by the final Tier 2/Gasoline Sulfur rule. A
general willingness to pay equation for improved visibility (measured in deciviews) was
developed as a function of the baseline level of visibility, the magnitude of the visibility
improvement, and household income. The behavioral parameters of this equation were taken
from analysis of the Chestnut and Rowe data. These parameters were used to calibrate WTP for
the visibility changes resulting from the final Tier 2/Gasoline Sulfur rule. The method for
developing calibrated WTP functions is based on the approach developed by Smith, et al. (1999).
Available evidence indicates that households are willing to pay more for a given visibility
improvement as their income increases (Chestnut 1997). The benefits estimates here incorporate
Chestnut's estimate that a one percent increase in income is associated with a 0.9 percent
increase in WTP for a given change in visibility.
Using the methodology outlined above, EPA estimates that the total willingness to pay
for the visibility improvements in Class I areas brought about by the final Tier 2/Gasoline Sulfur
rule is $371 million. This value includes the value to households living in the same state as the
Class I area as well as values for all households in the U.S. living outside the state containing the
Class I area. A complete presentation of this method can be found in the benefits TSD for this
RIA (Abt Associates, 1999).
For the alternative calculation for residential visibility, the McClelland study's results
were used to calculate the parameter for the effect of deciview changes on WTP. The WTP
equation was then run for the population affected by the final Tier 2/Gasoline Sulfur rule. The
results indicate that improvements to residential visibility provide an economic benefit of $581
million dollars for the continental U.S.26 A complete presentation of this method can be found in
the benefits TSD for this RIA (Abt Associates, 1999).
One major source of uncertainty for the visibility benefit estimate is the benefits transfer
process used. Judgments used to choose the functional form and key parameters of the
estimating equation for willingness to pay for the affected population could have significant
effects on the size of the estimates. Assumptions about how individuals respond to changes in
visibility that are either very small, or outside the range covered in the Chestnut and Rowe study,
could also affect the results.
26 The McClelland et al. (1990) study examined visibility changes in two Eastern cities, Chicago and
Atlanta. Transferring these values to residential visibility changes in the Western U.S. may introduce greater
uncertainty than transferring the values to other Eastern cities. As such, an additional alternate calculation showing
the value of residential visibility just for the Eastern U.S. is included in Table VII-18.
VII-58
-------
Chapter VII: Benefit-Cost Analysis
2. Agricultural and Forestry Benefits
Reduced levels of ground-level ozone resulting from the final Tier 2/Gasoline Sulfur rule
will have generally beneficial results on agricultural crop yields and commercial forest growth.
Well-developed techniques exist to provide monetary estimates of these benefits to agricultural
producers and to consumers. These techniques use models of farmers' planting decisions, yield
response functions, and agricultural supply and demand. The resulting welfare measures are
based on predicted changes in market prices and production costs.
The economic value associated with varying levels of yield loss for ozone-sensitive
commodity crops is analyzed using the AGSEVI® agricultural benefits model (Taylor et al., 1993).
AGSEVI® 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 final Tier 2/Gasoline Sulfur rule) that affect commodity crop yields or production costs27.
The benefits TSD for this RIA also provides further details on AGSEVI® (Abt Associates, 1999).
The measure of benefits calculated by the model is the net change in consumer and
producer 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 consumer and producer surplus on a crop-by-crop basis28. 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 final Tier 2/Gasoline Sulfur rule.
The model employs biological exposure-response information derived from controlled
experiments conducted by the National Crop Loss Assessment Network (NCLAN, 1996). For
the purpose of our analysis, we analyze changes for the six most economically significant crops
for which dose-response functions are available: corn, cotton, peanuts, sorghum, soybean, and
winter wheat.29 For some crops there are multiple dose-response functions, some more sensitive
to ozone and some less. Our primary estimate assumes that crops are evenly mixed between
relatively sensitive and relatively insensitive varieties. The primary estimate of the net change in
economic surplus resulting from changes in ozone associated with the Tier 2/Gasoline Sulfur rule
27AGSIM° is designed to forecast agricultural supply and demand out to 2010. We were not able to adapt
the model to forecast out to 2030. Instead, we apply percentage increases in yields from decreased ambient ozone
levels in 2030 to 2010 yield levels, and input these into a agricultural sector model held at 2010 levels of demand
and supply. It is uncertain what impact this assumption will have on net changes in surplus.
28 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.
29 The total value for these crops in 1997 was $57 billion.
VII-59
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
is $217 million.
Similar models exist for forest products. Ozone also has been shown conclusively to
cause discernible injury to forest trees (Fox and Mickler, 1996). Once the effects of changes in
ozone concentrations on tree growth are predicted, econometric models of forest product supply
and demand can be used to estimate changes in prices, producer profits and consumer surplus.
Our analysis does not attempt to quantify commercial forestry benefits due to difficulties in
obtaining C-R functions relating ozone exposure and tree growth. An additional welfare benefit
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 to reliably establish 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.
Urban ornamentals represent an additional vegetation category likely to experience some
degree of negative 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. This is therefore a potentially important welfare effects category. However, information
and 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.
The final Tier 2/Gasoline Sulfur rule, by reducing NOX emissions, will also reduce
nitrogen deposition on agricultural land and forests. There is some evidence that nitrogen
deposition may have positive effects on agricultural output through passive fertilization. Holding
all other factors constant, farmers' use of purchased fertilizers or manure may increase as
deposited nitrogen is reduced. Estimates of the potential value of this possible increase in the use
of purchased fertilizers are not available, but it is likely that the overall value is very small
relative to other health and welfare effects. The share of nitrogen requirements provided by this
deposition is small, and the marginal cost of providing this nitrogen from alternative sources is
quite low. In some areas, agricultural lands suffer from nitrogen over-saturation due to an
abundance of on-farm nitrogen production, primarily from animal manure. In these areas,
reductions in atmospheric deposition of nitrogen from PM represent additional agricultural
benefits.
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,
VII-60
-------
Chapter VII: Benefit-Cost Analysis
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.
3. Benefits from Reductions in Materials Damage
The final Tier 2/Gasoline Sulfur rule is expected to produce economic benefits in the
form of reduced materials damage. There are two important categories of these benefits.
Household soiling refers to the accumulation of dirt, dust, and ash on exposed surfaces. Criteria
pollutants also have corrosive effects on commercial/industrial buildings and structures of
cultural and historical significance. The effects on historic buildings and outdoor works of art
are of particular concern because of the uniqueness and irreplaceability of many of these objects.
Previous EPA benefit analyses including that for the Tier 2 Proposal RIA, have been able
to provide quantitative estimates of household soiling damage. Following an SAB
recommendation (EPA-SAB-Council-ADV-003, 1998), EPA has determined that the existing
data (based on consumer expenditures from the early 1970's) is too out of date to provide a
reliable enough estimate of current household soiling damages. However, a calculation is made
for inclusion in the alternative calculations table (Table VII-18).
EPA is unable to estimate any benefits to commercial and industrial entities from reduced
materials damage. Nor is EPA able to estimate the benefits of reductions in PM-related damage
to historic buildings and outdoor works of art. Existing studies of damage to this latter category
in Sweden (Grosclaude and Soguel, 1994) indicate that these benefits could be an order of
magnitude larger than household soiling benefits.
4. Benefits from Reduced Ecosystem Damage
The effects of air pollution on the health and stability of ecosystems are potentially very
important, but are at present poorly understood and difficult to measure. The reductions in NOX
caused by the final rule could produce significant benefits. Excess nutrient loads, especially 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
VII-61
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
(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 (Haire et al., 1992).
Reductions in nitrogen loadings are estimated for twelve eastern estuaries (including two
on the Gulf Coast). These estimated reductions are described earlier in this Chapter. Four of
these estuaries have established consensus goals for reductions in annual nitrogen loads,
indicating an intention of reaching these goals through implementation of controls on nitrogen
sources. These four estuaries and their reduction goals are listed in Table VII-13.
Table VII-13. Reduction Goals and Nitrogen Loads to Selected Eastern Estuaries
(tons per year)
Estuary
Albemarle/Pamlico Sound
Chesapeake Bay
Long Island Sound
Tampa Bay
Total Nitrogen Loadings
25,300
185,000
53,700
3,900
Nitrogen Loadings from
Atmospheric Deposition
11,000
49,500
13,200
2,100
Overall Reduction Goal
7,600
35,600
31,460
100
Source: U.S. EPA, 1998
Estimated reductions in deposition of atmospheric nitrogen to these four estuaries are
listed in Table VII-14, along with the percentage of the reduction goal accounted for by these
reductions. These figures suggest that the reductions in nitrogen deposition resulting from the
final Tier 2/Gasoline Sulfur rule will provide significant progress towards meeting nitrogen
reduction goals in several of these estuaries.
VII-62
-------
Chapter VII: Benefit-Cost Analysis
Table VII-14. Estimated Annual Reductions in Nitrogen Loadings in Selected Eastern
Estuaries for the Final Tier 2/Gasoline Sulfur Rule in 2030
(tons per year)
Estuary
Albemarle/Pamlico Sound
Chesapeake Bay
Long Island Sound
Tampa Baya
Change in Nitrogen Loadings
-2,013
-3,080
-1,144
-484
% of Estuary Nitrogen Reduction
Goal
26.5%
8.7%
3.6%
over 100%
" Tampa Bay had a very low nitrogen loadings reduction goal. As such, the Tier 2 rule provides more reductions than are necessary to achieve
the stated goal.
Direct C-R functions relating changes in nitrogen loadings to changes in estuarine
benefits are not available. The preferred WTP 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
value of changes in water quality exist at present, calculation of a WTP measure is not possible.
As stated earlier, an alternative is to use an avoided cost approach to estimate the welfare effects
of PM on estuarine ecosystems. The use of the avoided cost approach to establish the value of a
reduction in nitrogen deposition is problematic if there is not a direct link between reductions in
air deposited nitrogen and the abandonment of a costly regulatory program. However, there are
currently no readily available alternatives to this approach.
Based on the advice of the EPA Science Advisory Board, we use the avoided cost
approach only to derive an alternative calculation of the value of reductions in atmospheric
nitrogen loadings to estuaries (EPA-SAB-COUNCIL-ADV-00-002, 1999). The SAB believes
that the avoided cost approach for nitrogen loadings is valid only if the state and local
governments have established firm pollution reduction targets, and that displaced costs measured
in the study represent measures not taken because of the CAAA (EPA-SAB-COUNCIL-ADV-
00-002, 1999). Because the nitrate reduction targets in the studied estuaries are not firm targets,
and there is not assurance that planned measures would be undertaken in the absence of the
CAAA, we are currently unable to provide a meaningful primary estimate. However, the
avoided cost estimate is presented in the table of alternative calculations (Table VII-18).
If better models of ecological effects can be defined, EPA believes that progress can be
made in estimating WTP measures for ecosystem functions. These estimates would be superior
to avoided cost estimates in placing economic values on the welfare changes associated with air
pollution damage to ecosystem health. For example, if nitrogen or sulfate loadings can be linked
to measurable and definable changes in fish populations or definable indexes of biodiversity,
VII-63
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
then CVM studies can be designed to elicit individuals' willingness to pay for changes in these
effects. This is an important area for further research and analysis, and will require close
collaboration among air quality modelers, natural scientists, and economists.
5. Estimated Values for Welfare Endpoints
Applying the valuation methods described above to the estimated changes in ozone and
PM in 2030 yields estimates of the value of changes in visibility and agricultural yields. These
estimates are presented in Table VII-15. All of the monetary benefits are in constant 1997
dollars.
We are unable to provide primary monetized estimates of residential visibility, household
soiling, materials damage, nitrogen deposition and commercial forestry benefits, in addition to
the other welfare effects listed in Table VII-1. These unmonetized benefits are indicated by
placeholders, labeled B3 to B9. The estimate of total monetized welfare benefits is thus equal to
the subset of monetized welfare benefits plus Bw, the sum of the unmonetized welfare benefits.
Total monetized welfare related benefits are around $590 million. Monetized welfare
benefits are roughly one fortieth the magnitude of monetized health benefits. However, due to
the difficulty in quantifying and monetizing welfare benefits, a higher proportion of welfare
benefits
VII-64
-------
Chapter VII: Benefit-Cost Analysis
Table VII-15. Estimated Annual Monetary Values for Welfare Effects Associated With
Improved Air Quality Resulting from the Tier 2/Gasoline Sulfur Rule in 2030
Endpoint
PM-related Endpoints
Recreational Visibility (86 Class I areas in California, the Southeast and the Southwest)
Residential Visibility
Household Soiling
Materials Damage
Nitrogen Deposition to Estuaries
Other PM-related welfare effects1"
Ozone-related Endpoints
Commercial Agricultural Benefits (6 major crops)
Commercial Forestry Benefits
Other ozone-related welfare effects'1
CO-related welfare effects1"
HAPS-related welfare effects*
Total Monetized Welfare-related Benefitsc
Monetary Benefits
(millions 1997$)a
$370
B5
B6
B7
B8
B9
$220
BIO
Bn
B12
B13
$590+6^,
"Rounded to the nearest 10 million.
b A detailed listing of unquantified PM, ozone, CO, and HAPS related welfare effects is provided in Table VII-1.
0 Bw is equal to the sum of all unmonetized welfare categories, i.e. B5+B6+...+B13.
are not monetized. It is thus inappropriate to conclude that welfare benefits are unimportant just
by comparing the estimates of the monetized benefits.
Alternative calculations for recreational visibility, residential visibility, household soiling,
and nitrogen deposition are presented in Table VII-18 later in this chapter.
F.
Total Benefits
We provide our preferred estimate of benefits for each health and welfare endpoint and
the resulting preferred estimate of total benefits. To obtain this estimate, we aggregate 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 is the sum of the
VII-65
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
separate effects estimates. Total monetized benefits associated with the final Tier 2/Gasoline
Sulfur rule are listed in Table VD-16, along with a breakdown of benefits by endpoint. Note that
the value of endpoints known to be affected by ozone and/or PM that we are not able to monetize
are assigned a placeholder value, e.g. Bl3 B2, etc. Unquantified physical effects are indicated by a
U. The estimate of total benefits is thus the sum of the monetized benefits and a constant, B,
equal to the sum of the unmonetized benefits, B1+B2+...+Bn.
A comparison of the incidence column to the monetary benefits column reveals that there
is not always a close correspondence between the number of incidences avoided for a given
endpoint and the monetary value associated with that endpoint. This reflects the fact that many
of the less severe health effects, while more common, are valued at a lower level than the more
severe health effects.
Our preferred estimate of total monetized benefits for the final Tier 2/Gasoline Sulfur rule
is $25 billion, of which $23 billion is the benefits of reduced premature mortality risk from PM
exposure. Total monetized benefits are dominated by the benefits of reduced mortality risk.
Mortality related benefits account for over 90 percent of total monetized benefits followed by
chronic bronchitis (3 percent). However, the adoption of a value for the projected reduction in
the risk of premature mortality is the subject of continuing discussion within the economic and
public policy analysis community within and outside the Administration. In response to the
sensitivity on this issue, we provide estimates reflecting two alternative approaches. The first
approach — supported by some in the above community and preferred by EPA — uses a Value of
a Statistical Life (VSL) approach developed for the Clean Air Act Section 812 benefit-cost
studies. This VSL estimate of $5.9 million (1997$) was derived from a set of 26 studies
identified by EPA using criteria established in Viscusi (1992), as those most appropriate for
environmental policy analysis applications.
An alternative, age-adjusted approach is preferred by some others in the above
community both within and outside the Administration. This approach was also developed for
the Section 812 studies and addresses concerns with applying the VSL estimate -reflecting a
valuation derived mostly from labor market studies involving healthy working-age manual
laborers- to PM-related mortality risks that are primarily associated with older populations and
those with impaired health status. This alternative approach leads to an estimate of the value of a
statistical life year (VSLY), which is derived directly from the VSL estimate. It differs only in
incorporating an explicit assumption about the number of life years saved and an implicit
assumption that the valuation of each life year is not affected by age.30 The mean VSLY is
30 Specifically, the VSLY estimate is calculated by amortizing the $5.9 million mean VSL estimate over the 35 years of
life expectancy associated with subjects in the labor market studies. The resulting estimate, using a 5 percent discount rate, is
$360,000 per life-year saved in 1997 dollars. This annual average value of a life-year is then multiplied times the number of
years of remaining life expectancy for the affected population (in the case of PM-related premature mortality, the average number
of $ life-years saved is 14.
VII-66
-------
Chapter VII: Benefit-Cost Analysis
$360,000 (1997$); combining this number with a mean life expectancy of 14 years yields an age-
adjusted VSL of $3.6 million (1997$).
Both approaches are imperfect, and raise difficult methodological issues which are
discussed in depth in the recently published Section 812 Prospective Study, the draft EPA
Economic Guidelines, and the peer-review commentaries prepared in support of each of these
documents. For example, both methodologies embed assumptions (explicit or implicit) about
which there is little or no definitive scientific guidance. In particular, both methods adopt the
assumption that the risk versus dollars trade-offs revealed by available labor market studies are
applicable to the risk versus dollar trade-offs in an air pollution context.
EPA currently prefers the VSL approach because, essentially, the method reflects the
direct, application of what EPA considers to be the most reliable estimates for valuation of
premature mortality available in the current economic literature. While there are several
differences between the labor market studies EPA uses to derive a VSL estimate and the
particulate matter air pollution context addressed here, those differences in the affected
populations and the nature of the risks imply both upward and downward adjustments. For
example, adjusting for age differences may imply the need to adjust the $5.9 million VSL
downward as would adjusting for health differences, but the involuntary nature of air pollution-
related risks and the lower level of risk-aversion of the manual laborers in the labor market
studies may imply the need for upward adjustments. In the absence of a comprehensive and
balanced set of adjustment factors, EPA believes it is reasonable to continue to use the $5.9
million value while acknowledging the significant limitations and uncertainties in the available
literature. Furthermore, EPA prefers not to draw distinctions in the monetary value assigned to
the lives saved even if they differ in age, health status, socioeconomic status, gender or other
characteristic of the adult population.
Those who favor the alternative, age-adjusted approach (i.e. the VSLY approach)
emphasize that the value of a statistical life is not a single number relevant for all situations.
Indeed, the VSL estimate of $5.9 million (1997 dollars) is itself the central tendency of a number
of estimates of the VSL for some rather narrowly defined populations. When there are
significant differences between the population affected by a particular health risk and the
populations used in the labor market studies - as is the case here - they prefer to adjust the VSL
estimate to reflect those differences. While acknowledging that the VSLY approach provides an
admittedly crude adjustment (for age though not for other possible differences between the
populations), they point out that it has the advantage of yielding an estimate that is not
presumptively biased. Proponents of adjusting for age differences using the VSLY approach
fully concur that enormous uncertainty remains on both sides of this estimate - upwards as well
as downwards - and that the populations differ in ways other than age (and therefore life
expectancy). But rather than waiting for all relevant questions to be answered, they prefer a
process of refining estimates by incorporating new information and evidence as it becomes
VII-67
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
available.
The estimates of benefits for the final Tier 2/Gasoline Sulfur rule using the different
approaches for premature mortality valuation are presented in Table VII-17. The VSL approach
-the approach EPA prefers - yields a monetized benefit estimate of $25.5 billion. The
alternative, age-adjusted approach yields monetary benefits of $14 billion. The final Tier
2/Gasoline Sulfur rule is expected to affect populations in the entire continental U.S31. Given a
projected U.S. population in 2030 of 300 million, annual monetized per capita benefits (using
EPA's preferred approach for valuing reductions in premature mortality) are over $84 in 2030.
Assuming an average household size of 2.6 persons, this translates to over $218 per household in
2030.
31 The Tier 2/Gasoline Sulfur standards will not apply to vehicles in California, however, populations in
California are expected to receive some benefits from the Tier 2/Gasoline Sulfur standards due to reductions in
pollutants transported into the state from other regions.
VII-68
-------
Chapter VII: Benefit-Cost Analysis
Table VII-16. EPA Preferred Estimate of Annual Quantified and Monetized Benefits
Associated With Improved Air Quality Resulting from the Tier 2/Gasoline Sulfur Rule in
2030
Endpoint
Premature mortality ^ (adults, 30 and over)
Chronic asthma (adult males, 27 and over)
Chronic bronchitis
Hospital Admissions from Respiratory Causes
Hospital Admissions from Cardiovascular Causes
Emergency Room Visits for Asthma
Acute bronchitis (children, 8-12)
Lower respiratory symptoms (children, 7-14)
Upper respiratory symptoms (asthmatic children, 9-1 1)
Shortness of breath (African American asthmatics, 7-
12)
Work loss days (adults, 18-65)
Minor restricted activity days /Acute resp. symptoms
Other health effects4
Decreased worker productivity
Recreational visibility (86 Class I Areas)
Residential visibility
Household soiling damage
Materials damage
Nitrogen Deposition to Estuaries
Agricultural crop damage (6 crops)
Commercial forest damage
Other welfare effectsf
Pollutant
PMb
Ozone
PM
Ozone and PM
Ozone and PM
Ozone and PM
PM
PM
PM
PM
PM
Ozone and PM
Ozone, PM, CO, HAPS
Ozone
PM
PM
PM
PM
Nitrogen
Ozone
Ozone
Ozone, PM, CO, HAPS
Avoided
Incidence0'11
(cases/year)
4,300
400
2,300
2,200
800
1,200
7,900
87,100
86,500
17,400
682,900
5,855,000
U!+U2+U3+U4
—
—
—
—
—
—
—
—
—
Monetized Total8*
Monetary
Benefits"
(millions 1997$)
$23,380
$10
$730
$20
$10
$<1
$<1
$<5
$<5
$<1
$70
$270
B!+B2+B3+B4
$140
$370
B5
B6
By
B8
$220
B9
Bio+Bn+B12+B13
$25,220+B
* Premature mortality associated with ozone is not separately included in this analysis. It is assumed that the Pope, et al.
premature mortality captures both PM mortality benefits and any mortality benefits associated with other air pollutants.
valuation assumes the 5 year distributed lag structure described earlier.
b PM reductions are due to reductions in NOx and SO2 resulting from the Tier 2/Gasoline Sulfur rule.
0 Incidences are rounded to the nearest 100.
dThe U; are the incidences for the unquantified category i.
C-R function for
Also note that the
VII-69
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
' Dollar values are rounded to the nearest 10 million.
f A detailed listing of unquantified PM, ozone, CO, and HAPS related heatlh and welfare effects is provided in Table VII-1.
8 B is equal to the sum of all unmonetized categories, i.e. Bj+Bj+.^+B^.
h These estimates are based on the EPA preferred approach for valuing reductions in premature mortality, the VSL approach. This approach and
an alternative, age-adjusted approach - the VSLY approach - are discussed more fully in section F.
Table VII-17. Final Tier 2/Gasoline Sulfur Rule: 2030 Monetized Benefits Estimates for
Alternative Premature Mortality Valuation Approaches
(Billions of 1997 dollars)
Premature Mortality Valuation Approach
Value of statistical life (VSL)
($5.9 million per life saved)3
Value of statistical life-years (VSLY)
($360,000 per life-year saved, which implies $3.6 million per
life saved, based on the mean of 14 life-years saved)a>b
PM Mortality Benefits
$23.4
$11.9
Total Monetized
Benefits
$25.2 + B
$13.7 + B
a Premature mortality estimates are determined assuming a 5 year distributed lag, which applies 25 percent of the incidence in year 1 and 2, and
then 16.7 percent of the incidence in years 3, 4, and 5.
b The VSLY estimate is calculated by amortizing the $5.9 million mean VSL estimate over the 35 years of life expectancy associated with
subjects in the labor market studies used to obtain the VSL estimate. The resulting estimate, using a 5 percent discount rate, is $360,000 per
life-year saved in 1997 dollars. This approach is discussed more fully in section F above.
In addition to the preferred estimate, in Table VII-18 we present alternative calculations
representing how the value for individual endpoints or total benefits would change if we were to
make a different assumption about an element of the benefits analysis. For example, this table
can be used to answer questions like "What would total benefits be if we were to use the
Dockery, et al. C-R function to estimate avoided premature mortality?" This table provides
alternative calculations both for valuation issues (e.g. the correct value for a statistical life saved)
and for physical effects issues (e.g., how reversals in chronic illnesses are treated). This table is
not meant to be comprehensive. Rather, it reflects some of the key issues identified by EPA or
commentors as likely to have a significant impact on total benefits. Accompanying Table VII-18
is a brief discussion of each of the alternative calculations.
While Table VII-18 provides alternative calculations for specific alternative assumptions,
there are some parameters to which total benefits may be sensitive but for which no or limited
credible scientific information exists to determine plausible values. Sensitivity analyses for these
parameters are presented in Appendix VII-A. Issues examined in this appendix include
alternative specifications for the lag structure of PM related premature mortality and impacts of
assumed thresholds on the estimated incidence of avoided premature mortality. Also, this
appendix contains several illustrative endpoint calculations for which the scientific uncertainty is
too great to provide a reasonable estimate for which inclusion would lead to double-counting of
benefits. These include premature mortality associated with daily fluctuations in PM, infant
VII-70
-------
Chapter VII: Benefit-Cost Analysis
mortality associated with PM, and premature mortality associated with daily fluctuations in
ozone.
We have simulated a distribution around our preferred estimate to characterize
uncertainty in the total benefit estimate due to measurement uncertainty (i.e. variance of
estimated C-R functions and valuation functions) holding all other potentially uncertain inputs
constant. Based on the simulated distribution, we have included calculations of the 5th and 95th
percentiles of the distribution of benefits in Table VII-18. This provides an estimate of how
sensitive the preferred estimate of total benefits would be to measurement errors (i.e. statistical
uncertainty around C-R and valuation functions) if all other factors could be treated as certain.
However, these do not represent the actual range of benefits, given the large number of uncertain
factors for which we are not able to provide uncertainty estimates. 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).
VII-71
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table VII-18. Alternative Benefits Calculations for the Tier 2/Gasoline Sulfur Rule in 2030
Alternative
Calculation
Description of Estimate
Impact on Preferred
Benefit Estimate
(million 1997$)
5th percentile of
"measurement"
uncertainty
distribution
Estimate of total monetized benefits at the 5th percentile of a
distribution generated using Monte Carlo simulation
assuming measurement error is the only source of uncertainty
in the benefits estimates.
-$20,300 (-81%)
95th percentile of
"measurement"
uncertainty
distribution
Estimate of total monetized benefits at the 95th percentile of a
distribution generated using Monte Carlo simulation
assuming measurement error is the only source of uncertainty
in the benefits estimates.
+$33,900 (+134%)
PM-related premature
mortality based on
Dockery et al.
The Dockery et al. study provides an alternative estimate of
the relationship between chronic PM exposure and mortality.
The number of avoided mortality incidences increases from
4,300 to 9,800 (128%)
+$30,200 (+120%)
Value of avoided
premature mortality
incidences based on
statistical life years.
Calculate the incremental number of life-years lost from
exposure to changes in ambient PM and use the value of a
statistical life year based on a $5.9 million value of a
statistical life
-$11,500 (-46%)
Reversals in chronic
bronchitis treated as
lowest severity cases
Instead of omitting those cases of chronic bronchitis that
reverse after a period of time, they are treated as being cases
with the lowest severity rating. The number of avoided
chronic bronchitis incidences increases from 2,300 to 4,300
(87%)
+$280 (+1%)
Value of visibility
changes in all Class I
areas
Values of visibility changes at Class I areas in California, the
Southwest, and the Southeast are transferred to visibility
changes in Class I areas in other regions of the country.
+$180 (+1%)
Value of visibility
changes in Eastern
U.S. residential areas
Value of visibility changes outside of Class I areas are
estimated for the Eastern U.S. based on the reported values
for Chicago and Atlanta from McClelland, et al. (1990)
+$420 (+2%)
Value of visibility
changes in Western
U.S. residential areas
Values of visibility changes outside of Class I areas are
estimated for the Western U.S. based on the reported values
for Chicago and Atlanta derived from McClelland et al.
(1990)
+$130 (+1%)
Household soiling
damage
Value of decreases in expenditures on cleaning are estimated
using values derived from Manuel, et al. (1983)
VII-72
-------
Chapter VII: Benefit-Cost Analysis
Alternative
Calculation
Avoided costs of
reducing nitrogen
loadings in east coast
estuaries
Description of Estimate
Estuarine benefits in 12 east coast estuaries from reduced
atmospheric nitrogen deposition are approximated using the
avoided costs of removing or preventing loadings from
terrestrial sources.
Impact on Preferred
Benefit Estimate
(million 1997$)
+$160 (+1%)
The 5th and 95th percentile alternative calculations (rows 1 and 2 of Table VII-18) are
estimated by holding air quality changes, population estimates, and other factors, including the
parameters examined in Table VII-18, constant and determining the distribution of total benefits
that would be generated by a large number of random draws from the distributions of C-R
functions and economic valuation functions. These alternative calculations thus show how the
preferred estimate of benefits changes in response to uncertainty in the measurement of C-R and
valuation functions.
The Dockery, et al. estimate of the relationship between PM exposure and premature
mortality (row 3 of Table VII-18) is a plausible alternative to the Pope, et al. The SAB has noted
that "the study had better monitoring with less measurement error than did most other studies"
(EPA-SAB-COUNCIL-ADV-99-012, 1999). However, the Dockery study had a more limited
geographic scope (and a smaller study population) than the Pope, et al. study. The demographics
of the Pope,et al. study population, i.e. largely white and middle-class, may also produce a
downward bias in the Pope PM mortality coefficient, because short-term studies indicate that the
effects of PM tend to be significantly greater among groups of lower socioeconomic status. The
Dockery study also covered a broader age category (25 and older compared to 30 and older in the
Pope study) and followed the cohort for a longer period (15 years compared to 8 years in the
Pope study). For these reasons, the Dockery study is considered to be a plausible alternative
estimate of the avoided premature mortality incidences associated with the final Tier 2/Gasoline
Sulfur rule.
The value of statistical life years alternative calculation (row 4 of Table VII-18)
recognizes that individuals who die from air pollution related causes tend to be older than the
average age of individuals in the VSL studies used to develop the $5.9 million value.
The treatment of reversals in chronic bronchitis incidences is addressed in row 5 of Table
VII-18. Reversals are defined as those cases where an individual reported having chronic
bronchitis at the beginning of the study period but reported not having chronic bronchitis in
follow-up interviews at a later point in the study period. Since, by definition, chronic diseases
are long-lasting or permanent, if the disease goes away it is not chronic. However, we have not
captured the benefits of reducing incidences of bronchitis that are somewhere in-between acute
and chronic. One way to address this is to treat reversals as cases of chronic bronchitis that are at
VII-73
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
the lowest severity level. These cases thus get the lowest value for chronic bronchitis.
The alternative calculation for recreational visibility (row 6 of Table VII-18) is an
estimate of the full value of visibility in the entire region affected by the final Tier 2/Gasoline
Sulfur rule. The Chestnut and Rowe study from which the primary valuation estimates are
derived only examined WTP for visibility changes in the southeastern portion of the affected
region. In order to obtain estimates of WTP for visibility changes in the northeastern and central
portion of the affected region, we have to transfer the southeastern WTP values. This introduces
additional uncertainty into the estimates. However, we have taken steps to adjust the WTP
values to account for the possibility that a visibility improvement in parks in one region, is not
necessarily the same environmental quality good as the same visibility improvement at parks in a
different region. This may be due to differences in the scenic vistas at different parks,
uniqueness of the parks, or other factors, such as public familiarity with the park resource. To
take this potential difference into account, we adjusted the WTP being transferred by the ratio of
visitor days in the two regions20 A complete discussion of the benefits transfer method used to
generate this alternative estimate is provided in the benefits TSD for this RIA (Abt Associates,
1999)
The alternative calculation for residential visibility (row 7 of Table VII-18) is based on
the McClelland, et al. study of WTP for visibility changes in Chicago and Atlanta. As discussed
in section F.I, the residential visibility estimates from the available literature have been
determined by the SAB to be inadequate for use in a primary estimate in a benefit-cost analysis.
However, EPA recognizes that residential visibility is likely to have some value and the
McClelland, et al. estimates are the most useful in providing an estimate of the likely magnitude
of the benefits of residential visibility improvements.
The alternative calculation for household soiling (row 8 of Table VII-18) is based on the Manuel
et al. study of consumer expenditures on cleaning and household maintenance. This study has
been cited as being "the only study that measures welfare benefits in a manner consistent with
economic principals (Desvouges et al., 1998). However, the data used to estimate household
soiling damages in the Manuel, et al. study is from a 1972 consumer expenditure survey and as
such may not accurately represent consumer preferences in 2007. EPA recognizes this limitation,
but believes the Manuel, et al. estimates are still useful in providing an estimate of the likely
magnitude of the benefits of reduced household soiling by particulate matter.
The alternative calculation for the avoided costs of reductions in nitrogen loadings (row 9 of
Table VII-18) is constructed by examining the avoided costs to surrounding communities of
reduced nitrogen loadings for three case study estuaries (EPA, 1998).21 The three case study
estuaries are chosen because they have agreed upon nitrogen reduction goals and the necessary
nitrogen control cost data. The values of atmospheric nitrogen reductions are determined on the
basis of avoided costs associated with agreed upon controls of nonpoint water pollution sources.
Benefits are estimated using a weighted-average, locally-based cost for nitrogen removal from
VII-74
-------
Chapter VII: Benefit-Cost Analysis
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. The avoided cost measure is likely to be an underestimate of the value of
reduced nitrogen loadings in eastern estuaries because: 1) the twelve estuaries represent only
about fifty percent of the total watershed area in the eastern U.S.; and 2) costs avoided are not
good proxies for willingness-to-pay. The details of the nitrogen deposition benefits calculation
are provided in the benefits TSD for this RIA (Abt Associates, 1999).
G. Summary of Cost Results
Since the benefits assessment has been performed on the basis of a 2030 fleet of Tier 2
vehicles, consistent costs were developed using the same basis. 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 2030 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 to multiply the cost effectiveness ratio (dollars per
ton) times the emission reduction estimates for the benefits assessment.
The cost-effectiveness value that was used in our calculation of applicable costs was
calculated as the ratio of the net present value of vehicle and fuel costs divided by the net present
value of emission reductions for an average vehicle meeting our Tier 2 standards, as described in
Section VI. However, the cost-effectiveness value used for our benefit-cost analysis differed in
several ways from those in Table VI-8. These differences ensured that the cost-effectiveness
value represented the same set of assumptions that were used when we developed the emission
inventories for use in the air quality modeling that formed the basis of the benefits analysis.
Specifically, we did not include the larger LDT4 trucks weighing greater than 8500 Ib GVWR,
and we did not include any effects of catalyst irreversibility. We also focused on the "long-term"
cost-effectiveness as described above, since this value best represents the cost-effectiveness in
VII-75
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
2030.
Finally, adjustments for two factors relating to fuel consumption needed to be made to
enable us to arrive at a final value for the adjusted cost. First the cost effectiveness value we
calculated did not account for the approximately 20 percent drop in the per-gallon fuel cost
occurring about 2020 (due to the lowering cost of desulfurization technology discussed in
Chapter V). Second, our cost effectiveness value also did not include the effects of expected
improvements in fleet fuel economy prior to 2020, also discussed in Chapter V. Improvements
in light duty truck and passenger car fuel economy are expected to reduce 2020 and beyond per-
mile fuel consumption by a little over 10 percent compared to that used in our cost effectiveness
calculation. Adjustments to account for the impact of these factors on total cost are included in
the table below.
There is one factor related to calculation of the adjusted cost which we were not able to
quantitatively account for in this analysis. This relates to the increasing rate of mileage
accumulation per vehicle over time. Fleet wide VMT is generally growing faster than vehicle
sales, indicating a gradual growth in VMT per vehicle. However, our per-vehicle cost
effectiveness is based on the current distribution of VMT with age of the vehicle, providing a
conservative basis for our cost-effectiveness calculations. By 2030, this assumption is likely to
yield substantially higher cost-effectiveness values than are appropriate. At this time we have no
specific estimate of the impact of this growth in per-vehicle VMT, so no adjustment has been
made to account for its existence. The adjusted cost would be lower if this factor were accounted
for.
The resulting adjusted costs are somewhat greater than the actual annual cost of the
program, reflecting the time value adjustment and lack of correction for the increase in VMT per
vehicle. Thus, the costs presented in this section do not represent actual annual costs of the Tier
2/gasoline sulfur program for 2030. Rather, they represent an approximation of the steady-state
cost per ton that would likely prevail in that time period. Except for the VMT adjustment, 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 per-vehicle costs are larger in the early years of the program while
the benefits are smaller. The resulting adjusted cost value is given in Table VII-19.
VII-76
-------
Chapter VII: Benefit-Cost Analysis
Table VII -19. Adjusted Cost for Comparison to Benefits
Cost per ton
ratio
$2,107
fuel cost
adjustment
0.87
fuel economy
adjustment
0.9
Tons of
NOx +
NMHC
3,204,600
Adjusted Cost
(billions of
1997 dollars)
$5.3
I. Comparison of Costs and Benefits
Benefit-cost analysis provides a valuable framework for organizing and evaluating
information on the effects of environmental programs. When used properly, benefit-cost analysis
helps illuminate important potential effects of alternative policies and helps set priorities for
closing information gaps and reducing uncertainty. According to economic theory, the efficient
alternative policy maximizes net benefits to society (i.e., social benefits minus social costs).
However, not all relevant costs and benefits can be captured in any analysis. Executive Order
12866 clearly indicates that unquantifiable or nonmonetizable categories of both costs and
benefits should not be ignored. There are many important unquantified and unmonetized costs
and benefits associated with reductions in emissions, including many health and welfare effects.
Potential benefit categories that have not been quantified and monetized are listed in Table VII-1
of this chapter.
In addition to categories that cannot be included in the calculated net benefits, there are
also practical limitations for the comparison of benefits to costs in this analysis, which have been
discussed throughout this chapter. Several specific limitations deserve to be mentioned again
here:
The state of atmospheric modeling is not sufficiently advanced to provide a workable
"one atmosphere" model capable of characterizing ground-level pollutant exposure for all
pollutants of interest (e.g., ozone, particulate matter, carbon monoxide, nitrogen
deposition, etc). Therefore, the EPA must employ several different pollutant models to
characterize the effects of alternative policies on relevant pollutants. Also, not all
atmospheric models have been widely validated against actual ambient data. In
particular, since a broad-scale monitoring network does not yet exist for fine particulate
matter (PM2 5), atmospheric models designed to capture the effects of alternative policies
on PM2 5 are not fully validated. Additionally, significant shortcomings exist in the data
that are available to perform these analyses. While containing identifiable shortcomings
and uncertainties, EPA believes the models and assumptions used in the analysis are
reasonable based on the available evidence.
VII-77
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Another dimension adding to the uncertainty of this analysis is time. Thirty years is a
very long time over which to carry assumptions. Projected growth in population and
VMT over the 30-year period may have a significant effect on the benefits estimates.
Pollution control technology has advanced considerably in the last 10 years and can be
expected to continue to advance in the future. Yet there is no clear way to model this
advance for use in this analysis. In addition, there is no clear way to predict future
meteorological conditions, or the growth in emissions from other sources over time.
Again, EPA believes that the assumptions to capture these elements are reasonable based
on the available evidence.
Qualitative and more detailed discussions of the above and other uncertainties and
limitations are included in detail in earlier sections. Where information and data exist,
quantitative characterizations of these uncertainties are included (in this chapter and in
Appendix VII-A). However, data limitations prevent an overall quantitative estimate of
the uncertainty associated with final estimates. Nevertheless, the reader should keep all
of these uncertainties and limitations in mind when reviewing and interpreting the results.
• The preferred benefit estimate does not include the monetary value of several known
ozone and PM-related welfare effects, including commercial forest growth, residential
visibility, household soiling and materials damage, and deposition of nitrogen to sensitive
estuaries.
Nonetheless, if one is mindful of these limitations, the relative magnitude of the benefit-
cost comparison presented here can be useful information. Thus, this section summarizes the
benefit and cost estimates that are potentially useful for evaluating the efficiency of the final Tier
2 rulemaking.
The estimated adjusted cost of implementing the final Tier 2 program is $5.3 billion
(1997$), while the estimate of monetized benefits using EPA's preferred approach for
monetizing reductions in PM-related premature mortality - the VSL approach - are $25.2
billion (1997$). Monetized net benefits using EPA's preferred method for valuing avoided
incidences of premature mortality are approximately $19.9 billion (1997$). Using the
alternative, age-adjusted approach - the VSLY approach - total monetized benefits are projected
to be around $13.8 billion (1997$). Monetized net benefits using this approach are
approximately $8.5 billion (1997$). Therefore, implementation of the Tier 2 program will
provide society with a net gain in social welfare. Tables VII-20 and Vn-21 summarize the costs,
benefits, and net benefits for the two alternative valuation approaches.
VII-78
-------
Chapter VII: Benefit-Cost Analysis
Table VII-20. 2030 Annual Monetized Costs, Benefits, and Net Benefits for the Final Tier
2/Gasoline Sulfur Rule: EPA Preferred Estimates Using the Value of Statistical Lives
Saved Approach to Value Reductions in Premature Mortalitya
Adjusted compliance costs
Monetized PM-related benefitsM
Monetized Ozone-related benefits'"
Monetized net benefits0'*1
Billion 1997$
$5.3
$24.7+BPM
$0.5+B0zone
S19.9+B
" For this section , all costs and benefits are rounded to the nearest 100 million. Thus, figures presented in this chapter may not exactly equal
benefit and cost numbers presented in earlier sections of the chapter.
b Not all possible benefits or disbenefits are quantified and monetized in this analysis. Potential benefit categories that have not been quantified
and monetized are listed in Table VII-1. Unmonetized PM- and ozone-related benefits are indicated by BPM. And B0zone, respectively.
0 B is equal to the sum of all unmonetized benefits, including those associated with PM, ozone, CO, and HAPS.
d These estimates are based on the EPA preferred approach for valuing reductions in premature mortality, the VSL approach. This approach and
an alternative, age-adjusted approach - the VSLY approach - are discussed more fully in section F.
Table VII-21. 2030 Annual Monetized Costs, Benefits, and Net Benefits for the Final Tier
2/Gasoline Sulfur Rule: Alternative Estimates Using the Value of Statistical Life Years
Saved Approach to Value Reductions in Premature Mortalitya
Adjusted compliance costs
Monetized PM-related benefitsM
Monetized Ozone-related benefits'"
Monetized net benefitsc>d
Billion 1997$
$5.3
$13.3+BPM
$0.5+B0zon(,
$8.5+B
* For this section , all costs and benefits are rounded to the nearest 100 million. Thus, figures presented in this chapter may not exactly equal
benefit and cost numbers presented in earlier sections of the chapter.
b Not all possible benefits or disbenefits are quantified and monetized in this analysis. Potential benefit categories that have not been quantified
and monetized are listed in Table VII-1. Unmonetized PM- and ozone-related benefits are indicated by BPM. And B0zone, respectively.
0 B is equal to the sum of all unmonetized benefits, including those associated with PM, ozone, CO, and HAPS.
d The VSLY estimate is calculated by amortizing the $5.9 million mean VSL estimate over the 35 years of life expectancy associated with
subjects in the labor market studies used to obtain the VSL estimate. The resulting estimate, using a 5 percent discount rate, is $360,000 per
life-year saved in 1997 dollars. This approach is discussed more fully in section F above.
H. 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.L. Hwang, RJ. Burchette, T. Vancuren and P.K. Mills. 1995a. Estimated Long-
VII-79
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Term Ambient Concentrations of Pm(10) and Development of Respiratory Symptoms in a
Nonsmoking Population. Archives of Environmental Health. 50(2): 139-152.
Abt Associates Inc., 1999b. Memorandum to Lisa Conner, US EPA, Office of Air Quality
Planning and Standards. "Derivation of 2030 Population Estimates for the Tier II Analysis."
September.
Abt Associates Inc., 1999a. 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.
Abt Associates, Inc. 1999. Tier 11 Final Rule: Air Quality Estimation, Selected Health
and Welfare Benefits Methods, and Benefit Analysis Results, Prepared for the U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards; Research
Triangle Park, N.C., November.
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)
American Lung Association, 1999. Chronic Bronchitis. Web site available at:
http://www.lungusa.org/diseases/lungchronic.html
Blumenschein, Karen and Magnus Johannesson. 1998. "Relationship Between Quality of Life
Instruments, Health State Utilities, and Willingness to Pay in Patients with Asthma." Annals of
Allergy, Asthma, and Immunology 80:189-194.
Chestnut, L.G. 1997. Draft Memorandum: Methodology for Estimating Values for Changes in
Visibility at National Parks.; April 15.
Chestnut, L.G. and R.D. Rowe. 1990a. Preservation Values for Visibility Protection at the
National Parks: Draft Final Report. Prepared for Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, Research Triangle Park, NC and Air Quality
Management Division, National Park Service, Denver, CO.
Chestnut, L.G. and R.L. Dennis. 1997. AEconomic Benefits of Improvements in Visibility: Acid
Rain Provisions of the 1990 Clean Air Act Amendments.@ Journal of Air and Waste
Management Association 47:395-402.
Chestnut, L.G., and R.D. Rowe. 1990b. ANew National Park Visibility Value Estimates.® In
Visibility and Fine Particles., Transactions of an AWMA/EPA International Specially
Conference, C. V. Mathai, ed. Air and Waste Management Association, Pittsburgh.
VII-80
-------
Chapter VII: Benefit-Cost Analysis
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.
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).
Desvouges, W.H., F. R. Johnson, and H.S. Banzhaf. 1998. Environmental Policy Analysis with
Limited Information: Principles and Applications of the Transfer Method. Edward Elgar,
Northhampton, MA.
Dickie, M. et al. 1991. Reconciling Averting Behavior and Contingent Valuation Benefit
Estimates of Reducing Symptoms of Ozone Exposure (draft), as cited in Neumann, I.E., Dickie,
M.T., and R.E. Unsworth. 1994. Industrial Economics, Incorporated. Memorandum to Jim
DeMocker, U.S. EPA, Office of Air and Radiation. March 31.
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 six U.S. cities. NEnglJ
Med. 329(24): 1753-1759.
Dockery, D.W., J. Cunningham, A.I. Damokosh, L.M. Neas, J.D. Spengler, P. Koutrakis, J.H.
Ware, M. Raizenne and F.E. Speizer. 1996. Health Effects of Acid Aerosols On North American
Children-Respiratory Symptoms. Environmental Health Perspectives. 104(5): 500-505.
Douglas, S.G. andR.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 Particulate 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.
Elixhauser, A., R.M. Andrews, and S. Fox. 1993. Clinical Classifications for Health Policy
Research: Discharge Statistics by Principal Diagnosis and Procedure. Agency for Health Care
Policy and Research (AHCPR), Center for General Health Services Intramural Research, U.S.
VII-81
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Department of Health and Human Services.
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.
EPA-SAB-COUNCIL-ADV-99-05, 1999. An SAB Advisory on the Health and Ecological
Effects Initial Studies of the Section 812 Prospective Study: Report to Congress: Advisory by the
Health and Ecological Effects Subcommittee, February.
EPA-SAB-COUNCIL-ADV-99-012, 1999. The Clean Air Act Amendments (CAAA) Section
812 Prospective Study of Costs and Benefits (1999): Advisory by the Health and Ecological
Effects Subcommittee on Initial Assessments of Health and Ecological Effects; Part 1. July.
EPA-SAB-COUNCIL-ADV-00-001, 1999. The Clean Air Act Amendments (CAAA) Section
812 Prospective Study of Costs and Benefits (1999): Advisory by the Health and Ecological
Effects Subcommittee on Initial Assessments of Health and Ecological Effects; Part 2. October.
EPA-SAB-COUNCIL-ADV-00-002, 1999. The Clean Air Act Amendments (CAAA) Section
812 Prospective Study of Costs and Benefits (1999): Advisory by the Advisory Council on Clean
Air Compliance Analysis: Costs and Benefits of the CAAA
EPA-SAB-COUNCIL-ADV-98-003, 1998. Advisory Council on Clean Air Compliance Analysis
Advisory on the Clean Air Act Amendments (CAAA) of 1990 Section 812 Prospective Study:
Overview of Air Quality and Emissions Estimates: Modeling, Health and Ecological Valuation
Issues Initial Studies.
Fox, S., and R.A. Mickler, 1995. Impact of Air Pollutants on Southern Pine Forests Ecological
Studies 118. Springer Verlag: New York.
Grosclaude, P. and N.C. Soguel. 1994. Valuing Damage to Historic Buildings Using a
Contingent Market: A Case Study of Road Traffic Externalities. Journal of Environmental
Planning and Management, 37: 279-287.
Holland, M., D. Forster, and M. Wenborn. 1999. Economic Evaluation of Proposals Under the
UNECE Multi-effects and Multi-pollutant Protocol. Prepared for: European Commission, DGXI,
Brussels and Luxembourg. January. Report no. AEAT-4587.
Industrial Economics, Incorporated (ffic). 1994. Memorandum to Jim DeMocker, Office of Air
and Radiation, Office of Policy Analysis and Review, U.S. Environmental Protection Agency,
March 31.
VII-82
-------
Chapter VII: Benefit-Cost Analysis
Krupnick, AJ. and M.L. Cropper. 1992. The Effect of Information on Health Risk Valuations.
Journal of Risk and Uncertainty 5(2): 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.
Lang, C., G. Yarwood, F. Lalonde, and R. Bloxam. 1995. Environmental and Health Benefits of
Cleaner Vehicles and Fuels. Prepared for: Canadian Council of Ministers of the Environment
Task Force on Cleaner Vehicles and Fuels, Winnipeg, Manitoba. October.
Latimer and Associates, 1996. Particulate 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.
Loehman, E.T., S.V. Berg, A.A. Arroyo, R.A. Hedinger, J.M. Schwartz, M.E. Shaw, R.W.
Fahien, V.H. De, R.P. Fishe, D.E. Rio, W.F. Rossley, and A.E.S. Green. 1979. "Distributional
Analysis of Regional Benefits and Cost of Air Quality Control." Journal of Environmental
Economics and Management 6: 222-243.
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, NC.
McClelland, G., W. Schulze, D. Waldman, J. Irwin, D. Schenk, T. Stewart, L. Deck, and M.
Thayer. 1993. Valuing Eastern Visibility: A Field Test of the Contingent Valuation Method.
Prepared for Office of Policy, Planning and Evaluation, U.S. Environmental Protection Agency.
September.
McDonnell, W.F., D.E. Abbey, N. Nishino and M.D. Lebowitz. 1999. Long-term ambient ozone
concentration and the incidence of asthma in nonsmoking adults: the ahsmog study. Environ
Res. 80(2 Ptl): 110-21.
NCLAN. 1988. Assessment of Crop Loss from Air Pollutants. (Eds. Walter W. Heck, O. Clifton
Taylor and David T. Tingey) Elsevier Science Publishing Co.: New York,. Pp. 1-5. (ERL,GB
639).
O'Conor, Richard M. and Glenn C. Blomquist, 1997. "Measurement of Consumer-Patient
Preferences Using a Hybrid Contingent Valuation Method." Journal of Health Economics
16:667-683.
VII-83
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Ostro B.D. and S. Rothschild. 1989. Air Pollution and Acute Respiratory Morbidity: An
Observational Study of Multiple Pollutants. Environmental Research 50:238-247.
Ostro, B.D., MJ. Lipsett, J.K. Mann, H. Braxton-Owens, and M.C. White. 1995. "Air Pollution
and Asthma Exacerbations Among African American Children in Los Angeles." Inhalation
Toxicology.
Ostro, B.D. 1987. Air Pollution and Morbidity Revisited: a Specification Test. J. Environ.
Econ. Manage. 14: 87-98.
Pope, C.A., HI, MJ. Thun, M.M. Namboodiri, D.W. Dockery, J.S. Evans, F.E. Speizer, and
C.W. Heath, Jr. 1995. Particulate Air Pollution as a Predictor of Mortality in a Prospective
Study of U.S. Adults. Am. J. Respir. Crit. Care Med. 151: 669-674.
Pope, C.A., HI, D.W. Dockery, J.D. Spengler, and M.E. Raizenne. 1991. Respiratory Health and
PM10 Pollution: a Daily Time Series Analysis. Am. Rev. Respir. Dis. 144: 668-674.
Rowe, R.D. and L.G. Chestnut. 1986. "Oxidants and Asthmatics in Los Angeles: A Benefits
Analysis—Executive Summary." Prepared by Energy and Resource Consultants, Inc. Report to
the U.S. EPA, Office of Policy Analysis. EPA-230-09-86-018. Washington, D.C. March.
Schwartz, J. 1993. Parti culate Air Pollution and Chronic Respiratory Disease. Environmental
Research 62: 7-13.
Schwartz, J., Dockery, D.W., Neas, L.M, Wypij, D., Ware, J.H., Spengler, J.D., Koutrakis, P.,
Speizer, F.E., and Ferris, Jr., E.G. 1994. Acute Effects of Summer Air Pollution on
Respiratory Symptom Reporting in Children. Am. J. Respir. Crit. Care Med. 150: 1234-1242.
Sisler, J.F. 1996. Spatial and Seasonal Patterns and Long Term Variability of the Composition of
the Haze in the United States: An Analysis of Data from the IMPROVE Network. Cooperative
Institute for Research in the Atmosphere, Colorado State University; Fort Collins, CO; July.
Sisler, J.F., 1996. Spatial and Seasonal Patterns and Long Term Variability of the Composition
of the Haze in the United States: An Analysis of Data from the IMPROVE Network. 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, D.H., D.C. Malone, K.A. Lawson, L. J. Okamoto, C. Battista, and W.B. Saunders, 1997.
"A National Estimate of the Economic Costs of Asthma. Am. J. Respir. Crit. Care Med. 156:
787-793.
Smith, V.K., G.Van Houtven, and S. Pattanayak. 1999. Benefit Transfer as Preference
Calibration. Resources for the Future Discussion Paper 99-36.
VII-84
-------
Chapter VII: Benefit-Cost Analysis
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.
Tolley, G.S. et al. 1986. Valuation of Reductions in Human Health Symptoms and Risks.
University of Chicago. Final Report for the U.S. Environmental Protection Agency. January.
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, 1997e. 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. See EPA Air Docket A-96-
56, Document No. VI-B-09-(r).
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).
U.S. EPA, 1996a. Proposed Methodology for Predicting PM2.5 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, 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.
U.S. Environmental Protection Agency, 1999. Regulatory Impact Analysis for the Final
Regional Haze Rule. Office of Air Quality Planning and Standards, Research Triangle Park, NC.
April 22, 1999.
U.S. Department of Commerce, Bureau of Economic Analysis. BEA Regional Projections to
2045: Vol. 1, States. Washington, D.C. U.S. Govt. Printing Office, July 1995.
U.S. Department of Commerce, Economics and Statistics Administration. 1992. Statistical
Abstract of the United States, 1992: The National Data Book. 112th Edition, Washington, D.C.
VII-85
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
U.S. Environmental Protection Agency, 1998a. Regulatory Impact Analysis for the NOx SIP
Call, FIP and Section 126 Petitions, Volume 2: Health and Welfare Benefits. Prepared by:
Innovative Strategies and Economics Group, Office of Air Quality Planning and Standards,
Research Triangle Park, N.C. December.
U.S. Department of Health and Human Services, Centers for Disease Control and Prevention,
National Center for Health Statistics. 1994. Vital Statistics of the United States, 1990. Volume
II-Mortality. Hyattsville, MD.
U.S. Department of Health and Human Services, Centers for Disease Control and Prevention.
1999. National Center for Environmental Health Web Site. Available at
http://www.cdc.gov/nceh/programs/asthma/ataglance/asthmaag2.htm
U.S. EPA. 1993. External Draft, Air Quality Criteria for Ozone and Related Photochemical
Oxidants. Volume II. U.S. EPA, Office of Health and Environmental Assessment. Research
Triangle Park, NC. EPA/600/AP-93/004b.3v.
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. Environmental Protection Agency, 1996a. Review of the National Ambient Air Quality
Standards for Ozone: Assessment of Scientific and Technical Information. Office of Air Quality
Planning and Standards, Research Triangle Park, N.C.; EPA report no. EPA/4521R-96-007.
U.S. Environmental Protection Agency, 1996b. Review of the National Ambient Air Quality
Standards for Particulate Matter: Assessment of Scientific and Technical Information. Office of
Air Quality Planning and Standards, Research Triangle Park, N.C.; EPA report no. EPA/4521R-
96-013.
U.S. Environmental Protection Agency, 1997a. Regulatory Impact Analysis for Particulate
Matter and Ozone National Ambient Air Quality Standards and Proposed RH Rule. Prepared by:
Innovative Strategies and Economics Group, Office of Air Quality Planning and Standards,
Research Triangle Park, N.C. July.
U. S. Environmental Protection Agency, 1997b. The Benefits and Costs of the Clean Air Act,
1970 to 1990. Prepared for U.S. Congress by U.S. EPA, Office of Air and Radiation/Office of
Policy Analysis and Review, Washington, D.C.
U.S. Environmental Protection Agency, 1998b. The Regional NOx SIP Call & Reduced
Atmospheric Deposition of Nitrogen: Benefits to Selected Estuaries, September.
VII-86
-------
Chapter VII: Benefit-Cost Analysis
U. S. Environmental Protection Agency, 1999a. The Benefits and Costs of the Clean Air Act,
1990-2010. Prepared for U.S. Congress by U.S. EPA, Office of Air and Radiation/Office of
Policy Analysis and Review, Washington, D.C., November; EPA report no. EPA-410-R-99-001
U.S. Environmental Protection Agency, 1999b. Regulatory Impact Analysis for the Final
Regional Haze Rule. Prepared by: Office of Air Quality Planning and Standards, Office of Air
and Radiation, April.
U.S. Environmental Protection Agency, 1999c. Regulatory Impact Analysis: Control of Air
Pollution from New Motor Vehicles: Tier 2 Motor Vehicle Emissions Standards and Gasoline
Sulfur Control Requirements. Prepared by: Office of Mobile Sources, Office of Air and
Radiation, April.
U.S. Environmental Protection Agency, 1996a. Review of the National Ambient Air Quality
Standards for Ozone: Assessment of Scientific and Technical Information. Office of Air Quality
Planning and Standards, Research Triangle Park, N.C.; EPA report no. EPA/4521R-96-007.
U.S. Environmental Protection Agency, 1996b. Review of the National Ambient Air Quality
Standards for P articulate Matter: Assessment of Scientific and Technical Information. Office of
Air Quality Planning and Standards, Research Triangle Park, N.C.; EPA report no. EPA/4521R-
96-013.
Valigura, Richard A., and W.T. Luke, R.S. Artz, B.B. Hicks. 1996. Atmospheric input to
coastal areas: reducing the uncertainties. NOAA Coastal Ocean Program, Decision Analysis
Series No. 9. NOAA Atmospheric Resources Laboratory, Silver Spring, MD.
Valiela, G. Collins, and J. Kremer, K. Lajtha, M. Geist, B. Seely, J. Brawley, C.H. Sham. 1997.
"Nitrogen loading from coastal watersheds to receiving estuaries: new method and application."
Ecological Applications 7(2), pp. 358-380.
Viscusi, W.K. 1992. Fatal Tradeoffs: Public and Private Responsibilities for Risk. (New York:
Oxford University Press).
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: 32-51.
VII-87
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Appendix VII-A
Supplementary Benefit Estimates and Sensitivity Analyses of Key
Parameters in the Benefits Analysis
A. Introduction and Overview
In chapter VII, we estimated the benefits of the final Tier 2/gasoline sulfur rule using the
most comprehensive set of endpoints available. For some health endpoints, this meant using a
dose-response function that linked a larger set of effects to a change in pollution, rather than
using dose-response functions for individual effects. For example, the minor restricted activity
days/any of 19 acute respiratory symptoms endpoint covers most of the symptoms used to
characterize asthma attacks and days of moderate or worse asthma. For premature mortality, we
selected a dose-response function that captured reductions in incidences due to both long and
short-term exposures to ambient concentrations of particulate matter (PM). In addition, the
premature mortality dose-response function is expected to capture at least some of the mortality
effects associated with exposure to ozone. This effect is described more fully below in section
A.2.
In order to provide the reader with a fuller understanding of the health effects associated
with reductions in air pollution associated with the final Tier 2/gasoline sulfur rule, this appendix
provides estimates for those health effects which, if included in the primary estimate, could result
in double-counting of benefits. For some endpoints, such as ozone mortality, additional research
is needed to provide separate estimates of the effects for different pollutants, i.e. PM and ozone.
These supplemental estimates should not be considered as additive to the primary estimate of
benefits. Supplemental estimates included in this appendix include premature mortality
associated with short-term exposures to PM and ozone, asthma attacks, and occurrences of
moderate or worse asthma symptoms. In addition, an estimate of the avoided incidences of
premature mortality in infants is provided. Because the Pope, et al. estimate applies only to
adults, avoided incidences of infant mortality are additive to the primary benefits estimate.
Table VII-19 in Chapter VII reports the results of alternative calculations based on
plausible alternatives to the assumptions used in deriving the primary estimate of benefits. In
addition to these calculations, two important parameters, the length and structure of the potential
lag in mortality effects and thresholds in PM health effects, have been identified as key to the
analysis, and are explored in this appendix through the use of sensitivity analyses.
B. Supplementary Benefit Estimates
In the primary estimate, we use the Pope et al. study to provide the C-R function
relating premature mortality to long-term PM exposure. In the primary analysis, we assume that
VII-88
-------
Chapter VII: Benefit-Cost Analysis
this mortality occurs over a five year period, with 25 percent of the deaths occurring in the first
year, 25 percent in the second year, and 16.7 percent in each of the third, fourth, and fifth years.
Studies examining the relationship between short-term exposures and premature mortality can
reveal what proportion of premature mortality is due to immediate response to daily variations in
PM. There is only one short-term study (presenting results from 6 separate U.S. cities) that uses
PM25 as the metric of PM (Schwartz et al., 1996). As such, the supplemental estimate for
premature mortality related to short-term PM exposures is based on the pooled city-specific,
short-term PM2 5 results from Schwartz, et al.
In the Tier 2 Proposal RIA, we estimated avoided incidences of ozone-related premature
mortality for the primary benefits estimate. Based on recent advice from the SAB (EPA-SAB-
Council-ADV-99-012, 1999), we have converted this endpoint to a supplemental estimate to
avoid potential double-counting of benefits captured by the Pope, et al. PM premature mortality
endpoint32. There are many studies of the relationship between ambient ozone levels and daily
mortality levels. The supplemental estimate is calculated using results from only four U.S.
studies (Ito and Thurston, 1996; Kinney et al., 1995; Moolgavkar et al., 1995; and Samet et al.,
1997), based on the assumption that demographic and environmental conditions on average
would be more similar between these studies and the conditions prevailing when the Tier
2/gasoline sulfur rule is implemented. 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. We combined
these studies using probabilistic sampling methods 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). The estimated incidences of short-term premature mortality
are valued using the value of statistical lives saved method, as described in Chapter VII.
The estimated effect of PM exposure on premature mortality in infants (post neo-natal) is
based on a single U.S. study (Woodruff, et al, 1997) which, on recommendation of the Science
Advisory Board, was deemed too uncertain to include in the primary analysis. Adding this
endpoint to the primary benefits estimate would result in an increase in total benefits.
32While the growing body of epidemiological studies suggests that there may be a positive relationship
between ozone and premature mortality, there is still substantial uncertainty about this relationship. Because the
evidence linking premature mortality and paniculate matter is currently stronger than the evidence linking premature
mortality and ozone, it is important that models of the relationship between ozone and mortality include a measure
of paniculate matter as well. Because of the lack of monitoring data on fine particulates or its components, however,
the measure of paniculate matter used in most studies was generally either PM10 or TSP or, in some cases, Black
Smoke. If a component of PM, such as PM 2.5 or sulfates, is more highly correlated with ozone than with PM or
TSP, and if this component is also related to premature mortality, then the apparent ozone effects on mortality
could be at least partially spurious. Even if there is a true relationship between ozone and premature mortality, after
taking paniculate matter into account, there would be a potential problem of double counting in this analysis if the
ozone effects on premature mortality were added to the PM effects estimated by Pope et al., 1995, because, as noted
above, the Pope study does not include ozone in its model.
VII-89
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
As noted in Chapter VII, asthma affects over seven percent of the U.S. population. One
study identifies a statistical association between air pollution and the development of asthma in
some non-smoking adult men. Other studies identify a relationship between air quality and
occurrences of acute asthma attacks or worsening of asthma symptoms. Supplemental estimates
are provided for two asthma related endpoints. Occurrence of moderate or worse asthma
symptoms in adults is estimated using a C-R function derived from Ostro, et al. (1991). Asthma
attacks in children are estimated using a C-R function derived from Whittemore and Korn
(1980). Both asthma attacks and occurrence of moderate or worse asthma symptoms are valued
at $39 per incidence, based on the mean of average WTP estimates for the four severity
definitions of a "bad asthma day," described in Rowe and Chestnut (1986), a study which
surveyed asthmatics to estimate WTP for avoidance of a "bad asthma day," as defined by the
subjects.
Table VII-A-1 presents estimated incidences and values for the supplemental endpoints
listed above. The supplemental estimate of 1,200 avoided incidences of premature mortality
from short-term exposures to PM indicates that these incidences are approximately 25 percent of
the total premature mortality incidences estimated using the Pope, et al., study (4,300). This
lends support for the assumption that 25 percent of the premature deaths predicted to be avoided
in the first year using the Pope study should be assigned to the first year after a reduction in
exposure.
The infant mortality estimate indicates that exclusion of this endpoint does not have a
large impact, either in terms of incidences (13) or monetary value (approximately $80 million).
Estimates of the value for separate asthma endpoints are well under the estimate of the value of
all respiratory symptoms. All of these supplemental estimates support the set of endpoints and
assumptions chosen as the basis of the primary benefits estimate described in Chapter VII.
VII-90
-------
Chapter VII: Benefit-Cost Analysis
Table VII-A-1
Supplemental Benefit Estimates for the Final Tier 2 Rule for the 2030 Analysis Year
Endpoint
Premature mortality (short-term exposures)
Premature mortality (short-term exposures)0
Premature mortality in infant population
Asthma attacks
Asthma attacks c
Moderate or Worse Asthma
Pollutant
PM
Ozone
PM
PM
Ozone
PM
Avoided
Incidence"
(cases/year)
1,200
500
<100
77,000
188,100
79,500
Monetary
Benefits'1
(millions 1997$)
$6,320
$2,670
$80
$<10
$10
$<10
* Incidences are rounded to the nearest 100.
' Dollar values are rounded to the nearest 10.
C. Sensitivity Analyses
As discussed in Chapter Vn, there are two key parameters of the benefits analysis for
which there are no specific values recommended in the scientific literature. These parameters,
the lag between changes in exposure to PM and reductions in premature mortality and the
threshold in PM-related health effects, are investigated in this section through the use of
sensitivity analyses, we perform an analysis of the sensitivity of benefits valuation to the lag
structure by considering a range of assumptions about the timing of premature mortality. To
examine the threshold parameter, we show how the estimated avoided incidences of PM-related
premature mortality are distributed with respect to the threshold.
1. Alternative Lag Structures
As noted by the SAB (EPA-SAB-COUNCIL-ADV-00-001, 1999), "some of the mortality
effects of cumulative exposures will occur over short periods of time in individuals with
compromised health status, but other effects are likely to occur among individuals who, at
baseline, have reasonably good health that will deteriorate because of continued exposure. No
animal models have yet been developed to quantify these cumulative effects, nor are there
epidemiologic studies bearing on this question." However, they also note that "Although there is
substantial evidence that a portion of the mortality effect of PM is manifest within a short period
of time, i.e., less than one year, it can be argued that, if no a lag assumption is made, the entire
mortality excess observed in the cohort studies will be analyzed as immediate effects, and this
will result in an overestimate of the health benefits of improved air quality. Thus some time lag is
appropriate for distributing the cumulative mortality effect of PM in the population." In the
VII-91
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
primary analysis, based on SAB advice, we assume that mortality occurs over a five year period,
with 25 percent of the deaths occurring in the first year, 25 percent in the second year, and 16.7
percent in each of the third, fourth, and fifth years. Readers should note that the selection of a 5
year lag is not supported by any scientific literature on PM-related mortality. Rather it is
intended to be a best guess at the appropriate distribution of avoided incidences of PM-related
mortality.
Although the SAB recommended the five-year distributed lag be used for the primary
analysis, the SAB has also recommended (EPA-SAB-COUNCIL-ADV-00-001, 1999) that
alternative lag structures be explored as a sensitivity analysis. Specifically, they recommended an
analysis of 0, 8, and 15 year lags. The 0 year lag is representative of EPA's assumption in
previous RIAs. The 8 and 15 year lags are based on the study periods from the Pope and
Dockery studies, respectively33. However, neither the Pope or Dockery studies assumed any lag
structure when estimating the relative risks from PM exposure. In fact, the Pope and Dockery
studies do not contain any data either supporting or refuting the existence of a lag. Therefore,
any lag structure applied to the avoided incidences estimated from either of these studies will be
an assumed structure. The 8 and 15 year lags implicitly assume that all premature mortalities
occur at the end of the study periods, i.e. at 8 and 15 years. We also present two additional lags:
a 15 year distributed lag with the distribution skewed towards the early years and a 15 year
distributed lag with the distribution skewed towards the later years. This is to demonstrate how
sensitive the results are not only to the length of the lag, but also to the shape of the distribution
of incidences over the lag period. It is important to keep in mind that changes in the lag
assumptions do not change the total number of estimated deaths, but rather the timing of those
deaths.
The estimated impacts of alternative lag structures on the monetary benefits associated
with reductions in PM-related premature mortality (estimated with the Pope, et al. C-R function)
are presented in Table VII-A-2. These estimates are based on the value of statistical lives saved
approach, i.e. $5.9 million per incidence, and assume a 5 percent discount rate over the lag
period. The results using the primary 5-year lag are repeated here for comparison. The table
reveals that the length of the lag period is not as important as the distribution of incidences
within the lag period. A 15 year distributed lag with most of the incidences occurring in the early
years reduces monetary benefits less than an 8 year lag with all incidences occurring at the eighth
year. Even with an extreme lag assumption of 15 years, benefits are reduced by less than half
relative to the no lag and primary (5 year distributed lag) benefit estimates.
"Although these studies were conducted for 8 and 15 years, respectively, the choice of the duration of the
study by the authors was not likely due to observations of a lag in effects, but is more likely due to the expense of
conducting long-term studies or the amount of satisfactory data that could be collected during this time period.
VII-92
-------
Chapter VII: Benefit-Cost Analysis
Table VII-A-2
Sensitivity Analysis of Alternative Lag Structures for PM-related Premature Mortality
Lag
5 -year distributed
None
8 year
15 year
15 year distributed - skewed early
15 year distributed - skewed late
Description
Primary estimate, incidences are
distributed with 25% in the 1st and
2nd years, and 16.7% in the
remaining 3 years.
Incidences all occur in the first year
Incidences all occur in the 8th year
Incidences all occur in the 15th year
Incidences are distributed with 30%
in the 1st year, 25% in the 2nd year,
15% in the 3rd year, 6% in the 4th
year, 4% in the 5th year, and the
remainder 20% distributed over the
last 10 years.
Incidences are distributed with 4%
in the 1 1th year, 6% in the 12th year,
15% in the 13th year, 25% in the
14th year, and 30% in the 15th year,
with the remaining 20 % distributed
over the first 10 years.
Monetary Benefit
(millions 1997$)
$23,400
$25,400
$18,000
$12,800
$22,700
$14,800
Percent of
Primary
Estimate
100%
109%
77%
55%
97%
63%
2.
PM Health Effect Threshold
In developing its primary estimate of benefits for previous analyses, EPA has assumed a
PM health effects threshold equal to the lowest observed level in a given epidemiological study
or anthropogenic background when no lowest observed level is reported (Hubbell, 1998). Recent
advice from the SAB (EPA-SAB-Council-ADV-99-012, 1999) is that there is currently no
scientific basis for selecting a threshold of 15 |ig/m3 or any other specific threshold for the PM
related health effects considered in this analysis. The most important health endpoint that would
be impacted by a PM threshold is premature mortality, as measured by the Pope, et al. (1995) C-
R function. Pope et al. did not explicitly include a threshold in their analysis. However, if the
true mortality C-R relationship has a threshold, then Pope et al.'s slope coefficient would likely
have been underestimated for that portion of the C-R relationship above the threshold. This
would likely lead to an underestimate of the incidences of avoided cases above any assumed
VII-93
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
threshold level. It is difficult to determine the size of the underestimate without data on a likely
threshold and without re-analyzing the Pope et al. data. Nevertheless, it is illustrative to show at
what threshold levels benefits are significantly affected.
Any of the PM-related health effects estimated in the primary analysis could have a
threshold; however a threshold for PM-related mortality would have the greatest impact on the
overall benefits analysis. Figure A-l shows the effect of incorporating a range of possible
thresholds, using 2030 PM levels and the Pope et al. (1995) study.
The distribution of premature mortality incidences in Figure A-l indicates that over
ninety percent of the premature mortality related benefits of the final Tier 2/gasoline sulfur rule
are due to changes in PM concentrations occurring above 10 |ig/m3, and around seventy-five
percent are due to changes above 12 |ig/m3, the lowest observed level in the Pope, et al. study.
Over fifty percent of avoided incidences are due to changes occurring above the PM2 5 standard
of 15 |ig/m3.
VII-94
-------
Chapter VII: Benefit-Cost Analysis
4500
o
8
u
=
HH
•a
a>
."2
'S
5 10 15 20 25 30
Assumed Effect Threshold (Annual Mean PM2.5 (ug/m3))
35
Figure VII-A-1
Impact of PM Health Effects Threshold on Avoided Incidences of Premature Mortality
Estimated with the Pope Concentration-Response Function
VII-95
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
D. References
Abt Associates, Inc. 1999. Section 126 Final Rule: Air Quality Estimation, Selected Health
and Welfare Benefits Methods, and Benefit Analysis Results, Prepared for the U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards; Research
Triangle Park, N.C., November.
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 six U.S. cities." New
England Journal of Medicine. 329(24): 1753-1759.
EPA-SAB-COUNCIL-ADV-00-001, 1999. The Clean Air Act Amendments (CAAA) Section
812 Prospective Study of Costs and Benefits (1999): Advisory by the Health and Ecological
Effects Subcommittee on Initial Assessments of Health and Ecological Effects; Part 2. October.
EPA-SAB-COUNCIL-ADV-99-012, 1999. The Clean Air Act Amendments (CAAA) Section
812 Prospective Study of Costs and Benefits (1999): Advisory by the Health and Ecological
Effects Subcommittee on Initial Assessments of Health and Ecological Effects; Part 1. July.
Hubbell, B. 1998. Memorandum to the Files. Preliminary Estimates of Benefits of the NOx SIP
Call. October.
Ito, K. and G.D. Thurston. 1996. Daily PMlO/mortality associations: an investigations of at-risk
subpopulations. Journal of Exposure Analysis and Environmental Epidemiology. 6(1): 79-95.
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.
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.
Ostro, B.D., M.J. Lipsett, M.B. Wiener and J.C. Seiner. 1991. Asthmatic Responses to Airborne
Acid Aerosols. Am J Public Health. 81(6): 694-702.
Pope, C.A., M.J. Thun, M.M. Namboodiri, D.W. Dockery, J.S. Evans, F.E. Speizer and C.W.
Heath. 1995. Parti culate air pollution as a predictor of mortality in a prospective study of U.S.
adults. Am J Respir Crit Care Med. 151(3): 669-674.
Rowe, R.D. and L.G. Chestnust. 1986. Oxidants and Asthmatics in Los Angeles: A Benefits
Analysis — Executive Summary. Prepared for U.S. Environmental Protection Agency, Office of
Policy Analysis. Prepared by Energy and Resource Consultants, Inc. Washington, DC. EPA-
230-09-86-018. March.
VII-96
-------
Chapter VII: Benefit-Cost Analysis
Samet, J.M., S.L. Zeger, I.E. Kelsall, J. Xu and L.S. Kalkstein. 1997. Air Pollution, Weather,
and Mortality in Philadelphia 1973-1988. Health Effects Institute. Cambridge, MA. March.
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.
Whittemore, A.S. and E.L. Korn. 1980. Asthma and Air Pollution in the Los Angeles Area. Am
J Public Health. 70: 687-696.
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.
VII-97
-------
Chapter VIM: Regulatory Flexibility Analysis
Chapter VIII: Regulatory Flexibility
This chapter presents our Final Regulatory Flexibility Analysis (FRFA) which evaluates
the impacts of our Tier 2 and gasoline sulfur standards on small businesses. Prior to issuing our
proposal last May, we analyzed the potential impacts of our program on small businesses. As a
part of this analysis, we convened a Small Business Advocacy Review Panel, as required under
the Regulatory Flexibility Act as amended by the Small Business Regulatory Enforcement
Fairness Act of 1996 (SBREFA). Through the Panel process, we gathered advice and
recommendations from small entity representatives (SERs) who would be affected by our
proposed vehicle and fuel standards. The report of the Panel has been placed in the rulemaking
record.1 A month after our proposal was published in the Federal Register, we held four public
hearings to obtain feedback on it. The small business provisions of today's action reflect changes
to the proposed program based upon updated analyses as well as comments heard at the public
hearings and those submitted in writing during the comment period.
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
FRFA 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 RFA was amended by SBREFA to ensure that concerns regarding small entities are
adequately considered during the development of new regulations that affect them. Although we
are not required by the CAA to provide special treatment to small businesses, the RFA requires
us to carefully consider the economic impacts that our rules will have on small entities.
Specifically, the RFA requires us to determine, to the extent feasible, our rule's economic impact
on small entities, explore regulatory options for reducing any significant economic impact on a
substantial number of such entities, and explain our ultimate choice of regulatory approach.
vm-i
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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-refmery 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.
B. Description of Affected Entities
Our Tier 2/gasoline sulfur program will primarily affect manufacturers of LDVs, LDTs,
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
CFR Part 121). However, we have identified several companies within these industries that are
small businesses as defined by SBA. 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.
vm-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
NAICS1
Codes
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 corporate-wide to qualify as a SBA small business. Based on our small business
analysis, we believe that some small refiners will have greater difficulty than larger refiners in
vm-3
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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.
2. Small Petroleum Marketers
While refiners are more affected by our gasoline sulfur control program than any other
industry, some marketers of gasoline, many of which are small by SBA's definition1, may also be
impacted. However, this impact appears to be limited to new or expanded requirements for
reporting the sulfur content of gasoline samples.
3. Small Certifiers of Covered Vehicles
In addition to the major vehicle manufacturers, three distinct categories of businesses
relating to LDVs and LDTs exist that are covered by Tier 2 emission standards. Some
companies in each of these categories are small businesses according to SB A regulations.
Small Independent Commercial Importers
Independent Commercial Importers (ICIs) 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
'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.
vm-4
-------
Chapter VIM: Regulatory Flexibility Analysis
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 will 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
designated as SVMs. This status allows vehicle models to be certified under a slightly simpler
certification process. More stringent Tier 2 standards will be relatively more difficult for small
manufacturers to achieve than larger manufacturers because research and development resources
are more limited. Less than five current SVMs meet the SBA 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
PADDHI
PADDIV
Per-Gallon Cost
(cents/gallon)
1.47
2.58
Operating Cost
($million/year)
3
6
Capital Cost
($million/year)
14
28
Costs for a small refinery located in PADD n to produce 30 ppm gasoline would fall
between the costs for a refinery in PADD in 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
vm-5
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
The types and number of small entities to which the proposed rule would apply are
described in Table Vin-3, below.
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 15
Several Hundred
Approximately 15
We have estimated that small refiners produce approximately four percent of all gasoline
in the U.S., excluding California. In most cases, gasoline produced by small refiners is mixed
with substantial amounts of other gasoline prior to retail distribution (due to the nature of the
gasoline distribution system).
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 requiring 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 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
vm-6
-------
Chapter VIM: Regulatory Flexibility Analysis
difficult to adapt the present RFG and anti-dumping reports to include the information required
under today's proposed rule, we are requiring 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 is relatively short and will require only the minimum
information necessary to demonstrate compliance with the applicable sulfur standards. Parties
are 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
are also 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
Our Tier 2 emission standards and gasoline sulfur control regulations 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,
we are not aware of any area where the new regulations would duplicate, overlap, or conflict with
any existing federal, state, or local regulations.
G. Regulatory Alternatives
We considered a wide range of options and regulatory alternatives for providing small
businesses flexibility in complying with the Tier 2 vehicle emission and gasoline sulfur
standards. The regulatory flexibilities we are providing for small businesses are described below.
1. Small Refiner Interim Sulfur Standards
Upon careful review of the comments received on the proposal as well as the
recommendations of the Small Business Advocacy Review Panel, we have determined that
regulatory relief in the form of delayed compliance dates is appropriate to allow small refiners,
both foreign and domestic, to comply with our regulations without disproportionate burdens.
From 2004-2007, when other refiners must meet the 30/80 standards (or the standards listed in
Table IV.C-1 of the preamble if they are participating in our ABT program), refiners meeting the
corporate employee and capacity limits are allowed to comply with somewhat less stringent
requirements. These interim annual-average standards, beginning in 2004, for qualifying small
refiners are shown in Table VHI-4 below.
VIH-7
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table VIII-4. Temporary Gasoline Sulfur Requirements
for Small Refiners in 2004-2007
Refinery
Baseline Sulfur Level
(ppm)
Oto30
31 to 200
201 to 400
401 to 600
601 and above
Temporary Sulfur Standards
(vvm)
Average
30 ppm
Baseline Level
200 ppm
50% of baseline
300
Cap
300 ppm
300 ppm
300 ppm
Factor of 1.5 times the average standard
450
The cap standards for the first two "bins" of refineries (that is those with baseline sulfur
levels from zero to 30 and 31 to 200) have been relaxed somewhat based on comments that the
proposed standards for these two bins were more stringent than the options under discussion for
all other refiners. We believe that small refiners should be able to meet the average standards
without much, if any, change to their operations but the more lenient cap will give them some
flexibility for turnarounds or unexpected equipment shutdowns. In addition, small refiners may
also use credits or allotments in 2004-2007 to comply with their average standards.
Compliance with these standards is based on a refiner's demonstration that it meets our
specific small refiner criteria. Refiners who qualify as a small refiner under our definition must
establish a sulfur baseline for each of their participating refineries. Section IV.C.2. of the
preamble explains these requirements in more detail.
Based on comments received on the proposal, we are also making four other changes to
our small refiner program.
1. We are revising the employee number criterion.
2. We are adopting a cap on the corporate crude oil capacity for a refining company
to qualify as a small business under today's regulations.
3. We are allowing small refiners to use credits and allotments to meet their
average standard (as specified in Table Vni-1) in addition to allowing them to
generate credits (from 2000 through 2007) and allotments (2003 only).
4. We are requiring that small refiners expecting to apply for a hardship extension
establish a compliance plan to demonstrate their commitment to produce low
sulfur gasoline (described in subsection a below).
vm-8
-------
Chapter VIM: Regulatory Flexibility Analysis
In regard to the employee number criterion, we are modifying how the employee number
is determined, based on comments received from SBA. As mentioned above, our proposed
definition applied to any petroleum refining company having no more than 1,500 employees
throughout the corporation as of January 1, 1999. We selected that date to prevent companies
from "gaming" the system. However, as SBA pointed out in its comments, the Small Business
Act regulations specify that, where the number of employees is used as a size standard, as we
proposed for small refiners, size determination is based on the average number of employees for
all pay periods during the preceding 12 months. Since we intended to use SBA's size standard in
our proposal, we are incorporating that definition correctly in today's action. It is also worth
mentioning that SBA shares our concerns about preventing companies from gaming the system
and that it solved this problem specifically by using the average employment over 12 months. In
addition, the averaging concept was designed to properly address firms with seasonal
fluctuations, according to SBA.
Second, we're amending the small refiner definition to include a corporate capacity cap.
We believe such a corporate capacity limitation is necessary to ensure that only truly small
businesses benefit from the relaxed interim standards. Furthermore, we received many
comments we should adopt a threshold based on crude capacity as specified in the Clean Air Act
and used in past EPA fuel programs.
As proposed, small refiners will be permitted to generate and trade sulfur credits and
allotments if they reduce sulfur levels early in 2000-2003. In today's action, we are also allowing
them to generate credits in 2004-2007. Furthermore, they may use credits that they generate in
2000-2007 and/or purchase credits from another refinery to meet their average standard during
2004-2007. A small refiner may sell credits in 2004 and beyond provided they honor the credit
life provisions explained in the ABT section (Section IV.C. I.e., above) while at the same time
meeting the small refinery standards.
a. Extensions Beyond 2007 for Qualifying Small Refiners
We are also finalizing our proposed hardship relief provision for qualifying small refiners
with some added detail based on the feedback we received during the public comment period.
Beginning January 1, 2008, all small companies' refineries must meet the national sulfur
standard of 30 ppm on average and the 80 ppm cap, except small refineries that apply for and
receive an extension of their 2007 standards. An extension will provide a given small refinery up
to an additional two years to comply with the national standards. An extension must be
requested in writing and must specify the factors that demonstrate a significant economic
hardship to qualify the refinery for such an extension. Factors considered for an extension could
include, but are not limited to, the refinery's financial position; its efforts to procure necessary
equipment and to obtain design and engineering services and construction contractors; the
vm-9
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
availability of desulfurization equipment, and any other relevant factors.
In order for us to consider an extension, a refiner must submit a detailed request for an
extension by January 1, 2007 demonstrating that it has made best efforts to obtain necessary
financing, and must provide detailed information regarding any lack of success in obtaining
financing. In addition, the refiner must meet the compliance plan requirements described below.
If we determine that the refiner has made the best efforts possible to achieve compliance with the
national standards by January 1, 2008, but has been unsuccessful for reasons beyond its control,
we will consider granting the hardship extension initially for the 2008 averaging period. If
further relief is appropriate for good reasons, we will consider a further extension through the
2009 averaging period but in no case will this relief be provided unless the refiner can
demonstrate conclusively that it has financing in place and that it will be able to complete
construction and meet the national gasoline sulfur standards no later than December 31, 2009.
b. Compliance Plans for Demonstrating a Commitment to Produce Low Sulfur
Gasoline
This final rule includes a compliance plan provision for those refiners who may seek a
hardship extension of their approved interim standards. This provision requires that those
refiners with approved interim standards who seek a hardship extension must submit a series of
reports to EPA discussing and describing their progress toward producing gasoline that meets the
30/80 ppm standards by January 1, 2008. We expect that small refiners will need to begin
preparations to meet the national standards in 2008 by 2004. However, we understand that the
potential exists for some small refiners to face additional hardship circumstances that will
warrant more time to meet the standards. For this reason, we have adopted provisions (see
above) allowing refiners subject to the interim standards to petition us and make a showing that
additional time is needed to meet the national standards. To properly evaluate hardship
applications, we are requiring demonstrations of good faith efforts towards assessing the
economic feasibility, along with the business and technical practicality of ultimately producing
low sulfur gasoline. Such progress reports must be submitted for a refiner to receive
consideration in any future determinations regarding hardship extensions. However, these
reports are not required from refiners who will not be seeking a hardship extension.
By June 1, 2004, such refiners would need to submit preliminary information in the form
of a report outlining its time line for compliance and a project plan discussing areas such as
permits, engineering plans (e.g., design and construction), and capital commitments for making
the necessary modifications to produce low sulfur gasoline. Documents showing activities and
progress in these areas should be provided if available.
By no later than June 1, 2005, these small refiners would need to submit a report to us
VHI-10
-------
Chapter VIM: Regulatory Flexibility Analysis
stating in detail progress to date based on their time line and project plan. This should include
copies of approved permits for construction of the equipment, contracts for design and
construction, and any available evidence of having secured the necessary financing to complete
the required construction. If any difficulties in meeting this requirement are anticipated, the
refiner must submit a detailed report of all efforts to date and the factors that may cause delay,
including costs, specification of engineering or other design work still needed and reasons for
delay, specification of equipment needed and any reasons for delay, potential equipment
suppliers and history of negotiations, and any other relevant information. If unavailability of
equipment is a factor, the report must include a discussion of other options considered, and the
reasons these other options are not feasible.
In addition, the small refiner would need to provide evidence by June 1, 2006, that on-site
construction has begun at its refinery(s) and that absent unforeseen circumstances or problems,
they will be producing complying gasoline (30/80 ppm) by January 1, 2008. While the
submission of these progress reports is evidence of a refiner's good faith efforts to comply by
2008, it does not bind the refiner to make gasoline in 2008. There are several reasons why a
refiner may choose to exit the gasoline-production business in 2008 that go beyond the low sulfur
gasoline requirement.
As a result of a refiner's efforts in moving toward compliance with the 2008 standards,
for market, economic, business, or technical reasons, the company could choose not to make
gasoline in 2008. Although we do not believe this will be the likely outcome for small refiners,
we cannot preclude it. Any refiner that makes such a determination in its progress reports will
have until 2008 to transition out of gasoline production, but will not be considered for a hardship
extension.
2. Small Certifiers of Covered Vehicles
During the SBREFA process and in the Tier 2 proposal, we discussed compliance
flexibilities for small certifiers of cars and light trucks who are subject to the Tier 2 standards.
The final rule includes several provisions that should ease the compliance burden for some, if not
all, of these companies.
Small certifiers will benefit from the provisions we are finalizing for all SVMs, not just
those that are also small entities according to the SBA definition (primarily Ids and alternative
fuel converters). One of these provisions allows SVMs to opt-out of the interim standards during
the phase-in years and to then comply with the final standards with 100 percent of their vehicles
at the end of the phase-in period. Another aspect of the final rule is a one-year hardship
provision for SVMs that will allow these manufacturers to apply for an additional year to meet
any of the 100 percent phase-in requirements. Finally, SVMs will be allowed to participate in the
vm-ii
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
averaging, banking, and trading (ABT) program. (Although our proposal did not include ICIs in
the ABT program, the final rule includes them, albeit with slightly different requirements
because of the unique nature of their business). See Section V of the Preamble of this rule for
more information on the treatment of small certifiers.
VHI-12
-------
Chapter VIM: Regulatory Flexibility Analysis
Chapter VUI 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.
vm-13
-------
Appendix A. Summary of Emission Inventories for 47 States and
for Four Urban Areas
In this appendix, emissions are summarized for the 47 States and for four urban
areas: New York, Chicago, Atlanta, and Charlotte. VOC, NOx, SOx, PM10, and PM2.5
emissions are summarized in Tables A-3 to A-7. Abbreviations used in these tables are
written out in Table A-2. The emissions presented here are mass emissions inventories.
These are distinct from the emissions input files that are used by the UAM-V and
REMSAD modeling systems, which generate air quality estimates from emissions and
meteorological inputs. Those emissions input files contain hour-by-hour emissions
produced under the specific meteorological conditions of the ozone episode being modeled.
The mass emissions files, on the other hand, which are summarized in this appendix,
represent average or typical conditions for the whole year and for an "ozone season day".
Detailed information on how both these sets of files were prepared is in the "Procedures for
Developing Base Year and Future Year Mass and Modeling Inventories for the Tier 2 Final
Rulemaking," September 1999, which is in Air Docket A-97-10.
As indicated in the titles of the tables, emissions of VOC and NOx are presented as
"annualized ozone season tons," which were generated by multiplying the "ozone season
day" variable in each mass emissions file by 365. For highway mobile sources, ozone
season day emissions represent a typical July day. The electrical generation mass emissions
files contain June, July, and August ozone season days; for consistency with the highway
mobile sources, the July day was chosen. The other mass emissions files contain only a
single ozone-season-day variable.
The urban areas coincide closely with ozone nonattainment or maintenance areas, as
described on EPA's Greenbook website
(http://www.epa.gov/oar/oaqps/greenbk/oindex.html). They depart slightly from the
nonattainment/maintenance definitions where partial counties are included in those
definitions. For the emissions analysis presented here, only whole counties are
summarized. In general, if only part of a county is included in a
nonattainment/maintenance area, the whole county is included in the emissions analysis.
However, if a county is in two nonattainment/maintenance areas, it is assigned to the
nonattainment/maintenance area containing most of the county's population. Counties
included in each urban area are listed in Table 1.
A-l
-------
Table A-l. Counties included in the emissions for four urban areas.
Chicago
Cook Co
Du Page Co
Grundy Co
Kane Co
Kendall Co
Lake Co
McHenry Co
Will Co
Lake Co
Porter Co
Atlanta
Cherokee Co
Clayton Co
Cobb Co
Coweta Co
De Kalb Co
Douglas Co
Fayette Co
Forsyth Co
Fulton Co
Gwinnett Co
Henry Co
Paul ding Co
Rockdale Co
IL
IL
IL
IL
IL
IL
IL
IL
IN
IN
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
GA
New York
Fairfield Co
Bergen Co
Essex Co
Hudson Co
Hunterdon Co
Middlesex Co
Monmouth Co
Morris Co
Ocean Co
Passaic Co
Somerset Co
Sussex Co
Union Co
Bronx Co
Kings Co
Nassau Co
New York Co
Orange Co
Queens Co
Richmond Co
Rockland Co
Suffolk Co
Westchester Co
Charlotte
Gaston Co
Mecklenburs Co
CT
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NJ
NY
NY
NY
NY
NY
NY
NY
NY
NY
NY
NC
NC
A-2
-------
Table A-2. Full description of abbreviated entries in the emissions tables in this appendix.
Table Entry
Full description
Stationary and Area
S & A w/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Stationary and area sources including natural and
miscellaneous categories.
Stationary and area sources without the natural and
miscellaneous categories
Nonroad mobile sources.
Heavy-duty highway vehicles and motorcycles.
Light-duty highway vehicles.
Emissions of all highway vehicles
The total inventory without the natural and
miscellaneous categories.
The percentage of the total inventory that light-duty
highway vehicles emit
The percentage that light-duty highway vehicles emit of
the total inventory without the natural and
miscellaneous categories.
A-3
-------
Table A-3. 47-State non-attainment area, FRM inventory, annualized ozone season tons.
VOC and NOx are in annualized ozone season tons. SOx and PM are in true annual tons.
47 -State VOC
Category
Stationary and Area
S & A w/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
%LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
11,948,272 11,030,732 13,632,729
11,094,679 10,028,585 12,608,048
3,536,201 2,487,618 2,218,832
537,939 370,071 473,224
3,577,079 2,026,945 2,108,765
4,115,018 2,397,016 2,581,988
19,599,492 15,915,366 18,433,549
18,745,898 14,913,219 17,408,868
18.251 12.736 11.440
19.082 13.592 12.113
2007
1 1 ,030,732
10,028,585
2,487,618
365,537
1 ,883,438
2,248,975
15,767,326
14,765,178
1 1 .945
12.756
Control Case
% Red. 2030
13,632,729
12,608,048
2,218,832
470,697
7.080 1,707,797
2,178,494
0.930 18,030,055
0.993 17,005,374
9.472
10.043
% Red.
19.014
2.189
2.318
47-State NOx
Category
Stationary and Area
S & A w/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
%LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
11,768,720 8,690,766 8,765,165
1 1 ,483,283 8,279,298 8,339,423
6,215,200 5,950,291 5,675,190
2,372,476 1 ,484,071 1 ,286,233
3,908,147 3,095,698 3,704,747
6,280,624 4,579,769 4,990,979
24,264,544 19,220,826 19,431,335
23,979,107 18,809,358 19,005,593
16.106 16.106 19.066
16.298 16.458 19.493
2007
8,690,766
8,279,298
5,950,291
1 ,452,462
2,239,227
3,691 ,688
18,332,746
17,921,277
12.214
12.495
Control Case
% Red. 2030
8,765,165
8,339,423
5,675,190
1 ,251 ,626
27.667 909,196
2,160,821
4.620 16,601,177
4.721 16,175,435
5.477
5.621
% Red.
75.459
14.565
14.891
47-State SOx
Category
Stationary and Area
S & A w/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
%LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
18,374,124 16,599,944 15,172,033
18,362,566 16,585,319 15,156,406
923,392 1,132,670 1,589,760
82,226 103,036 166,550
197,754 216,626 313,998
279,980 319,662 480,548
19,577,496 18,052,276 17,242,341
19,565,938 18,037,651 17,226,714
1.010 1.200 1.821
1.011 1.201 1.823
2007
16,599,944
16,585,319
1,111,538
92,700
22,847
115,547
17,827,028
17,812,404
0.128
0.128
Control Case
% Red. 2030
15,172,033
15,156,406
1,562,831
149,592
89.453 32,982
182,574
1.248 16,917,439
1.249 16,901,811
0.195
0.195
% Red.
89.496
1.884
1.886
A-4
-------
Table A-3. 47-State non-attainment area, FRM inventory, annualized ozone season tons.
VOC and NOx are in annualized ozone season tons. SOx and PM are in true annual tons.
47-State PM10
Category
Stationary and Area
S & A w/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
%LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
30,941,836 30,917,199 32,270,152
2,319,354 2,244,054 2,530,559
428,101 479,433 657,057
160,473 87,063 106,758
92,160 97,269 137,077
252,633 184,332 243,835
31,622,569 31,580,964 33,171,043
3,000,087 2,907,819 3,431,450
0.291 0.308 0.413
3.072 3.345 3.995
2007
30,917,199
2,244,054
479,433
86,107
74,574
160,682
31,557,314
2,884,169
0.236
2.586
Control Case
% Red. 2030
32,270,152
2,530,559
657,057
105,603
23.332 103,169
208,773
0.075 33,135,981
0.813 3,396,388
0.311
3.038
% Red.
24.736
0.106
1.022
47-State PM2.5
Category
Stationary and Area
S & A w/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
%LD of Total
% LD of Total w/o Nat. & Misc.
1996
7,233,601
1 ,61 1 ,450
379,532
140,393
57,452
197,846
7,810,978
2,188,827
0.736
2.625
Base Case
2007
7,665,679
1 ,736,330
426,976
72,544
56,939
129,484
8,222,138
2,292,789
0.693
2.483
2030
8,420,947
2,074,096
589,326
86,436
80,818
167,253
9,177,527
2,830,675
0.881
2.855
2007
7,665,679
1 ,736,330
426,976
71,783
35,340
107,123
8,199,778
2,270,429
0.431
1.557
Control
% Red.
37.933
0.272
0.975
Case
2030
8,420,947
2,074,096
589,326
85,451
48,388
133,839
9,144,112
2,797,261
0.529
1.730
% Red.
40.127
0.364
1.180
A-5
-------
Table A-4. Atlanta non-attainment area, FRM inventory, annualized ozone season tons.
VOC and NOx are in annualized ozone season tons. SOx and PM are in true annual tons.
Atlanta VOC
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
105,393 103,128 137,039
101,924 99,710 133,614
34,776 22,828 23,391
10,535 6,598 8,725
85,071 26,719 28,751
95,606 33,317 37,477
235,775 159,273 197,906
232,305 155,855 194,481
36.081 16.775 14.528
36.620 17.143 14.784
2007
103,128
99,710
22,828
6,485
23,964
30,449
156,405
152,987
15.322
15.664
Control Case
% Red. 2030
137,039
133,614
23,391
8,658
10.308 20,157
28,815
1.800 189,244
1.840 185,820
10.651
10.847
% Red.
29.893
4.377
4.454
Atlanta NOx
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
41,959 31,275 29,887
41 ,309 30,485 29,093
59,541 58,261 57,388
35,900 25,048 23,404
80,446 58,592 75,723
116,346 83,639 99,126
217,846 173,176 186,401
217,196 172,386 185,607
36.928 33.834 40.623
37.038 33.989 40.797
2007
31,275
30,485
58,261
24,339
40,605
64,945
154,481
153,691
26.285
26.420
Control Case
% Red. 2030
29,887
29,093
57,388
22,573
30.698 13,372
35,945
10.795 123,220
10.845 122,426
10.852
10.923
% Red.
82.340
33.895
34.040
Atlanta SOx
Category
Stationary and Area
S & A w/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
49,955 81,482 75,655
49,871 81,397 75,570
10,726 14,056 21,056
1,260 1,721 2,973
4,079 5,566 8,591
5,339 7,287 11,564
66,021 102,825 108,275
65,937 102,740 108,190
6.179 5.413 7.934
6.187 5.417 7.941
2007
81,482
81 ,397
13,848
1,500
504
2,004
97,334
97,249
0.518
0.519
Control Case
% Red. 2030
75,655
75,570
20,765
2,586
90.936 778
3,364
5.340 99,784
5.345 99,699
0.780
0.780
% Red.
90.945
7.842
7.848
A-6
-------
Table A-4. Atlanta non-attainment area, FRM inventory, annualized ozone season tons.
VOC and NOx are in annualized ozone season tons. SOx and PM are in true annual tons.
Atlanta PM10
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
218,941 239,672 285,950
7,933 8,745 10,076
5,808 6,945 10,028
2,474 1 ,530 1 ,877
1,947 2,147 3,204
4,420 3,677 5,080
229,169 250,295 301,058
18,162 19,368 25,185
0.849 0.858 1 .064
10.719 11.086 12.721
2007
239,672
8,745
6,945
1,509
1,602
3,111
249,729
18,802
0.641
8.520
Control Case
% Red. 2030
285,950
10,076
10,028
1,852
25.391 2,355
4,207
0.226 300,185
2.922 24,312
0.785
9.688
% Red.
26.484
0.290
3.467
Atlanta PM2.5
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007
47,873
6,738
5,183
2,160
1,219
3,379
56,435
15,300
2.160
7.966
54,594
7,956
6,222
1,278
1,254
2,532
63,347
16,710
1.980
7.505
2030
68,015
9,803
9,041
1,516
1,899
3,415
80,472
22,259
2.359
8.530
2007
54,594
7,956
6,222
1,262
740
2,001
62,817
16,179
1.177
4.571
Control
% Red.
41 .025
0.838
3.176
Case
2030
68,015
9,803
9,041
1,496
1,079
2,574
79,631
21,418
1.355
5.037
% Red.
43.177
1.045
3.777
A-7
-------
Table A-5. Charlotte non-attainment area, FRM inventory.
VOC and NOx are in annualized ozone season tons. SOx and PM are in true annual tons.
Charlotte VOC
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
34,288 33,411 44,570
32,956 32,087 43,234
11,807 7,463 7,844
1,413 1,255 1,672
12,535 7,506 7,177
13,949 8,761 8,849
60,044 49,635 61,263
58,712 48,311 59,927
20.877 15.122 11.715
21.351 15.537 11.977
2007
33,411
32,087
7,463
1,235
6,723
7,958
48,832
47,508
13.768
14.151
Control Case
% Red. 2030
44,570
43,234
7,844
1,661
10.431 4,873
6,534
1.618 58,947
1.662 57,611
8.267
8.459
% Red.
32.103
3.780
3.864
Charlotte NOx
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
27,348 17,022 18,358
27,141 16,814 18,150
18,720 18,526 18,954
6,530 4,510 4,569
14,151 12,406 15,143
20,681 16,915 19,712
66,749 52,463 57,024
66,541 52,256 56,816
21.200 23.646 26.556
21.266 23.740 26.653
2007
17,022
16,814
18,526
4,388
8,632
13,019
48,567
48,360
17.773
17.849
Control Case
% Red. 2030
18,358
18,150
18,954
4,418
30.422 2,719
7,137
7.426 44,450
7.456 44,242
6.117
6.146
% Red.
82.044
22.051
22.132
Charlotte SOx
Category
Stationary and Area
S & A w/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
56,355 71,844 61,130
56,355 71,844 61,130
2,848 3,899 6,100
258 340 579
859 1,147 1,716
1,117 1,487 2,295
60,320 77,230 69,526
60,320 77,230 69,526
1.424 1.485 2.468
1.424 1.485 2.468
2007
71,844
71 ,844
3,833
296
104
400
76,078
76,077
0.137
0.137
Control Case
% Red. 2030
61,130
61,130
6,010
503
90.936 155
658
1.492 67,798
1.492 67,798
0.229
0.229
% Red.
90.945
2.484
2.484
A-8
-------
Table A-5. Charlotte non-attainment area, FRM inventory.
VOC and NOx are in annualized ozone season tons. SOx and PM are in true annual tons.
Charlotte PM10
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
43,590 52,887 69,383
2,858 2,928 3,029
1 ,486 1 ,841 2,850
429 238 367
386 442 643
815 680 1,010
45,891 55,407 73,244
5,159 5,449 6,889
0.842 0.797 0.878
7.490 8.106 9.330
2007
52,887
2,928
1,841
235
330
565
55,292
5,334
0.596
6.183
Control Case
% Red. 2030
69,383
3,029
2,850
362
25.342 471
834
0.208 73,067
2.110 6,713
0.645
7.023
% Red.
26.651
0.241
2.558
Charlotte PM2.5
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
10,963 13,389 17,563
2,186 2,535 2,996
1 ,355 1 ,679 2,603
373 1 95 297
238 258 379
611 453 676
12,929 15,520 20,842
4,152 4,666 6,275
1.843 1.660 1.819
5.739 5.520 6.043
2007
13,389
2,535
1,679
193
152
345
15,412
4,558
0.987
3.336
Control Case
% Red. 2030
17,563
2,996
2,603
293
40.962 216
508
0.696 20,674
2.316 6,107
1.042
3.529
% Red.
43.167
0.805
2.675
A-9
-------
Table A-6. Chicago non-attainment area, FRM inventory.
VOC and NOx are in annualized ozone season tons. SOx and PM are in true annual tons.
Chicago VOC
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
409,386 384,462 475,490
408,536 383,598 474,604
97,946 64,689 61,501
10,479 6,507 7,887
91,524 26,895 28,181
102,003 33,402 36,069
609,334 482,553 573,059
608,484 481,689 572,174
15.020 5.574 4.918
15.041 5.584 4.925
2007
384,462
383,598
64,689
6,452
25,029
31,480
480,631
479,767
5.207
5.217
Control Case
% Red. 2030
475,490
474,604
61,501
7,837
6.940 20,250
28,086
0.398 565,077
0.399 564,191
3.584
3.589
% Red.
28.145
1.393
1.395
Chicago NOx
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
257,347 195,557 209,579
257,243 195,453 209,475
155,390 149,783 144,669
43,072 27,950 25,858
107,714 62,072 73,150
150,786 90,023 99,008
563,523 435,363 453,257
563,419 435,259 453,152
19.114 14.258 16.139
19.118 14.261 16.142
2007
195,557
195,453
149,783
27,605
47,980
75,585
420,925
420,821
1 1 .399
1 1 .402
Control Case
% Red. 2030
209,579
209,475
144,669
25,332
22.703 15,414
40,746
3.316 394,994
3.317 394,890
3.902
3.903
% Red.
78.928
12.854
12.857
Chicago SOx
Category
Stationary and Area
S & A w/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
283,888 331,893 342,165
283,888 331,892 342,164
17,378 23,142 35,152
1 ,662 1 ,885 2,989
5,983 2,910 4,082
7,645 4,796 7,071
308,911 359,830 384,388
308,911 359,829 384,388
1 .937 0.809 1 .062
1 .937 0.809 1 .062
2007
331 ,893
331 ,892
22,555
1,791
669
2,460
356,908
356,908
0.187
0.187
Control Case
% Red. 2030
342,165
342,164
34,386
2,838
77.021 937
3,775
0.812 380,326
0.812 380,326
0.246
0.246
% Red.
77.041
1.057
1.057
A-10
-------
Table A-6. Chicago non-attainment area, FRM inventory.
VOC and NOx are in annualized ozone season tons. SOx and PM are in true annual tons.
Chicago PM10
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
259,792 274,604 314,362
62,898 65,902 76,380
9,188 11,025 16,309
2,902 1 ,508 2,078
2,671 2,820 3,867
5,573 4,329 5,945
274,553 289,957 336,616
77,658 81,256 98,634
0.973 0.973 1.149
3.440 3.471 3.921
2007
274,604
65,902
1 1 ,025
1,496
2,370
3,866
289,494
80,793
0.819
2.933
Control Case
% Red. 2030
314,362
76,380
16,309
2,060
15.970 3,200
5,260
0.160 335,931
0.569 97,949
0.952
3.267
% Red.
17.265
0.204
0.695
Chicago PM2.5
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
88,998 94,550 109,471
44,137 46,377 53,563
8,474 10,093 14,938
2,525 1 ,238 1 ,677
1,634 1,603 2,212
4,159 2,841 3,889
101,631 107,484 128,297
56,770 59,311 72,389
1 .608 1 .491 1 .724
2.878 2.703 3.056
2007
94,550
46,377
10,093
1,228
1,184
2,412
107,055
58,882
1.106
2.011
Control Case
% Red. 2030
109,471
53,563
14,938
1,662
26.129 1,590
3,252
0.399 127,660
0.723 71,753
1.245
2.216
% Red.
28.123
0.496
0.880
A-11
-------
Table A-7. New York non-attainment area, FRM inventory.
VOC and NOx are in annualized ozone season tons. SOx and PM are in true annual tons.
New York VOC
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
517,305 459,828 529,668
514,215 456,656 526,465
151,282 96,183 92,176
16,893 10,794 12,683
149,763 37,125 40,442
166,656 47,918 53,125
835,243 603,930 674,968
832,153 600,758 671,765
17.930 6.147 5.992
17.997 6.180 6.020
2007
459,828
456,656
96,183
10,707
34,591
45,297
601,309
598,137
5.753
5.783
Control Case
% Red. 2030
529,668
526,465
92,176
12,612
6.826 29,008
41,619
0.434 663,463
0.436 660,260
4.372
4.393
% Red.
28.273
1.705
1.713
New York NOx
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
165,367 122,713 120,317
165,006 122,298 119,900
218,910 220,952 231,676
77,788 48,094 42,383
177,258 88,878 113,645
255,047 136,973 156,028
639,324 480,638 508,021
638,963 480,223 507,604
27.726 18.492 22.370
27.742 18.508 22.389
2007
122,713
122,298
220,952
47,471
68,366
115,837
459,502
459,087
14.878
14.892
Control Case
% Red. 2030
120,317
119,900
231 ,676
41,487
23.079 24,089
65,576
4.397 417,569
4.401 417,152
5.769
5.775
% Red.
78.803
17.805
17.819
New York SOx
Category
Stationary and Area
S & A w/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
323,294 257,903 265,644
323,290 257,900 265,640
70,567 77,015 97,854
2,925 3,257 5,008
10,288 5,044 6,829
13,213 8,301 11,836
407,074 343,220 375,335
407,071 343,216 375,331
2.527 1.470 1.819
2.527 1.470 1.819
2007
257,903
257,900
76,065
3,087
1,115
4,202
338,170
338,166
0.330
0.330
Control Case
% Red. 2030
265,644
265,640
96,559
4,743
77.891 1,503
6,245
1.471 368,448
1.471 368,444
0.408
0.408
% Red.
77.996
1.835
1.835
A-12
-------
Table A-7. New York non-attainment area, FRM inventory.
VOC and NOx are in annualized ozone season tons. SOx and PM are in true annual tons.
New York PM10
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
466,339 497,623 575,429
49,070 42,324 44,825
19,127 22,525 32,535
5,045 2,603 3,455
4,614 4,723 6,215
9,659 7,326 9,670
495,126 527,474 617,634
77,857 72,175 87,031
0.932 0.895 1 .006
5.926 6.544 7.142
2007
497,623
42,324
22,525
2,581
3,952
6,534
526,681
71,382
0.750
5.537
Control Case
% Red. 2030
575,429
44,825
32,535
3,426
16.317 5,120
8,546
0.150 616,510
1 .098 85,906
0.830
5.960
% Red.
17.622
0.182
1.292
New York PM2.5
Category
Stationary and Area
S & Aw/o Nat. and Misc.
Nonroad
HD & MC Highway
LD Highway
Total Highway
Total Inventory
Total without Natural and Misc.
% LD of Total
% LD of Total w/o Nat. & Misc.
Base Case
1996 2007 2030
128,173 131,512 151,036
36,042 30,438 32,426
15,308 19,599 28,781
4,386 2,141 2,789
2,824 2,700 3,569
7,211 4,841 6,359
150,692 155,951 186,176
58,561 54,877 67,565
1.874 1.731 1.917
4.823 4.920 5.283
2007
131,512
30,438
19,599
2,123
1,980
4,103
155,213
54,139
1.276
3.657
Control Case
% Red. 2030
151,036
32,426
28,781
2,764
26.665 2,546
5,310
0.473 185,127
1.345 66,517
1.375
3.828
% Red.
28.671
0.563
1.551
A-13
-------
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 a significant, 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.11
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), on
average, by 40 percent for NMHC and 134 percent for NOx emissions compared to 30 ppm
sulfur fuel. New data generated since the NPRM on similar LEVs and ULEVs show that when
these vehicles were driven on high sulfur (330 ppm) fuel for a few thousand miles, the NMHC
and NOx emission increase due to high sulfur fuel increased by 149 percent and 47 percent,
respectively. In other words, instead of the previous estimated 40 percent and 134 percent
increases in NMHC and NOx emissions, respectively, the newer estimates would be 100 percent
and 197 percent, respectively.
In this section, we are concerned with the impact of sulfur under a broader range of
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
irreversibility). The third describes the vehicle testing programs which have attempted to
measure the irreversibility of the sulfur impact. The fourth presents criteria for evaluating the
wide range of sulfur irreversibility data which are available. Finally, the fifth describes EPA's
projections of the degree of sulfur irreversibility for various vehicle types (e.g., Tier 1 vehicle,
LEVs, and Tier 2 vehicles).
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.
B-l
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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.
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. In the NPRM we estimated that, based on data from test
programs conducted by EPA and the automotive and oil industries, LEV and ULEV vehicles
could experience, on average, a 40 percent increase in NMHC and 134 percent increase in NOx
emissions when operated on 330 ppm sulfur fuel (our estimate in the NPRM of the current
national average sulfur level) compared to 30 ppm sulfur fuel. New data generated since the
NPRM on similar LEVs and ULEVs show that when these vehicles were driven on high sulfur
(330 ppm) fuel for a few thousand miles, the NMHC and NOx emission increase due to high
sulfur fuel increased by 149 percent and 47 percent, respectively. In other words, instead of the
previous estimated 40 percent and 134 percent increases in NMHC and NOx emissions,
respectively, more realistic estimates would be 100 percent and 197 percent, respectively. The
calculations resulting in these sensitivity values are described below in this section. Also, new
data generated since the NPRM for late model LEV and ULEV vehicles that meet the Federal
and California supplemental federal test procedure (SFTP) standards and also have very low FTP
emission levels, indicate that, on average, a 51 percent increase in NMHC and a 242 percent
increase in NOx emissions when operated for a short period of time on 330 ppm compared to 30
ppm could be realized.
Table A-l lists new sulfur sensitivity data for several late model LEV and ULEV vehicles
that meet the Federal and California supplemental federal test procedure (SFTP) standards and
also have very low FTP emission levels when sulfur is increased from 30 ppm to 350 ppm.
B-2
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Table B-l. Sulfur Sensitivity: New Data Between 30 ppm and 350 ppm
Vehicle
DaimlerChrysler Caravan
Ford Expedition
Ford Windstar
FordF-150
Average
NMHC
87%
81%
12%
30%
51%
NOx
333%
42%
238%
249%
242%
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. We examined the sulfur sensitivity for vehicles in our sulfur
database that were at or below Tier 2 levels with those that were above Tier 2 standards. What
we found was that those vehicles meeting Tier 2 standards showed a higher degree of sensitivity
to sulfur than those with higher emission levels. However, at a 95 percent confidence level, there
was no statistical difference in sulfur sensitivity between the vehicles at or below Tier 2 emission
standards and those above Tier 2 standards. Thus, we have only projected that Tier 2 vehicles
will be just as sensitive as LEV and ULEV LDVs and not more so. Therefore, these should be
considered conservative estimates for Tier 2 vehicles.
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"2 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
2 "EPA Staff Paper on Gasoline Sulfur Issues," U.S. EPA, May 1998, EPA420-R-98-005
B-3
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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.
A second concern about the estimates of sulfur sensitivity used in the NPRM is that all of
the vehicles in the test programs used to develop the NPRM projections of sulfur sensitivities
were only exposed to high sulfur fuel for a few miles of driving prior to emission testing. This is
referred to as "short-term" sulfur exposure. 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. The short-term exposure in
the test programs typically consisted of only running several emission tests (FTP or LA4). Since
each FTP is approximately 18 miles in length, short-term exposure usually amounted to just
under 100 miles of operation, all of which was in a controlled laboratory environment.
To address this concern, API and EPA each conducted test programs testing a combined
total of six light-duty vehicles for sulfur sensitivity after short-term and long-term exposure to
sulfur. The vehicles were randomly selected by both API and EPA. The long-term exposure
consisted of between 1,500 and 3,000 miles of in-use operation over urban, rural and highway
roads. Two of the vehicles were 1999 models, while the other four were 1998 models. All six
were either LEV or ULEV vehicles. Three of the vehicles were equipped with catalyst systems
aged to either 50,000 or 100,000 miles. The other three vehicles had low mileage catalyst
systems aged to only 4,000 miles. Table A-2 describes the vehicles tested:
B-4
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Table B-2. Vehicles Tested After Short-Term vs. Long-Term
Exposure to Higher Sulfur Fuel
Make/Model
Model Year
Emission Level
Catalyst Aging (miles)
EPA Test Program
Honda Accord
Chevrolet Cavalier
1999
1999
ULEV
LEV
50,000
50,000
API Test Program
Nissan Altima
Ford Taurus
Honda Accord
Toyota Avalon
1998
1998
1998
1998
LEV
LEV
ULEV
LEV
100,000
4,000
4,000
4,000
All of the vehicles were tested for short-term exposure first. Each vehicle was FTP
baseline3 tested on low sulfur fuel (30 or 40 ppm). The number of tests used to establish the
baseline varied from two to four. The vehicles were then tested with the high sulfur fuel (EPA
at 350 ppm, API at 540 ppm). Again the number of tests ranged from two to four. Upon
completion of the short-term program, each vehicle was preconditioned several times with the
EPEFE sulfur purge cycle prior to beginning the long-term exposure program. Only the 1999
Honda Accord of the EPA test program reestablished a new baseline for the long-term
program-the other vehicles used the original short-term baseline. All of the vehicles were then
operated on the road with the high sulfur fuel from anywhere between 1,500 to 3,000 miles and
tested over the FTP to establish long-term high sulfur emission levels.
Sulfur sensitivity was determined by calculating the percent increase in average emissions
with the high sulfur fuel compared to the average emissions with the low sulfur fuel. For NOx
emissions, all six vehicles showed greater sulfur sensitivity after long-term exposure to high
sulfur fuel than after short-term exposure. For NMHC emissions, all of the vehicles except the
Altima and Avalon experienced greater sensitivity for long-term exposure. Only the Altima
showed lower sulfur sensitivity for CO emissions after long-term exposure. Table A-3 lists the
sulfur sensitivity results for all six vehicles:
3 Prior to baseline testing, each vehicle was preconditioned with a purge cycle based on the European
Programme for Emission, Fuels, and Engine Technologies (EPEFE) sulfur purge cycle, which uses a series of ten
wide-open throttle accelerations from 30 to 70 mph, in order to ensure there was no sulfur contamination prior to
baseline testing.
B-5
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
B-6
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Table B-3. Vehicle-by-Vehicle Short-Term vs. Long-Term Sulfur Sensitivity
Vehicle
Accord
(EPA
Vehicle)
Cavalier
Altima
Taurus
Accord
(API
Cat.
Age
50K
50K
50K
50K
100K
100K
4K
4K
4K
Sulfur
Aging
Short
Long
Short
Long
Short
Long
Short
Long
Short
Sulfur
Level
30 ppm
350
ppm
30 ppm
350
ppm
30 ppm
350
ppm
30 ppm
350
ppm
40 ppm
540
ppm
40 ppm
540
ppm
40 ppm
540
ppm
40 ppm
540
ppm
40 ppm
540
Tailpipe Emissions (g/mi)
NMHC
0.031
0.035
0.033
0.040
0.070
0.105
0.0.70
0.223
0.041
0.059
0.041
0.057
0.033
0.051
0.033
0.073
0.029
0.032
CO
0.351
0.478
0.330
0.731
1.778
4.048
1.778
7.224
0.788
1.058
0.788
0.987
0.522
0.832
0.522
1.310
0.285
0.299
NOx
0.092
0.155
0.090
0.234
0.068
0.303
0.068
0.324
0.061
0.112
0.061
0.132
0.075
0.101
0.075
0.117
0.100
0.192
Sulfur Sensitivity (%)
NMHC
12.0
21.7
49.3
216.6
43.9
39.0
54.5
121.2
10.3
CO
36.3
121.1
127.7
306.4
34.3
25.3
59.4
151.0
4.9
NOx
69.4
158.5
347.0
411.8
83.6
116.4
34.7
56.0
92.0
B-7
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Vehicle
Aval on
Cat.
Age
4K
4K
4K
Sulfur
Aging
Long
Short
Long
Sulfur
Level
40 ppm
540
ppm
40 ppm
540
ppm
40 ppm
540
oom
Tailpipe Emissions (g/mi)
NMHC
0.029
0.041
0.040
0.061
0.040
0.060
CO
0.285
0.465
0.406
0.541
0.406
0.734
NOx
0.100
0.245
0.068
0.116
0.068
0.142
Sulfur Sensitivity (%)
NMHC
41.4
52.5
50.0
CO
63.2
33.3
80.8
NOx
145.0
70.6
108.8
In order to quantify the difference between short-term and long-term exposure, we
averaged the low and high sulfur emissions for each pollutant for all of the vehicles and
determined a straight linear fleet average emissions for both low sulfur and high sulfur fuels.
The ratio of the long-term sensitivity to the short-term sulfur sensitivity was then determined. As
can be seen in table A-4, the percent increases from short-term to long-term are quite significant,
especially for NMHC emissions. The three vehicles with catalysts aged to 50,000 or 100,000
miles had, on average, long-term sensitivities greater than the three vehicles with 4,000 mile
catalysts. Therefore, the effects of long-term exposure to sulfur presented here may be
underestimated.
Table B-4. Percent Difference Between Short-Term vs. Long-Term Sulfur Sensitivity
Average
Short-Term
Long-Term
Sulfur Sensitivity (%)
NMHC
40.2
100.3
CO
75.7
178.7
NOx
111.3
163.4
Ratio of the Sensitivities (long-term to
short-term , in %)
NMHC
149.2
CO
136.0
NOx
46.8
To test whether this observed increase in sulfur sensitivity was statistically valid, we
calculated the ratio of short-term sulfur sensitivity (in percent) to long-term sulfur sensitivity (in
percent) for each vehicle. We then calculated the average and standard deviation of these ratios
and calculated 90percent and 95percent confidence intervals. At a 95percent confidence level,
B-8
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
the lower limits of the confidence intervals for NMHC and NOx pollutants exceeded 1.0. This
indicates that at least a 95percent confidence exists that the long-term sulfur sensitivity exceeds
that for short-term exposure. The same was true for CO emissions at a 90percent confidence
level.
We multiplied the short term sulfur sensitivities from the larger vehicle database by the
ratio of the long to short term sensitivities from the 6 vehicle database. This resulted in a sulfur
sensitivity of 100 percent for NMHC and 197 percent for NOx emissions when measured at 330
ppm fuel sulfur compared to 30 ppm.
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
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.
4 "EPA Staff Paper on Gasoline Sulfur Issues," U.S. EPA, May 1998, EPA420-R-98-005.
B-9
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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
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
B-10
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
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 no information on 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. 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
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.
Results from our Tier 2 technology assessment program indicate that there will be trade-
offs between NMOG and NOx control in order to meet Tier 2 emission standards, especially for
larger vehicles. For example, significant reductions in NOx can be achieved by improved EGR
strategies that don't necessarily rely on improvements to the catalyst, whereas reductions in
NMOG may rely more heavily on strategies to reduce catalyst light-off time as well as catalyst
light-off performance. Since sulfur doesn't affect emissions coming out of the engine, an
emission control strategy that focuses less on catalyst performance may not experience sulfur
B-ll
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
problems as readily as systems that depend more on the catalyst. Since these trade-offs will be
very model-specific, it is very difficult to determine the impact emission control strategies
needed to meet strict Tier 2 emission standards will have on sulfur reversibility tolerance.
Thus, the two changes in emission control design necessary to reverse sulfur, 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 evaluates the available sulfur irreversibility data from numerous sulfur
irreversibility test programs.
C. Results of Sulfur Irreversibility Test Programs
We have received data from seven test programs which evaluate the irreversibility of
sulfur's impact on vehicle emissions. These programs are summarized in the following seven
sections.
1. Pre-SFTP LEVs
All of the data generated for the NPRM was for Tier 1, LEV, and ULEV vehicles that
were not designed to meet the Federal, or California SFTP standards. The potential effect of the
SFTP standards on sulfur reversibility has already been discussed above. Therefore, we are
going to divide the data from the various test programs into two categories: pre-SFTP LEVs and
SFTP-compliant LEV and Tier 2 vehicles. The following is a summary of the programs in the
pre-SFTP category. Two of the programs were discussed in the RIA for the proposed rule.
There are also two new programs that were run after the NPRM.
a. Coordinating Research Council (CRC) Sulfur Irreversibility Program
The CRC sulfur irreversibility 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.
B-12
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Table B-5. 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.
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 irreversibility test program.
Table B-6. Sulfur Irreversibility: CRC Test Program (%)
B-13
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Vehicle
Manufac
Ford
Ford
Honda
Nissan
Toyota
Suzuki
Fleet
Estimate
Models
Taurus
Escort
Civic
Sentra
Camry
Metro
NMHC
Purge Cycle
LA4
31.0
0.0
6.0
1.0
0.0
0.0
3.0
US06
17.0
0.0
1.0
0.0
2.0
0.0
0.0
NOx
Purge Cycle
LA4
30.0
5.0
4.0
15.0
50.0
14.0
16.0
US06
5.0
0.0
3.0
12.0
0.0
13.0
5.0
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 irreversibility for NMHC, especially with the LA4
purge cycle (31 percent). The results for NOx indicate that with the LA4 purge cycle, the
average level of irreversibility is 16 percent with the Toyota Camry having irreversibility as high
as 50 percent. When using the US06 purge cycle, NOx emissions were far more reversible with
an average irreversibility of 5 percent. The Nissan Sentra and Suzuki Metro showed almost the
exact same level of irreversibility with both purge cycles.
b. American Petroleum Institute Sulfur Irreversibility Program
The API program5 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
5 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-14
-------
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. This will
underestimate reversibility since other programs have shown that emissions on low sulfur fuel
after exposure to high sulfur fuel continue to decrease after two tests. Thus, statistically
speaking, the API program is weaker than the CRC program. Examination of individual
emission test results shows significant variability occurred.
Table B-7 lists the vehicle tested in the API program.
Table B-7. 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)
Number of
Cylinders
6
6
6
4
8
8
6
Engine
Displacement
3.0L
2.3L
3.0L
2.4L
4.6L
4.6L
4.3L
B-15
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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
technology. This assumption will be discussed further below.
Table B-8 shows the sulfur irreversibility emission results for all of the vehicles when
tested with low mileage (4,000 mile) catalysts.
Table B-8. Sulfur Irreversibility: 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
0.0
76.9
28.6
0.0
0.0
53.7
33.3
32.4
US06
n/a *
0.0
57.1
n/a*
19.4
40.0
0.0
54.1
NOx
Purge Cycle
LA4
3.8
21.7
47.9
0.0
15.5
5.0
29.7
16.7
US06
n/a
2.2
0.0
n/a
28.2
0.0
0.0
7.7
* Vehicle not tested with US06 purge cycle.
The most obvious difference between the irreversibilities measured by API and those
found by CRC is that API's average NMHC irreversibility rate when using the LA4 as a purge
cycle is 32 percent, while CRC's average NMHC reversibility rate shows nearly full reversibility
at three percent irreversibility. The measured NOx irreversibilities (with the LA4 purge cycle)
were almost identical in the two programs, 17 percent for API compared to 16 percent for CRC.
However, it should be pointed out that API only performed two tests on low sulfur fuel.
API found much lower irreversibility using the US06 cycle as a purge cycle for NOx (7.7
B-16
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
percent). However, the opposite was true for NMHC (54.1 percent). This 54.1 percent
irreversibility is considerably higher 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 irreversibilities 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
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
irreversibility.
Table B-9 shows measured irreversibility 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 irreversibility with the US06 cycle. Irreversibility 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.
B-17
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table B-9. Sulfur Irreversibility: 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
32.5
n/a
n/a
n/a
32.5
US06
0.0
0.0
25.0
54.5
12.0
NOx
Purge Cycle
LA4
2.4
n/a
n/a
n/a
2.4
US06
0.0
5.5
0.0
0.0
0.0
Short-Term Exposure
NMHC
Purge Cycle
LA4
0.0
76.9
28.6
0.0
6.0
US06
n/a *
0.0
57.1
19.4
31.0
NOx
Purge cycle
LA4
3.8
21.7
47.9
15.5
22.3
US06
n/a
2.2
0.0
28.2
11.6
Table B-9 shows measured irreversibility for vehicles with catalysts bench aged to
represent 100,000 miles 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. Irreversibility of
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.
B-18
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Table B-10. Sulfur Irreversibility: 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
0.0
15.1
0.0
0.0
0.0
0.0
11.3
21.1
12.7
14.6
10.8
12.7
2,000 Mile Exposure to High Sulfur Fuel
98 Altima
n/a
0.0
n/a
6.1
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 irreversibility increased from 3.8 percent with the low mileage
catalyst to 11.3 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 reversible for both
NMHC and NOx emissions with the LA4 purge cycle. The second Altima, which had a 100,000
mile catalyst showed more irreversibility, only 15.1 percent for NMHC emissions and 21.1
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.
Table B-ll. Sulfur Sensitivity: API Test Program
Low Mileage Catalysts, Short-Term Exposure to High Sulfur Fuel (g/mi)
NMHC
NOx
B-19
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
FTP Test
Sulfur
Level
Sulfur
Exposure
Vehicle
Taurus
Accord
Aval on
Gr. Marq.
Average
Altima
30 ppm
—
540 ppm
Short-term
540 ppm
1,000 Mile
30 ppm
—
5 40 ppm
Short-term
5 40 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
c. Ford Sulfur Irreversibility Program
Ford tested two vehicles, a 1999 LEV Taurus and a 1999 LEV Explorer. Both vehicles
were equipped with 4K mile aged catalysts. Neither vehicle was designed to meet SFTP
emission standards. The vehicles were initially tested over the FTP cycle using low sulfur fuel
(35 ppm) to establish a baseline. A total of three FTPs were run. This was followed by testing
another three FTPs on a high sulfur fuel (450 ppm). Then, the fuel was switched back to the low
sulfur fuel and the vehicles ran a LA4 cycle immediately followed by a FTP. The LA4 cycle was
used to purge sulfur from the catalyst. Ford ran between three to five of these LA4/FTP
combinations for each vehicle. Ford repeated the entire procedure with the US06 cycle in place
of the LA4 cycle as a purge cycle. The following table lists the results of the Ford sulfur
irreversibility test program.
B-20
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Table B-12. Sulfur Irreversibility: Ford Test Program (%)
Model
Taurus
Explorer
Vehicle Type
LDV
LDT3
NMHC
Purge Cycle
LA4
12.0
91.0
US06
0.0
70.0
NOx
Purge cycle
LA4
7.0
n/a
US06
0.0
n/a
The Ford results were somewhat sporadic. The Taurus was mostly reversible over the
LA4 and fully reversible over the US06, whereas the Explorer was highly irreversible over the
LA4 cycle. They were unable to perform any tests over the US06 for the Explorer. Our biggest
concerns with the Ford data was that the vehicles were mistakenly equipped and tested with 4K
catalyst systems instead of 50K or 100K aged catalysts, and the data showed an enormous
amount of variability.
d. EPA Sulfur Irreversibility Test Program
After publication of the NPRM, we tested two 1999 LDVs which were supplied to us by
their manufacturer. One was a LEV Chevrolet Cavalier and the other was a ULEV Honda
Accord. Both vehicles were equipped 50K aged catalysts. The vehicles were initially tested over
the FTP cycle using low sulfur fuel (30 ppm) to establish a baseline. We tested until the
emission results stabilized, typically three to four FTPs (approximately 60 to 80 miles). This was
followed by FTP testing on a high sulfur fuel (350 ppm). Again, tests were run until emission
levels were stable. Upon completion of high sulfur testing, the fuel was switched back to the low
sulfur fuel and the vehicles ran a combination of LA4 cycle immediately followed by a FTP.
These tests were also run until emissions stabilized, not exceeding eight LA4 + FTP cycles.
Since neither vehicle was SFTP-compliant, we did not perform any tests with the US06 cycle as
a purge cycle.
In addition to short-term sulfur testing, we also tested the vehicles with long-term
exposure to sulfur. After completion of the short-term testing, the vehicles were driven over
several EPEFE purge cycles to ensure that any sulfur from the short-term testing was removed
from the catalysts. The vehicles were baseline tested on low sulfur fuel as before and then driven
for 2,000 to 3,000 miles on the road with high sulfur (350 ppm) fuel. FTP tests were performed
at 500 mile intervals. Upon the completion of high sulfur road aging, the vehicles ran a
combination of LA4 cycle immediately followed by a FTP, exactly as in the short-term testing.
Again, no US06 purge cycle testing was performed.
B-21
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
The following Table lists the results from our test program.
Table B-13. Sulfur Irreversibility: EPA Test Program,
Short-Term and Long-Term Exposure (%)
Short-Term
Model
Cavalier
Accord
Vehicle Type
LEV LDV
ULEV LDV
Long-Term
Cavalier
Accord
LEV LDV
ULEV LDV
NMHC
Purge Cycle
LA4
67.0
0.0
0.0
0.0
US06
n/a
n/a
n/a
n/a
NOx
Purge cycle
LA4
42.0
26.0
27.0
14.0
US06
n/a
n/a
n/a
n/a
The results for short-term exposure suggest that both vehicles were highly irreversible,
especially the Cavalier. The NMHC results for the Accord are most likely an anomaly since the
vehicle mistakenly had only one test performed on high sulfur fuel. The results for long-term
sulfur exposure are counter-intuitive, since NMHC emissions were fully reversible compared to
the low level of reversibility for the vehicles when tested after short-term exposure. NOx
emissions were slightly more reversible than for short-term exposure. However, even with long-
term exposure to sulfur, NOx emissions were still only partially reversible.
e. ATL Sulfur Irreversibility Program
ATL, under contract for us, tested two vehicles, a 1999 LEV Ford Windstar mini-van and
a 1999 LEV Ford Taurus. Both vehicles were procured from a rental agency in California and
had approximately 50K miles. Thus, they were equipped with catalysts which had been aged
with 50,000 miles of in-use driving, albeit at higher annual mileage rates than typical in-use
vehicles. Both vehicles had low emissions, especially the Taurus which had emissions below
Tier 2 levels.
ATL used the exact same test procedure as us for our in-house testing. The vehicles were
initially tested over the FTP cycle using low sulfur fuel (30 ppm) to establish a baseline. Tests
were run until emission results stabilized, typically three to four FTPs (approximately 60 to 80
miles). This was followed by FTP testing on a high sulfur fuel (350 ppm). Again, tests were run
until emission levels were stable. Upon completion of high sulfur testing, the fuel was switched
back to the low sulfur fuel and the vehicles ran a combination of LA4 cycle immediately
B-22
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
followed by a FTP. These tests were also run until emissions stabilized, not exceeding eight LA4
+ FTP cycles. Because we were anticipating new powertrain control modules (PCM) from Ford
that were to be equipped with SFTP-compliant calibrations, we did not run either vehicle over
the US06 cycle. The following Table summarizes the results.
Table B-14. Sulfur Irreversibility: ATL Test Program
Model
Taurus
Windstar
Vehicle Type
LDV
LDT2
NMHC
Purge Cycle
LA4
30.0
26.0
US06
n/a
n/a
NOx
Purge cycle
LA4
34.0
29.0
US06
n/a
n/a
As can be seen, both of these vehicles were partially reversible. The level of
irreversibility for both vehicles falls almost exactly in the middle of the data spread for all of the
pre-SFTP vehicles.
f. Irreversibility for Long-Term Sulfur Exposure
In section A., we discussed the effect long-term exposure to sulfur has on sulfur
sensitivity. We found that, based on a sample of six pre-SFTP LEV vehicles, long-term exposure
to high sulfur fuel resulted in an additional sensitivity in emissions to sulfur of 149 percent for
NMHC and 48 percent for NOx, above the original emission sensitivity levels when comparing
emissions from a fuel sulfur level of 30 ppm to a fuel sulfur level of 330 ppm. For example, if
baseline emissions were 0.10 g/mi NOx and high sulfur emissions were 0.15 g/mi NOx after
short term exposure, then high sulfur emissions would be about 0.175 g/mi NOx after long term
exposure. However, the data from these six vehicles indicates that when these vehicles were
operated again on 30 ppm sulfur fuel with an LA4 purge cycle, the extra emissions sensitivity
resulting from long-term exposure was completely recovered. In other words, all of the vehicles
showed the same or lower emissions on low sulfur fuel after long-term exposure to high sulfur
fuel as they did after short term exposure to high sulfur fuel. This would suggest that after long-
term exposure to sulfur, emissions are capable of recovering to short-term levels with only
moderate FTP-type driving. We are projecting this phenomenon to occur for both pre-SFTP and
SFTP-compliant vehicles, though this data is available only for pre-SFTP vehicles. Thus, there is
some uncertainty in applying it to SFTP-compliant Tier 2 vehicles, as well. However, the same
is true for the data showing a larger sulfur sensitivity after long-term exposure to high sulfur fuel.
2.
SFTP-compliant LEV and Tier 2 vehicles
B-23
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
The following three sections describe sulfur irreversibility test programs that utilize
SFTP-compliant, low emitting LEVs and prototype Tier 2 vehicles. All of these programs
occurred after publication of the NPRM. We are also quantifying irreversibility for NMHC and
NOx emissions together instead of independently, because per our discussion above and our own
experience developing the emission control strategy for the Expedition discussed below,
sensitivity and irreversibility of either pollutant appears to be very dependent on the particular
strategy chosen to reduce these emissions (particularly engine calibration and catalyst loading of
precious metals and oxygen storage).
a. DaimlerChrysler Sulfur Irreversibility Program
DaimlerChrysler tested a prototype "Tier 2-like" 3.3L Dodge Caravan that met SFTP
emission standards and was equipped with a 100K aged catalyst. DaimlerChrysler tested the
vehicle over a test procedure very similar to the short-term portion of our sulfur irreversibility
test program for the Cavalier and Accord. The only differences were that they tested a high
sulfur fuel level of 450 ppm instead of 350 ppm, and they performed reversibility testing with the
REP05 cycle in lieu of the US06 cycle for sulfur purging prior to FTP tests after operation on
high sulfur fuel. The following Table lists results for the Caravan.
Table B-15. Sulfur Irreversibility: DaimlerChrysler Test Program (%)
Vehicle
Caravan
NMHC
Purge Cycle
LA4
18.0
REP05
39.0
NOx
Purge Cycle
LA4
29.0
REP05
5.0
NMHC + NOx
Purge Cycle
LA4
27.0
REP05
11.0
The Caravan was partially reversible for NMHC, NOx, and NMHC + NOx. The vehicle
was more reversible after REP05 operation than LA4 operation for NOx, but not NMHC
emissions. NMHC + NOx emissions indicate significant irreversibility for LA4 operation and
moderate irreversibility even after REP05 operation.
b. EPA Sulfur Reversibility Program
In addition to the two pre-SFTP LEVs vehicles that we tested, we also tested a 1999 Ford
Expedition SUV from our Tier 2 technology demonstration program. The vehicle was equipped
with a 50K aged catalyst system. We modified the Expedition such that it met Tier 2 intermediate
useful life emissions standards (bin 4 - 0.075 g/mi NMOG, 0.05 g/mi NOx) as well as federal and
California SFTP standards with reasonable margins of safety. The modifications made consisted
of calibration changes and an advanced catalyst system (see Chapter IV. A of the RIA for a
B-24
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
detailed description of our Tier 2 test work with the Expedition). The vehicle was initially tested
over the FTP cycle using low sulfur fuel (30 ppm) to establish a baseline. We tested until the
emission results stabilized, typically three to four FTPs. This was followed by FTP testing on a
high sulfur fuel (350 ppm). Again, tests were run until emission levels were stable. Upon
completion of high sulfur testing, the fuel was switched back to the low sulfur fuel and the
vehicles ran a combination of LA4 cycle immediately followed by a FTP. These tests were also
run until emissions stabilized, not exceeding eight LA4 + FTP cycles. The entire test procedure
was repeated with the REP05 cycle in place of the LA4 cycle as a purge cycle. The following
Table lists the results for the Expedition.
Table B-16. Sulfur Irreversibility: EPA Test Program (%)
Vehicle
Expedition
NMHC
Purge Cycle
LA4
78.0
REP05
91.0
NOx
Purge Cycle
LA4
21.0
REP05
0.0
NMHC + NOx
Purge Cycle
LA4
65.0
REP05
70.0
The Expedition was partially reversible for NOx over LA4 conditions and fully reversible
over the REP05. However, the most interesting observation was that it was highly irreversible
for NMHC over all driving conditions, but especially for aggressive. The results for NMHC ,
although more severe for the Expedition, were similar to the Caravan in that both vehicles were
more irreversible over the REP05 than the LA4 cycle. Looking at NMHC + NOx results,
indicate that the Expedition was highly irreversible for all driving conditions.
c. ATL Sulfur Reversibility Program
ATL, under contract for us, tested two vehicles, a 1999 LEV Ford Windstar mini-van and
a 1999 LEV Ford F-150 pick-up truck. Both vehicles were procured from a rental agency in
California and had approximately 50K miles. Thus, they were equipped with catalysts which had
been aged with 50,000 miles of in-use driving, albeit at higher annual mileage rates than typical
in-use vehicles. Both vehicles were then equipped with new powertrain control modules (PCM)
with calibrations modified to meet SFTP emission standards, courtesy of Ford. Both vehicles
had low emissions, but not at Tier 2 emission levels. A third vehicle (Taurus) was procured by
ATL, but a SFTP-compliant PCM was not available, so it was not tested with the other two
vehicles.
ATL used the exact same test procedure as us for our in-house testing. The vehicles were
initially tested over the FTP cycle using low sulfur fuel (30 ppm) to establish a baseline. Tests
were run until emission results stabilized, typically three to four FTPs. This was followed by
B-25
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
FTP testing on a high sulfur fuel (350 ppm). Again, tests were run until emission levels were
stable. Upon completion of high sulfur testing, the fuel was switched back to the low sulfur fuel
and the vehicles ran a combination of LA4 cycle immediately followed by a FTP. These tests
were also run until emissions stabilized, not exceeding eight LA4 + FTP cycles. The entire test
procedure was repeated with the REP05 cycle in place of the LA4 cycle as a purge cycle. The
following Table summarizes the results.
Table B-17. Sulfur Irreversibility: ATL Test Program (%)
Vehicle
Windstar
F-150
NMHC
Purge Cycle
LA4
0.0
60.0
REP05
0.0
25.0
NOx
Purge Cycle
LA4
35.0
13.0
REP05
0.0
7.0
NMHC + NOx
Purge Cycle
LA4
31.0
18.0
REP05
0.0
9.0
For the LA4 cycle, both vehicles experienced considerable irreversibility for NMHC,
NOx, and NMHC + NOx, except for NMHC emissions with the Windstar. With the REP05
cycle, the Windstar was fully reversible for all pollutants, while the F-150 was still partially
reversible.
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 much of the available data. As
mentioned in the previous section, the sulfur reversibility testing would ideally have 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.
EPA established a number of criteria for evaluating the available data in order to project
likely levels of in-use sulfur irreversibility. 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 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.
B-26
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Development of the subsequent criteria are more complex, because the issues of SFTP
compliance, vehicles emissions performance 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. Four of the vehicles were equipped with "prototype" SFTP-compliant calibrations,
meaning they met both the Tier 1 and LEV US06 standards. Of these four vehicles, only one
(EPA Ford Expedition) met the SC03 standards as well as the US06 standards.
As discussed earlier, there will be considerable trade-offs in NMOG and NOx control in
order to meet strict Tier 2 emission standards. There can be considerable uncertainty associated
with balancing these trade-offs at very low emissions levels if the vehicle is periodically operated
on high sulfur fuels, making the ability to remove sulfur from the catalyst highly uncertain. For
example, a given catalyst today may be fully reversible for one pollutant and only partially
reversible for another. However, because of the trade-off in NMOG and NOx performance, the
modifications necessary to get that vehicle to meet both emission standards may result in the
opposite effect for reversibility; i.e., full reversibility for NMOG and partial reversibility for
NOx. Therefore, a very important criterion in conjunction with SFTP compliance is LEV
emission performance for pre-SFTP vehicles and "Tier 2-like" emission performance for SFTP-
compliant Tier 2 vehicles.
Likewise, for the bulk of the data which is for pre-SFTP LEVs, only the LA4 and US06
driving cycles were used in the test programs. 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 from the REP05 cycle.
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-use during aggressive driving.
The four SFTP-compliant vehicles used the REP05 cycle in lieu of the US06 cycle. The
REP05 cycle was developed by EPA to be is representative of aggressive driving that occurs
outside the LA4 or FTP cycle. All but one of the aggressive driving segments found in the US06
cycle were taken from the REP05. While each segment of the US06 cycle was taken from actual
in-use driving, the timing and combination of these segments is not representative of in-use
driving in the way REP05 is representative. As with the US06 cycle, however, the length of time
that the catalyst is exposed to both high temperatures and rich conditions could be much higher
than would occur in the real world, resulting in the removal of more sulfur than would be
removed in-use even during aggressive driving. Thus, while it is likely that typical vehicles will
B-27
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
experience the reversibility which was measured after driving over the REP05 cycle, we cannot
be certain that this is the case.
As mentioned in Section B, meeting the SFTP standards will require 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. 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:
Pre-SFTP Vehicles
1. Only use data from vehicles with aged catalyst systems,
2. Emphasize data from vehicles with LEV emission levels, and
3. Use data where the sulfur was purged using either the LA4 or US06 cycle.
SFTP-Compliant LEV and Tier 2 vehicles
1. Only use data from vehicles with aged catalyst systems,
2. Emphasize data from vehicles with emission levels appropriate for the LEV and Tier 2
standards,
3. Use data from vehicles that were modified to meet SFTP standards, and
4. Use data where the sulfur was purged using either the LA4 or REP05 cycle.
E. Projected Levels of Sulfur Irreversibility In-Use
1. Pre-SFTP Vehicles
Applying the first criterion developed in Section D. results in the retention of the CRC
and EPA data (Tables B-2 and B-9), 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 irreversibility of
sulfur effects for pre-SFTP vehicles do not rely on the API data except that in Table B-6.
B-28
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
Table B-18. Pre-SFTP Sulfur Irreversibility: Summary of Relevant Test Programs (%)
Models
CRC (6 vehicles)
EPA (2 vehicles)
ATL (2 vehicles)
API (2-3 vehicles)
Fleet Estimate
NMHC
Purge Cycle
LA4
0.0
75.0
28.0
0.0
14.0
US06
0.0
n/a
n/a
14.0
0.0
NOx
Purge Cycle
LA4
16.0
38.0
30.0
0.0
20.0
US06
4.0
n/a
n/a
12.0
5.0
For pre-SFTP vehicles (Tier 0, Tier 1 and NLEV), as described above, we decided to
utilize reversibility measurements using both the LA4 and US06 driving cycles. We decided to
project reversibility for these vehicles by taking the mid-point of the LA4 and US06 values for
NMHC and NOx, respectively. Therefore, for these vehicles, using the average of these test
results, we project that NMHC emissions are almost fully reversible at four percent
irreversibility, while NOx emissions are 12 percent irreversible.
2. SFTP-Compliant LEV and Tier 2 vehicles
The DaimlerChrysler, EPA, and ATL data all met the first criterion of aged catalyst
systems. The DaimlerChrysler vehicle was equipped with a 100K aged catalyst, while the
vehicles from the other two programs had 50K aged catalyst systems. As for emission
performance, all four vehicles were LEVs, meeting the LEV standards with considerable margins
of safely. The Expedition and Caravan were both modified in an attempt to meet Tier 2
standards. Only the Expedition actually met both Tier 2 NOx and NMOG emission standards
with a typical margin of safety. The Caravan was close to Tier 2 levels, but exceeded the NOx
standard. The F-150 was also close to Tier 2 levels, while the Windstar exceeded Tier 2 levels
by a significant margin. It should be noted, however, that all four vehicles are LDTs which were
certified to substantially higher emission standards than the Tier 2 standards in their baseline
configurations. Also, because these vehicles are LDTs, their catalyst temperatures are typically
higher than LDVs, which is good for removing sulfur from the catalyst.
All four vehicles from the three programs met the federal and California SFTP emission
standards for the aggressive driving portion (US06) with considerable margin. The Expedition
also met the SFTP emission standards for the air-conditioning portion of the test (SC03) as well.
B-29
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
As discussed above, trade-offs between NMOG and NOx control in order to meet Tier 2
standards combined with periodic operation on high sulfur fuel will result in uncertainty in the
ability to remove sulfur from the catalyst. Therefore, for SFTP-compliant Tier 2 vehicles, we
feel that the most appropriate way to analyze irreversibility is to evaluate NMHC and NOx
together (i.e., NMHC + NOx), rather than separately. The following Table shows the results for
the four vehicles from the DaimlerChrysler, EPA, and ATL test programs.
Table B-19. SFTP-Compliant Sulfur Irreversibility:
Summary of Relevant Test Programs (%)
Daimler-Chrysler Caravan
Ford Expedition
Ford Windstar
Ford F- 150
Average
Average w/ lower weights for
Caravan and Windstar
NMHC + NOx
FTP Purge
27.0
64.8
31.4
18.3
29.4
29.4
REP05 Purge
11.0
69.6
0.0
9.1
10.9
16.6
As can be seen, NMHC + NOx irreversibility is generally much lower after high speed,
aggressive driving than after more average city driving. As previously discussed, the REP05
cycle represents the top 28 percent of driving with the highest speeds and hardest accelerations.
Thus, most people will drive like the REP05 cycle at least part of the time; however, it is not
clear whether occasional driving like the REP05 cycle will provide all of the reversibility
enhancement that was provided by the entire REP05 cycle performed in sequence.
There is also still significant variability between the irreversibility of individual vehicles,
with the Expedition showing the highest irreversibility by far. This is significant for determining
a SFTP-compliant Tier 2 irreversibility estimate, because the Expedition is the only vehicle
which complies with both the NMHC and NOx Tier 2 standards with a reasonable amount of
headroom. The Windstar (@ 0.12 g/mi NOx) and the Caravan (@ 0.09 g/mi) exceed the Tier 2
NOx standard by significant margins, while the F-150 truck had NOx emissions just slightly
above the 0.07 g/mi standard.
Therefore, to determine an irreversibility estimate for SFTP-compliant Tier 2 vehicles, we
had to account for the differences in various vehicle's compliance with the Tier 2 standards. We
B-30
-------
Appendix B: Irreversibility of Sulfur's Emission Impact
accomplished this by reducing the weight given to the Windstar and Caravan. We recalculated
the average irreversibility by reducing the weight assigned to the Windstar and Caravan to one-
fourth and one-half of a vehicle, respectively. As shown in the table above, this has no impact on
the average irreversibility after FTP driving, but reduces that after REP05 driving modestly. The
REP05 cycle represents about 28 percent of all in-use driving. Due to roadway limitations, no
one can drive like the REP05 cycle 100% of the time (i.e., residential areas, congested streets,
etc.). Therefore, it is reasonable to project that the majority of vehicles are driven in this way at
least part of the time. However, it is likely that some vehicles are never or very rarely driven this
aggressively. Therefore, we project that roughly 75 percent of vehicles are driven regularly like
the REP05 cycle and that 25 percent are not. Thus, we decided to weigh the irreversibility after
FTP driving by 25 percent and that after REP05 driving by 75 percent. This results in an average
NMHC+NOx irreversibility of 20 percent.
We also wanted to focus on the irreversibility of the Expedition, since it was the only
vehicle meeting the Tier 2 standards with adequate headroom. The Expedition had irreversibility
levels of 65-70 percent over the two driving cycles. Since the lower irreversibility was seen over
the FTP, that figure (65 percent) seems reasonable for an estimate based solely on the
Expedition. Therefore, for Tier 2 vehicles, we project that irreversibility of NMHC+NOx
emissions will fall somewhere between the low level of 20 percent, based on all four vehicles,
and 65 percent based on the Expedition. For emission modeling and cost effective analyses, we
decided to use a midpoint estimate of 42.5 percent for Tier 2 vehicles. As for SFTP-compliant
LEV vehicles, we decided to use a straight average of the four vehicles weighing the
irreversibility after FTP driving by 25 percent and that after REP05 driving by 75 percent, similar
to what we did for Tier 2 vehicles. This resulted in an average NMHC+NOx irreversibility of 15
percent. As mentioned above, we project that Tier 0 and Tier 1 vehicles are fully reversible.
B-31
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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 Report Number M6.FUL.001
B-32
-------
Appendix C: Refinery Energy and Global Warming Impacts and Emissions
Appendix C: Refinery Energy and Global Warming
Impacts and Emissions
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 in 1997.
We simplified our methodology here relative to that used for the cost analysis. We estimate the
energy impact of desulfurizing gasoline under two scenarios. The first assumes that refiners will
use the improved desulfurization technologies discussed in Chapter IV.B. This represents the
near term scenario, when refiners are projected to use a mixture of proven, improved, and
adsorption technologies. The second scenario assumes that refiners will use the advanced
adsorption technologies also discussed in that chapter. This represents the long term scenario,
when refiners are projected to use only the adsorption technologies. We use 1997 as the base
year for this analysis because that is the base year of the cost calculation, which is a basis for this
analysis, and because of the uncertainty in projecting U.S. global warming gas emissions in
future years.
To determine the percentage increase in refinery energy consumed by desulfurizing
gasoline, we first established the baseline energy consumption by PADD 3 refineries using 1994
Energy Information Administration data, which is the most recent energy consumption data
available. We project the baseline energy consumption of PADD 3 refineries from 1994 to 1997
using an estimated increase in energy consumption of 6 percent, which is based on 2.05 percent
increase in increased refinery throughput per year. This PADD 3 energy consumption calculation
is summarized below in Table V-46. The energy consumed by PADD 3 refineries in 1997 is
estimated to be about 1,500 trillion BTUs.
C-l
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table C-l. 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
0 MBbls
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
The total amount of energy consumed to desulfurize gasoline down to 30 ppm target is
calculated by adding up the fuel gas, steam and electricity, in terms of British thermal units
(BTUs) consumed, for the desulfurization unit, hydrogen production and octane makeup . First
we estimated the energy consumed running both the CDTECH and Octgain 220 processing units.
Consistent with how the cost of desulfurization was estimated for these improved technologies,
each desulfurization technology was presumed to handle half of PADD 3's desulfurization needs.
Then the energy consumed for recovering octane and producing hydrogen demand is calculated.
C-2
-------
Appendix C: Refinery Energy and Global Warming Impacts and Emissions
For both CDTECH and Octgain, we presumed that reformer capacity1 was available to make up
the octane lost from desulfurization. We accounted for the amount of hydrogen that the reformer
would produce and by subtracting the reformer production from the hydrotreater demand, we
estimated the amount of hydrogen which had to be provided by an existing hydrogen plant.
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. A summary of the estimated CDTECH and Octgain energy and hydrogen
demands in PADD 3 is summarized in Tables V-47 and V-48, respectively.
1 According to the 1996 API/NPRA survey of gasoline quality and refinery operations, spare reformer
capacity is available in the U.S. If a particular refiner has no spare reformer capacity, then the energy consumed and
global warming emissions emitted can then be assumed to be emitted by an expanded reformer, or other octane
generating units which likely consume less energy than the reformer due to their less severe operating conditions.
C-3
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table C-2. Estimated Annual Energy and Hydrogen Demand of CDTECH
Desulfurization Units for Half of PADD 3
CDTECH Utility
Demands
Electricity
Fuel Gas
Hydrogen
Reformer
Electricity
Fuel Gas
Steam
Hydrogen Plant
Fuel Gas
Electricity
Total
Process
Demand
0.44 KwH/Bbl
41250 Btu/Bbl
102 Scf/Bbl
2.6 KwH/Bbl
0.06 FOE/Bbl
94 Lb/Bbl
0.11 FOE/Bbl
1.69 KwH/Bbl
Yearly
Throughput
273 MMBbls
273 MMBbls
273 MMBbls
Hydrogen
12 MMBbls feed
12 MMBbls feed
12 MMBbls feed
Hydrogen
8. 5 MMBbls H2
8. 5 MMBbls H2
BTU Conversion
Factor
3.41 MBtu/KwH
-
-
Produced
3.41 MBtu/KwH
6 MMBtu/Bbl
0.809MBtu/Lb
Produced
6 MMBtu/Bbl
6 MMBtu/Bbl
Energy and
Hydrogen
Consumed
410MMMBtu
11,250
MMMBtu
27,800 MMScf
12,320 MMScf
105 MMMBtu
43 50 MMMBtu
11 00 MMMBtu
14, 180 MMScf
5680MMMBtu
50 MMMBtu
22,950
MMMBtu
C-4
-------
Appendix C: Refinery Energy and Global Warming Impacts and Emissions
Table C-3. Estimated Annual Energy and Hydrogen Demand of OCTGAIN
Desulfurization Units for Half of PADD 3
OCTGAIN 220
Utility Demands
Electricity
Fuel Gas
Steam
Hydrogen
Splitter
Electricity
Fuel Gas
Steam
Reformer
Electricity
Fuel Gas
Steam
Hydrogen Plant
Fuel Gas
Electricity
Total
Process
Demand
1.55KwH/Bbl
44500 Btu/Bbl
0 Lb/Bbl
130 Scf/Bbl
0.17KwH/Bbl
0 FOE/Bbl
45 Lb/Bbl
2.6 KwH/Bbl
0.06 FOE/Bbl
94 Lb/Bbl
0.11 FOE/Bbl
1.69 KwH/Bbl
Yearly
Throughput
213 MMBbls
213 MMBbls
213 MMBbls
213 MMBbls
273 MMBbls
273 MMBbls
273 MMBbls
Hydrogen
1.2 MMBbls
1.2 MMBbls
1.2 MMBbls
Hydrogen
13.5MMBblsH2
13.5MMBblsH2
BTU Conversion
Factor
3.41 MBtu/KwH
-
0.809 MBtu/Lb
-
-
3.41 MBtu/KwH
6 MM Btu/Bbl
0.809 MBtu/Lb
Produced
3.41 MBtu/KwH
6 MM Btu/Bbl
0.809 MBtu/Lb
Produced
6 MMBtu/Bbl
6 MMBtu/Bbl
Energy and
Hydrogen
Consumed
1440 MMBtu
9470 MMMBtu
OMMMBtu
27,700 MMScf
80 MMMBtu
OMMMBtu
5960 MMMBtu
1,1 40 MMScf
1 1 MMMBtu
445 MMMBtu
94 MMMBtu
26,600 MMScf
9040 MMMBtu
75 MMMBtu
26,600
MMMBtu
We used the same methodology to estimate the energy consumed for the adsorption
technologies by Black and Veatch and Phillips, with each presumed to desulfurize half the
gasoline pool. Those estimates are summarized in Tables V-49 and V-50.
C-5
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table C-4. Estimated Annual Energy and Hydrogen Demand of Black and Veatch
Desulfurization Units for Half of PADD 3
Black and
Veatch Utility
Demands
Electricity
Fuel Gas
Steam
Hydrogen
Hydrogen Plant
Fuel Gas
Electricity
Diesel
Hydrotreater
Total
Process
Demand
1.92KwH/Bbl
24100Btu/Bbl
0 Lb/Bbl
14 Scf/Bbl
O.llFOE/Bbl
1.69KwH/Bbl
Yearly
Throughput
272 MMBbls
272 MMBbls
272 MMBbls
272 MMBbls
Hydrogen
13.5MMBblsH2
13.5MMBblsH2
BTU Conversion
Factor
3.41 MBtu/KwH
-
0.809 MBtu/Lb
-
Produced
6 MMBtu/Bbl
6 MMBtu/Bbl
Energy and
Hydrogen
Consumed
1790MMBtu
6580 MMMBtu
OMMMBtu
3900 MMScf
3900 MMScf
1330 MMMBtu
1 1 MMMBtu
11 70 MMMBtu
10,900
MMMBtu
C-6
-------
Appendix C: Refinery Energy and Global Warming Impacts and Emissions
Table C-5. Estimated Annual Energy and Hydrogen Demand of Phillips
Desulfurization Units for Half of PADD 3
Phillips Utility
Demands
Electricity
Fuel Gas
Steam
Hydrogen
Reformer
Electricity
Fuel Gas
Steam
Hydrogen Plant
Fuel Gas
Electricity
Total
Process
Demand
0 KwH/Bbl
49400 Btu/Bbl
0 Lb/Bbl
70 Scf/Bbl
2.6 KwH/Bbl
0.06 FOE/Bbl
94 Lb/Bbl
0.11 FOE/Bbl
1.69 KwH/Bbl
Yearly
Throughput
272 MMBbls
272 MMBbls
272 MMBbls
272 MMBbls
Hydrogen
9.3 MMBbls
9.3 MMBbls
9.3 MMBbls
Hydrogen
13.5MMBblsH2
13.5MMBblsH2
BTU Conversion
Factor
3.41 MBtu/KwH
-
0.809 MBtu/Lb
-
Produced
3.41 MBtu/KwH
6 MM Btu/Bbl
0.809 MBtu/Lb
Produced
6 MMBtu/Bbl
6 MMBtu/Bbl
Energy and
Hydrogen
Consumed
1790MMBtu
6580 MMMBtu
OMMMBtu
19,100MMScf
8590 MMScf
83 MMMBtu
3350 MMMBtu
847 MMMBtu
15,750 MMScf
3 5 80 MMMBtu
30 MMMBtu
21,400
MMMBtu
As these tables show, the average increase in energy consumption for the improved
gasoline desulfurization technologies to meet a 30 ppm gasoline sulfur program, including the
energy needed to provide hydrogen and make up octane loss, is estimated to be about 50 trillion
BTUs based on 1997 volumes. This increase in energy use is about 3.2 percent of the baseline
PADD 3 refining industry energy consumption.
For the U.S. outside of California, the refining industry is estimated to consume 3000
trillion BTUs per year.2 Thus the increase in energy demand for the U.S. refining industry, based
on PADD 3 and using the 3.2 percent factor calculated above, is estimated to be about 96 trillion
2This estimate is based on the presumption that PADD 3 consumes 50 percent of the energy in the U.S.
outside of California.
C-7
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
BTUs per year.
In future years (i.e., after 2019), assuming that all FCC gasoline desulfurization is
switched over to the adsorption desulfurization technologies, PADD 3 refiners are expected to
consume 32 trillion BTU's over the baseline, which represents a 2.1 percent increase in PADD 3
refinery energy demand. Projecting this PADD 3 energy demand to the entire U.S., and based on
the adsorption technologies, we estimate that the U.S. refining industry would consume an
additional 63 trillion BTUs of energy to desulfurize gasoline based on a 1997 baseline.
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
refinery energy and, thus, is presumed to be the source of all refinery emissions of carbon
dioxide. The carbon dioxide emission factor is estimated to be 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).1 For simplicity, this analysis assumes that all BTUs consumed
in a refinery are produced by these fuel sources. On this basis, CO2 emissions from all U.S.
refineries would increase by 6.7 million tons per year in the 1997 base year based on the
improved desulfurization technologies, and CO2 emissions would increase by 4.2 million tons in
the 1997 base year based on the adsorption technologies.
The increase in CO2 emissions for installing improved desulfurization technologies is a
one-time step increase in CO2 emissions which represents 0.12 percent of the U.S. CO2
emissions inventory, which is 5.4 billion tons of CO2 per year in based on 1997 emissons.2 This
increase also represents about 9 percent of the total projected increase in U.S. CO2 emissions in a
single year, which is about 70 million tons per year in 1997. Based on our presumption that the
adsorption technologies will replace the mix of proven and improved desulfurization
technologies (expected to occur after 2019), worldwide CO2 emissions are projected to be 0.08
percent higher than the baseline due to the U.S. desulfurization program, which is 60 percent (or
40 percent less) of the total projected one-time increase in U.S. CO2 emissions based on the mix
of desulfurization technologies which we presume to be used at the outset of the program. The
increase in energy consumed and carbon dioxide emissions is summarized in Table V-51.
-------
Appendix C: Refinery Energy and Global Warming Impacts and Emissions
Table C- 6. Increase in Energy Consumed and Carbon Dioxide Emissions Due to
Desulfurizing Gasoline (1997 energy use and emissions)
Increase in Energy Consumed (trillion Btu)
Percent Increase in U.S. Refining Industry Energy
Consumed
Increase in CO2 Emissions (million tons)
One-Time Percent Increase in Yearly U.S. CO2
Emissions
Improved
Desulfurization
Technologies
96
3.2
6.7
0.12
Adsorption
Desulfurization
Technologies
63
2.1
4.2
0.08
The desulfurization of gasoline will increase emissions of conventional pollutants from
refineries. At a minimum, desulfurizing gasoline will require the addition of a naphtha
hydrotreater. However, refiners will need a source of hydrogen and make up the loss of octane.
A likely way for refiners to make up that lost octane is to increase the feed to an existing
reformer. In addition to the octane boost, the reformer also provides a source of hydrogen. If the
amount of hydrogen provided by the reformer is insufficient, the refiner is assumed to either
produce more hydrogen from an existing hydrogen plant, or install another hydrogen plant to
provide the balance of hydrogen.
Each of these units require heaters to provide heat for reactions. Heaters emit oxide of
nitrogen (NOx) emissions, volatile organic carbon (VOC) emissions, carbon monoxide (CO),
particulates and sulfur oxide (SOx). VOC emissions are also emitted from leaks from pipes,
valves, pumps etc. of refinery units which process petroleum.
To estimate the increase in emissions from desulfurizing gasoline, we contacted Mobil
Oil to obtain emissions information from Mobil on their Octgain unit in operation at their Joliet,
Illinois refinery. Mobil provided information to us for a 25,000 barrel per day Octgain unit.3 The
information provided was expressed in aggregate emissions per year, emissions per pound of fuel
gas consumed, and emissions per hour. We used the hourly emissions rate with the hourly heater
energy consumption as a divisor. This ratio expressed the emissions rate in terms of the energy
consumed in the heater, which allowed us to base our emissions analysis on the PADD 3 energy
analysis summarized above.
Using the PADD 3 analysis above, we are presuming that the emissions per energy
consumed ratio, which we calculate in the Table C-7 below, is the same for all desulfurization
technologies. This assumption is reasonable since heaters should be essentially the same for any
C-9
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
refinery processing unit. Except for VOC emissions, we also assume that the emissions from the
other units that provide octane and hydrogen can be estimated by the emission factors in Table C-
7. Since hydrogen plants normally only react natural gas, and since the octane produced from the
reformer is likely from a unit already present in the refinery, we do not expect any additional
VOC emissions from these units. Consistent with that premise, we only account for the VOC
emissions from the hydrotreating unit. The VOC emission values reflect our presumption that
refiners have adopted, or will have adopted, leak detection and maintenance programs for
significant VOC emission control, which is consistent with the environmental programs in place,
or expected to be in place, in most refineries. We expressed the NOx emissions as a range; the
larger number reflects Mobil's estimate of NOx emissions from a heater with conventional
burners, and the lower number is our estimate of a heater with ultra-low burners (which
incorporates a 75 percent reductions in emissions). The information provided by Mobil Oil and
our recalculated emissions values are summarized in the following table.
Table C-7. Pollutant Emissions from an Octgain unit
Pollutant
NOx*
VOC**
CO
Particulates
SOx
Emissions (Ibs per hour)
6.19- 1.55
1.08
1.55
0.13
0.56
Emissions (Ibs per MMMbtu)
140-35
25
35
3.0
13
* Emissions based on two different types of burners, the upper range value is for a conventional
burner provided by Mobil Oil, lower range value is our estimate for a for an ultra-low NOx
burner which emits 75 percent less NOx per unit energy consumed.
** Emissions based on a unit monitored with a leak detection and maintenance program.
Next we revisit the increased energy demand due to desulfurizing gasoline for PADD 3
from Tables C-2 through C-5 above which are based on the four desulfurization technologies
which formed the basis for our energy analysis. Based on the estimated increased PADD 3
energy demand, we then estimate the increased energy demand and resulting emissions increases
for an average-sized refinery, which would be producing about 70,000 barrels per day gasoline or
refining about 133,000 barrels per day crude oil, and for the U.S. refining industry. Like the
PADD 3 energy analysis, we aggregate the two improved desulfurization technologies for one
estimate, and aggregate the two adsorption technologies for the second estimate.
The actual emission increases from any given refinery could be more or less than our
C-10
-------
Appendix C: Refinery Energy and Global Warming Impacts and Emissions
estimates, due to several refinery-specific factors, such as the specific type of gasoline
desulfurization technology chosen, the actual amount of gasoline produced per barrel of crude
oil, the refinery's baseline gasoline sulfur level, and the actual level of emissions control.
Therefore, EPA's estimates should in no way be viewed as the actual level of emission increases
from a given refinery, but rather, as a relative estimate of refinery emissions increases, for
general comparison with the benefits of the Tier 2/gasoline sulfur program. The true emission
increases can only be estimated by specific refineries as they prepare permit applications for
gasoline desulfurization projects. Refiners have also indicated that they may be making
modernizations, plant expansions, or debottlenecking changes along with gasoline
desulfurization changes. All of these changes, while not attributable to the Tier 2/sulfur program,
could nevertheless result in emission increases higher than EPA's estimates.
The emission estimates for an average-sized refinery and for the U.S. refining industry are
summarized in Table C-8 below.
C-ll
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Table C-8. Projected Increased Energy Use and Increased Pollutant Emissions of an
Average-Sized Refinery and the U.S. Refining Industry due to Meeting the 30 ppm
Gasoline Sulfur Standard (2004)
Technology Name
Energy Demand (MMBtu/year)
Improved Desulfurization
Technologies
CDTech & Octgain 220
Energy Demand (MMBtu/year)
Adsorption Desulfurization
Technologies
Black and Veatch & Phillips
Increased
Energy
Demand/or
PADD3
(MMMBtu/yr)
53,200
32,300
Projected Increase
in Energy Use and
Emissions for an
Average-Sized
Refinery
(Tons/year)
1830
NOx 130 -
32
VOC 11
CO 32
Particulate 3
SOx 12
1110
NOx 80 -
20
VOC 7
CO 20
Particulate 2
SOx 7
110,400
NOx 7730 -
1930
VOC 670
CO 1930
Particulate 165
SOx 695
72,500
NOx 5100-
1270
VOC 440
CO 1270
Particulate 110
SOx 460
C-12
-------
Appendix C: Refinery Energy and Global Warming Impacts and Emissions
This analysis shows that the projected increase in refinery emissions due to desulfurizing
gasoline is trivial compared to the significant emission reductions expected from the tighter
motor Tier 2 vehicle standards. For both NOx and VOC and considering emissions in 2010, the
increased emissions from refineries is less than one quarter of one percent of the decrease in NOx
emissions reductions due to motor vehicle standards. In future years, as Tier 2 vehicles phase
into the fleet, emission reductions from motor vehicles are expected to outpace emission
increases in the refinery, and as adsorption technologies replace proven and improved
desulfurization technologies in future years, the difference is expected to be even greater.
C-13
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Appendix C References
1. Wang, M.Q., GREET 1.4 - Transportation Fuel-Cycle Model, Center for Transportation
Research, Argonne National Laboratory.
2. Emissions of Greenhouse Gases in the United States 1998, Energy Information
Administration, October 1998.
3. Chuba, Mike, OCTGAIN Environmental Impact Evaluation, January 19, 1999.
C-14
-------
Appendix D: EPA's Legal Authority for Gasoline Sulfur Control
Appendix D: EPA's Legal Authority for Gasoline
Sulfur Control
We are adopting 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 our current our requirements that affect
gasoline sulfur content, and explain our bases for controlling gasoline sulfur under Section
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
D-l
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
sulfur content.1
All gasoline is subject to Section 21 l(f) of the Act, which prohibits fuel or fuel additive
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 4814-882), which limits the sulfur content
of unleaded gasoline to 0.1 weight percent (1000 ppm) sulfur.
B. How the 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 promulgating controls on 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
'Because sulfur is directly or indirectly controlled by EPA requirements, and will be controlled directly
under today's action, states are preempted from initiating sulfur control programs unless they are identical to the
federal requirements. See the discussion in Section IV.C. of the preamble on this subject.
2 Standard Specification for Automotive Spark-Ignition Engine Fuel
D-2
-------
Appendix D: EPA's Legal Authority for Gasoline Sulfur Control
more detail in Chapter HI above, 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 Tier 2 standards. The health and welfare
implications of the emissions of these compounds are discussed in greater detail in Section HI 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
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.
EPA has also satisfied the statutory requirement to consider "other technologically or
economically feasible means of achieving emission standards under section [202 of the Act]."
This provision has been interpreted as requiring consideration of establishing emissions
standards under § 202 prior to establishing controls or prohibitions on fuels or fuel additives
under § 211(c)(l)(A). See Ethyl Corp. v. EPA, 541 F.2d. 1, 31-32 (D.C. Cir. 1976). In Ethyl, the
court stated that § 21 l(c)(2)(A) calls for good faith consideration of the evidence and options, not
for mandatory deference to regulation under § 202 compared to fuel controls. Id. at 32, n.66.
In today's action, EPA for the first time is establishing standards for fuels and vehicles
together. Thus, it is first important to consider that the sulfur standards are not being adopted as
an alternative to vehicle emissions standards, but in addition to such standards, and as a
necessary prerequisite to ensuring that vehicles can meet the vehicle standards. In addition, the
Tier 2 standards being adopted today will begin phasing-in in 2004, but will not be fully phased
in for the fleet of new motor vehicles until 2009, and even at that time, many non-Tier 2 vehicles
will still be on the road. Thus, another point to consider is that the emissions standards under §
202 will achieve smaller emissions benefits in the early years of the program and will not achieve
their full emissions benefits for a number of years, while the sulfur standards will achieve
significant, immediate emissions benefits through reducing emissions from the existing fleet of
motor vehicles (primarily Tier 1 vehicles and LEVs), and will continue to achieve increasing
benefits as the fleet turns over to Tier 2 vehicles, especially in light of the expected increase in
vehicle miles travelled.
EPA has also considered emissions standards under § 202 that are more stringent than
Tier 2 as an alternative to regulating gasoline sulfur. However, for the reasons described in
Chapter IV, we conclude that the Tier 2 standards represent the level of emission control that is
economically and technologically feasible from new motor vehicles in the time frame over which
the standards are implemented. Moreover, we considered Tier 2 standards without control of
D-3
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
gasoline sulfur as an alternative to regulating gasoline sulfur. However, we have concluded that
the Tier 2 standards would not be feasible without control of gasoline sulfur. For these reasons,
we find that the alternatives of either more stringent vehicle standards, or Tier 2 standards
without sulfur control, are not technologically or economically feasible options to regulating
gasoline sulfur.
Our consideration of other technologically and economically feasible means of achieving
emissions standards under § 202 of the Act supports the conclusion that the sulfur standards
adopted today represent an appropriate exercise of the Agency's discretion under § 21 l(c)(l)(A),
even when the Tier 2 standards are considered.
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 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 adopted today). Thus,
Tier 0 and Tier 1 vehicles have higher emissions when they are exposed to sulfur levels
substantially higher than the sulfur standard adopted today. 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 gasoline sulfur control.
The sulfur impact on the catalysts used in later model vehicles is clearly significant. High
sulfur levels have been shown to significantly reduce the conversion efficiency of 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 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
D-4
-------
Appendix D: EPA's Legal Authority for Gasoline Sulfur Control
ULEV vehicles can experience, on average, a 100 percent increase in NMHC and 197 percent
increase in NOx emissions when operated on 330 ppm sulfur fuel compared to 30 ppm sulfur
fuel (see Appendix B for more details). This level of emissions increase is significant enough
that it would undermine the technical and economic feasibility of the Tier 2 standards.
This level of impact on emission control system efficiency would mean actual in-use
emissions reductions from the 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 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.
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, on average over their long-
term operation, have to use fuel with sulfur levels no greater than 30 ppm to avoid significant
impairment of their emissions control systems. Furthermore, short-term operation on gasoline
with sulfur levels higher than 80 ppm will have a significant adverse effect on the desired
emission performance and will significantly impair 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
D-5
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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.
Therefore, 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
D-6
-------
Appendix D: EPA's Legal Authority for Gasoline Sulfur Control
same level of emission standards without reducing commercial gasoline sulfur levels, as
explained in the next section.
In summary, 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 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 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 Tier 2 standards. While the impact on emissions
control systems of Tier 2 vehicles and LEVs is a sufficient basis to control gasoline sulfur under
Section 21 l(c)(l)(B), a similar analysis for Tier 0 and Tier 1 vehicles also supports 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.
End Of Moved Text
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 fuel control with such devices or systems
which are or will be in general use that do not require the fuel control. As described above, we
conclude that the emissions control systems expected to be used to meet the 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, and Chapters IV, V, and VII
above, including an analysis of the environmental benefits of the fuel control, an analysis of the
costs and the technological feasibility of controlling sulfur to the levels required by today's
action, and a cost-benefit analysis of the 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 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 general use in gasoline-powered vehicles that can meet the 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, as well as in Appendix B above, we cannot identify one or more factors that
definitively determine sulfur sensitivity, because sulfur sensitivity seems to be due to a
D-7
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
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, we anticipate that all the gasoline vehicle
technologies expected to be used to meet the Tier 2 standards will require the use of low sulfur
gasoline. If we do not control gasoline sulfur to the levels required by today's rule, we would not
be able to set Tier 2 standards as stringent as those we are adopting. Moreover, vehicles already
on the road would continue to emit at higher levels than they would if operated on low sulfur
fuel. These benefits from the existing fleet, which are particularly significant in the early years of
the program, cannot be achieved through new vehicle emissions standards. Consequently, we
conclude that the benefits that would be achieved through implementation of the vehicle and
sulfur control programs in today's rule cannot be achieved through the use of emission control
technology that is not sulfur-sensitive.
This also means that if we were to adopt vehicle emissions control standards without
controlling gasoline sulfur content, the standards would be significantly less stringent than those
based on what would be technologically feasible with current sulfur levels, rather than what is
feasible with lower sulfur. In such a situation, the cost of the vehicle emissions control
technology would likely be similar to the costs of meeting the 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 with low sulfur gasoline, 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. Thus, the same benefits achieved by today's
program could not be achieved through vehicle emissions standards alone, because of the
sensitivity of the emissions control technology to sulfur.
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 prohibiting sulfur
in gasoline, but rather controlling the levels of sulfur in gasoline, this finding is not required prior
to regulation. However, EPA does not believe that the sulfur standards adopted today 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
current sulfur levels.
We believe that gasoline formulated to meet the low sulfur standards will have a
D-8
-------
Appendix D: EPA's Legal Authority for Gasoline Sulfur Control
significant net benefit to public health due to reduced emissions of harmful compounds. 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 from the engine itself, we
believe that the improved catalyst performance enabled by the low sulfur fuel will more than
offset any slight increase in harmful emissions from the engine that might result, such that the net
emissions effect of the sulfur control is a substantial reduction in emissions, compared to
emissions from gasoline without the sulfur control.
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 sulfur 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
several options. All of these options, whether increasing the aromatics or olefms 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 action 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.
D-9
-------
Tier 2/Sulfur Regulatory Impact Analysis - December 1999
Appendix D References
1. EPA Staff Paper on Gasoline Sulfur Issues (EP A420-R-98-004), May 1998.
D-10
------- |